ChatGPT’s advanced voice mode showed us what’s possible with interactivity. Notebook LM’s audio summaries showed us what’s possible with grounded conversational audio generation.

Both are incredible accomplishments and useful tools for thought.

What’s still missing? Let’s time travel in our DeLorean and see what the future may hold…

1. Auto-generation. Today, sources need manual addition to Notebook LM, and audio needs regeneration. What if the product did this automatically? E.g., read my newsletter subscriptions or Slack messages from the last 24 hours and summarize them in audio. Additional sources: YouTube playlists, Notion folders, etc. Subject to constrains and customizations.
   
2. Customization. While both products accept custom instructions, they don’t strictly follow them. This makes it hard to optimize content for specific use cases. Users will likely choose different settings for different use cases. Likely, smart UI defaults/knobs would be helpful in addition to text/voice instructions. Potential dimensions:
   • Length
   • Tone and format
   • Number of speakers
   • Abstraction/summarization level
   • Educational density vs. entertainment/engagement
   • Strict source adherence vs. source-inspired exploration
     
3. Interactivity. It would be powerful to marry interactivity (advanced voice mode) with high-quality generated content (audio summaries). Imagine joining a pre-generated podcast about any topic as a third participant—either listening silently or interrupting to ask questions and redirect the conversation. This would make many GPUs go brr to achieve sufficiently low latency.

Some hallucination would still be acceptable for these use cases, so we’ll probably build them before we build truly reliable “nine nines” agentic behavior that is high-stakes.

Notes:

1. For those unfamiliar: advanced voice mode enables natural conversation with the LLM, and Notebook LM audio summaries create AI-generated “podcasts” about anything.

You can do this by fine-tuning OpenAI’s language models (GPT) on your own data. All you will need is the OpenAI API and two open-source libraries (LlamaIndex and Langchain). You can run this in a Google Colab (Jupyter) notebook for free. If this sounds interesting, here is how:

Steps

1. Open AI. Sign up for Open AI and get an API key.

2. Upload your content to GitHub.

3. Use a Colab notebook.

• To simplify things, we’ll use Google Colab. It runs in the cloud, so you won’t need to install anything on your laptop.
• To get started quickly, re-use a public Colab that has all the necessary code. This way you would only need to make a few changes. For example, you can copy this Colab (credit goes to Dan Shipper) to your Google Drive. For context, see Dan’s article in Lenny’s newsletter and this video walkthrough by Irina Nik.
• Replace the GitHub link with your own.
• Change runtime type to GPU (Menu: Runtime -> Change runtime type).
• Run all the Colab cells.
• Enter your OpenAI API code when prompted.
• Ask any question when prompted.

And that’s it!

You can find a copy of the complete code below:

! git clone https://github.com/[insert your repo].git

!pip install llama-index
!pip install langchain

from llama_index import SimpleDirectoryReader, GPTListIndex, readers, GPTSimpleVectorIndex, LLMPredictor, PromptHelper
from langchain import OpenAI
import sys
import os
from IPython.display import Markdown, display

def construct_index(directory_path):
    # set maximum input size
    max_input_size = 4096
    # set number of output tokens
    num_outputs = 2000
    # set maximum chunk overlap
    max_chunk_overlap = 20
    # set chunk size limit
    chunk_size_limit = 600 

    # define LLM
    llm_predictor = LLMPredictor(llm=OpenAI(temperature=0.5, model_name="text-davinci-003", max_tokens=num_outputs))
    prompt_helper = PromptHelper(max_input_size, num_outputs, max_chunk_overlap, chunk_size_limit=chunk_size_limit)
 
    documents = SimpleDirectoryReader(directory_path).load_data()
    
    index = GPTSimpleVectorIndex(
        documents, llm_predictor=llm_predictor, prompt_helper=prompt_helper
    )

    index.save_to_disk('index.json')

    return index

def ask_ai():
    index = GPTSimpleVectorIndex.load_from_disk('index.json')
    while True: 
        query = input("What do you want to ask? ")
        response = index.query(query, response_mode="compact")
        display(Markdown(f"Response: <b>{response.response}</b>"))

os.environ["OPENAI_API_KEY"] = input("Paste your OpenAI key here and hit enter:")

construct_index("context_data/data")

ask_ai()

Use Cases

I have found that ChatGPT is most consistently useful in the role of an expert, tutor, and editor.

1. Expert and Tutor

When: I want to learn X.

Expertise vs Tutoring

These are very similar conceptually, but I would make a small distinction:

• Expertise. At times, I may want to get a brief overview (tldr) of a topic. This could be viewed as a quick chat with an expert. For example, if I wanted a brief summary of the life extension industry, but didn’t want to dive too deep, I could think of it as a quick chat with a friend who works in the industry.
• Tutoring: In other instances, I may want to dive deeper into a specific topic and ask a series of questions. For instance, while I was learning about machine learning with PyTorch, I found ChatGPT to be very helpful as a tutor.”

How

• Know Where to Start. If I don’t know where to begin, a general introductory question might be helpful, such as “What are the most important topics to master in order to learn X?”
• Troubleshoot. If I encounter a specific problem or question, it’s best to address it directly. I’ve found that this approach works particularly well with programming questions, at least at a basic level.

Best Practices

• Creative Prompting: Sometimes, getting creative with prompts can help. For example, instead of asking, “Explain general relativity as if I’m five years old,” it may be helpful to try, “Explain general relativity as if I’m five years old. Please respond as a physics professor at MIT who is good at communicating complex ideas without dumbing them down.”
• Follow-up: In my experience, one of the main advantages of ChatGPT over traditional search engines and previous-generation AI assistants or chatbots is its interactivity and ability to maintain the thread of the conversation. So, just like with a real tutor, it may be helpful to ask for explanations in a different way, provide an example, elaborate on a particular point, or step back and explain how a certain concept fits into the big picture.

3. Editor

When: I want to change anything about a text.

How

Some examples include:

• Checking grammar
• Summarizing or rephrasing the text
• Getting ideas on what else to consider and think about
• Analyzing the argument. For example, “What are the likely influences behind this text?” or “What are the strongest counter-arguments to this?”

Context

Lessons and Challenges

1. Hallucinations and Validation. Of course, one must watch out for hallucinations. That’s why I find that using ChatGPT in combination with Google may work best. The first step may be asking ChatGPT to provide an initial overview of a topic to save time reading long articles from the Google’s organic results that have been SEO’d in and out. And the next step may be to search for specific sub-topics — sometimes using search syntax/operators to find exact matches or limit the search to certain websites.
2. Common Knowledge vs. Advanced/Niche Topics: I’ve found that ChatGPT is very good at summarizing and explaining common non-controversial knowledge, but is more likely to hallucinate or struggle when dealing with advanced or niche topics that may only be covered by a just few papers or blog posts. Another way to think about this is to consider how likely it is that this topic was prominent in the training dataset.

What I’ve Tested

Despite still relying on Google for the majority of my searches, I have come to the conclusion that ChatGPT — especially when combined with Google — is better or comparable for a specific subset of my use cases and queries. Depending on the question or query, I have also found Perplexity.ai, Neeva, You.com/chat, and Elicit.org to be worth trying out.

What’s Next

I am looking forward to trying Microsoft’s integration of ChatGPT with Bing and Google’s integration of Bard with Google Search. I find the intersection of search and conversational AI to be particularly intriguing, as these two tools are some of the best resources for organizing and retrieving information that we have today.

Examples

Please see — or skip — the three examples of actual “conversations” below:

Conversation 1. Expert: Industry Overview

– Me:

Can you give an overview of the life extension industry? I’m interested in understanding the structure of the ecosystem, encompassing both research and business aspects, as well as the major players, trends, and challenges.

– ChatGPT:

Sure! The life extension industry, also known as the anti-aging or longevity industry, is a relatively new and rapidly growing field that aims to extend the human lifespan and improve the health and well-being of individuals as they age.

The life extension industry encompasses a wide range of areas including pharmaceuticals, biotechnology, gerontology, and nutrition. Some of the key players in the industry include pharmaceutical companies, biotech firms, research institutions, and anti-aging clinics.

In terms of research, the focus is mainly on developing drugs, treatments, and interventions that can target the biological mechanisms of aging and reduce the risk of age-related diseases. This includes researching senolytics (drugs that can target and eliminate senescent cells), nutraceuticals (nutritional supplements that can help improve health), and gene therapy (using genetic manipulations to extend lifespan).

The business aspect of the industry involves the development, marketing, and sale of products and services aimed at the anti-aging and longevity market. This includes cosmetics, dietary supplements, and fitness equipment, as well as more cutting-edge technologies such as wearable health monitors and bioregenerative devices.

There are also a growing number of anti-aging clinics and spas that offer personalized health and wellness plans, and some also offer anti-aging treatments such as hormone therapy and intravenous infusions of nutrients.

The life extension industry is still in its infancy and faces many challenges, including regulatory hurdles, scientific skepticism, and public perception. Additionally, the high cost of many anti-aging treatments and lack of long-term data on their safety and efficacy can make it difficult for individuals to make informed decisions about whether to invest in these products and services.

Despite these challenges, the life extension industry is growing rapidly and attracting a lot of investment and attention. With advances in technology and an aging global population, it’s likely that the industry will continue to evolve and expand in the coming years.

– Me:
Can you name the most well-regarded, prominent, and promising players?

– ChatGPT:

Sure! Here are some of the most well-regarded, prominent, and promising players in the life extension industry:

1. Calico: Calico is a research-focused subsidiary of Alphabet Inc. (Google’s parent company) that is focused on the biology of aging and age-related diseases. Calico is one of the largest and most well-funded players in the life extension industry, and it has partnerships with a number of top biotech and pharmaceutical companies.
2. Unity Biotechnology: Unity Biotechnology is a biotech company that is focused on developing drugs that can target and eliminate senescent cells, which are thought to contribute to aging and age-related diseases.
3. Juvenescence: Juvenescence is a biotech company that is focused on developing drugs to extend lifespan and improve healthspan. The company has partnerships with a number of leading research institutions and is working on a number of promising drugs for age-related diseases.
4. Insilico Medicine: Insilico Medicine is a biotech company that is using AI and machine learning to develop drugs for age-related diseases. The company has partnerships with a number of leading pharmaceutical companies and is working on a number of promising drugs for age-related diseases.
5. Geroscience: Geroscience is a biotech company that is focused on developing drugs that can target the biology of aging and age-related diseases. The company has partnerships with a number of leading research institutions and is working on a number of promising drugs for age-related diseases.
6. Human Longevity: Human Longevity is a health and wellness company that is focused on using technology and data to improve health and extend lifespan. The company offers a range of products and services, including anti-aging treatments and health assessments, and it has partnerships with a number of leading healthcare organizations.

These are just a few of the most well-regarded, prominent, and promising players in the life extension industry. The field is constantly evolving and new players are entering the market all the time, so this list is by no means exhaustive.

Conversation 2. Tutor: Lambda Functions in Python

– Me: ELI5 lambda function in Python

– ChatGPT:

A lambda function in Python is a small anonymous function, which can take any number of arguments, but can only have one expression. The syntax for a lambda function is:

lambda arguments : expression

Lambda functions are used in cases where you need a quick throw-away function for a short period of time. They are often used as arguments to higher-order functions (functions that operate on other functions).

For example:

# A normal function to add two numbers
def add(x, y):
    return x + y

# An equivalent lambda function
add = lambda x, y : x + y

In the above example, add can be used like a normal function to add two numbers.

– Me: What are anonymous functions?

– ChatGPT:

An anonymous function is a function that has no name and is not bound to a variable. Anonymous functions are also called “lambda functions” in Python, because they are created using the lambda keyword. Anonymous functions are used when you need a small, one-time-use function for a short period of time.

Because anonymous functions have no name, they cannot be reused or referenced outside of the expression in which they are defined.

For example:

# An anonymous function to square a number
lambda x: x**2

# An anonymous function to add two numbers
lambda x, y: x + y

These anonymous functions can be assigned to variables, passed as arguments to other functions, or used in other ways, just like any other function.

– Me: So if it has no name, how do I call it?

– ChatGPT:

You can call an anonymous function by assigning it to a variable and then calling the variable as if it were a regular function. Alternatively, you can call the anonymous function directly by passing it as an argument to another function, or by using it in an expression.

For example:

# Assigning an anonymous function to a variable
square = lambda x: x**2
result = square(10)
print(result) # 100

# Calling an anonymous function directly
print((lambda x: x**2)(10)) # 100

# Using an anonymous function as an argument
numbers = [1, 2, 3, 4, 5]
squared_numbers = list(map(lambda x: x**2, numbers))
print(squared_numbers) # [1, 4, 9, 16, 25]

Conversation 3. Editor: Launch Announcement

Had ChatGPT been available in 2020, we could have used it to edit our product launch announcement. For example:

– Me:
Please respond as a highly skilled editor who is a native speaker of American English. Your writing is succinct, articulate, and accessible. Can you please edit and fix grammar in this text?

“Today my co-founder and I are launching the product we’ve been working on for months. It’s an odd time to launch anything but we’re proud of what we’ve built and we hope that our product can also help some people who were laid off recently. A Happy Job is a team-based and culture-centric recruiting platform. We help job-seekers — engineers at first — find teams that are actively hiring and fit their preferences. Search by product type, tech stack, team set-up, culture, values, working styles or perks.”

– ChatGPT:

Today, my co-founder and I are launching the product we have been working on for months. Despite the current circumstances, we are proud of what we have built and hope that our product can assist those who have recently been laid off.

“A Happy Job” is a team-focused and culture-oriented recruitment platform. We help job seekers – specifically engineers – find teams that are actively hiring and align with their preferences. You can search for teams based on product type, technical stack, team structure, culture, values, work style, and even perks.

In This Post

• Show Me the Numbers
• Can We Measure Progress?
• A National Dashboard
• Upstream Drivers of Long-Term Progress
• A Possible Set of 11 Metrics
• More Options
• Global Focus

Show Me the Numbers

As someone who has studied economics and business and spent over a decade working in the private sector, I’ve been long dismayed by how little focus there is on numbers in government and media reporting. To the extent politicians focus on numbers, they typically pick the ones that support their narrative and seem relevant to this week’s news cycle.¹

I wonder if there’s a better way. What if we reversed the order? What if we came together as a society and determined (1) what areas we would most like to improve, (2) the best way to measure the progress, and (3) the best way to report on it and future plans? What if we then used this framework to guide public conversation, media reporting, and perhaps even some White House press briefings?

The idea is not new, of course. And many people have already tried to popularize it. Recently, Andrew Young got many people excited when he proposed the idea of the American scorecard and — taking the idea to the extreme — suggested that the president use PowerPoint to report on progress.

So what’s the problem?

The problem isn’t as much that we aren’t measuring things.² The problem is threefold:

1. There is little national and global public consensus regarding which indicators matter most.
2. There is limited accountability when it comes to improving these numbers and publicly reporting on them in a structured way and with a regular cadence.
3. Media focus can sometimes be misplaced, haphazard, pessimistically biased, localist, and short-term focused.

Can We Measure Progress?

First, we might want to agree on what human progress is. And since this conversation can get rather complicated and heated quickly, it might make sense to focus on things things that people universally value. Here is a relevant quote from Steven Pinker’s Enlightenment Now:

“What is progress? You might think that the question is so subjective and culturally relative as to be forever unanswerable. In fact, it’s one of the easier questions to answer. Most people agree that life is better than death. Health is better than sickness. Sustenance is better than hunger. Abundance is better than poverty. Peace is better than war. Safety is better than danger. Freedom is better than tyranny. Equal rights are better than bigotry and discrimination. Literacy is better than illiteracy. Knowledge is better than ignorance. Intelligence is better than dull-wittedness. Happiness is better than misery. Opportunities to enjoy family, friends, culture, and nature are better than drudgery and monotony. All these things can be measured. If they have increased over time, that is progress.”

It’s true that “not everything that counts can be counted, and not everything that can be counted counts”. But we can still count most things that matter.

A National Dashboard

Imagine a regularly updated dashboard that shows 10 to 20 metrics that are widely viewed as important:

• Click on a metric and see the latest plan — a one-pager updated quarterly and aimed at improving the metric in the 1-year to 5-year horizon.
• Click on “previous plan” to see how well the plan from the last quarter worked — a one-page post-mortem.
• Ask yourself if the new plan is better.
• Submit your questions and upvote others’.
• Then hear the answers in the next live-streamed 30-min structured presentation and 30-min Q&A where questions with the highest number of upvotes minus downvotes get answered first.

I think for the idea to gain wide adoption in a given country, people would need to collectively answer several questions. When there’s consensus, there’s a chance to create accountability.

Here are just a few to ponder:

1. Metrics. What should we track?

• What matters most at the national level?
• What can be measured well and is hard to manipulate?
• How can the full set of metrics balance individual ones out?
• Can we track leading indicators as well as lagging indicators?
• What metrics are so straightforward that everyone can understand them?
• What incentives will be strengthened and weakened if we focus on these metrics — including perverse incentives?

2. Timeframe. What’s the right frequency and meaningful timeframe?
   
3. Accountability. Who should be accountable, who should present, how detailed should the presented vs. published plans be, etc.

Upstream Drivers of Long-Term Progress

One danger of looking at lagging indicators only is that they might not be actionable. And we might not take action necessary to ensure long-term progress. So we should also be asking ourselves what upstream drivers of progress are. In other words, we should consider leading indicators as well.

What drives human progress is a matter of ongoing debate, but let’s assume that the following three are at least a major part of the picture:

1. Universal global access to education and opportunity.
2. Fundamental research in universities and labs: theoretical physics, etc.
3. Real-life application and distribution of new technologies: new startup creation, engineering, corporate R&D, etc.

These three create a strong virtuous circle. More talented people → more foundational knowledge → more prosperity.

With all that in mind, I decided to do a fun thought exercise and come up with a possible set of metrics that could capture how we’re doing at the national level.

A Possible Set of 11 Metrics

Economy:

1. GDP (PPP) per capita growth
2. Growth in median net disposable income
3. Growth in median net worth

Well-being:

4. Median life expectancy and healthy life expectancy
5. Subjective life evaluation (aka happiness)
6. Depression rates
7. Crime rate
8. Suicide rates

Long-term/leading indicators:

9. Education quality
10. Funding of fundamental research
11. New business creation

For each metric report on:

1. How has the metric changed over time?
2. Where do we stand in the global ranking today?
3. How has our position changed over time?

So what could this hypothetical dashboard look like for the US in 2021? Let’s take a snapshot of these metrics.

Economy

1. GDP (PPP) per Capita Growth

Source: Worldbank.org

Source: International Monetary Fund, 2020

The GDP metric has many problems. Not all goods and services are equally valuable to humankind. However, it can be a good proxy and starting point. We know that growth, in itself, has been historically important for getting people out of the zero-sum game mindset. There have also been multiple attempts to account for the limitations of this metric.

2. Growth in Median Net Disposable Income

Looking at median numbers should help us partially account for inequality.

Source: OECD Statistics

An important question here is what constitutes “net”. What should we adjust for? For example, should we account for debt, rent, and healthcare?

If we want to create the right incentives to increase everyone’s prosperity, should we subtract rent payments, loan payments, and healthcare expenditures — given how much is spent on these three and given the differences among countries?

In other words, does a person who earns a post-tax income of $100K but spends $80 on rent, student loan payments, health insurance, and medical bills, really have $100K in disposable income?

3. Growth in Median Net Worth

Source: Credit Suisse’s annual Global Wealth Databook, 2019

I also wonder about the best way to adjust these numbers for the cost of living. Doing this might be more tricky than it might seem given the differences in how much different goods and services cost in various countries and in different regions/states.

For example, healthcare is notoriously expensive in the US and arguably more important than some other services and goods. In fact, healthcare, education, and housing — three things that people tend to value, all get more expensive despite the inflation rate staying relatively low:

Source: Aei.org and Bls.org

Well-Being

4. Median Life Expectancy and Healthy Life Expectancy

US life expectancy:

Source: United Nations

Source: World Health Organization

Healthy life expectancy is life expectancy minus years of living with disability.

Top 30 countries:

Source: World Health Organization, 2019

Source: Ourworldindata.org

In my opinion, the biggest challenge in tracking metrics such as life expectancy and other measures of well-being is that it is difficult to increase them rapidly. It may be more productive to ask questions such as “Where should we invest today to achieve the greatest improvement in 10 years?” But this timeframe does not align well with four-year terms. However, this issue can be resolved. For instance, long-term plans that include “fail fast” types of validation and testing can be developed.

5. Subjective Life Evaluation (AKA Happiness)

Source: World Happiness Report, 2021

I wasn’t able to find a chart showing how the US score reported by WHR has changed over time, but Gallup is reporting a similar metric:

Source: Gallup

6. Depression Rates

Source: Ourworldindata.org

Source: Ourworldindata.org

7. Crime Rates

Source: Pew Research

Source: Ourworldindata.org

Top 30 countries by homicide rate:

Source: United Nations

8. Suicide Rates

Source: Ourworldindata.org

Source: Ourworldindata.org

The top list looks noisy — likely due to underreporting.

Long-Term and Leading Indicators

9. Education Quality

This one is tricky to measure, just like healthcare quality. But there have been attempts. This is one of them:

Source: Ourworldindata.org

Top countries:

Source: Ourworldindata.org

It’s not easy to find a chart showing how the US scores and ranking have changed over time.

Another thing to consider is the universities. The US is dominating the list of best universities in the world, but there is not an easy way to quantify that dominance and track it over time. Perhaps citations in the relevant fields (e.g. STEM) can be considered as a proxy?

10. Funding of Fundamental Research

Total R&D funded by the US government (red line). It might be better to show it as a percentage of GDP:

Business-funded R&D is great, but it cannot replace government funding. Businesses and venture capital funds are quite risk-averse. Despite what VCs might claim, many may be reluctant to invest in technologies with long-term time horizons and unclear payoffs.

As % of GDP by country:

Source: Royalsociety.org

11. New Business Creation

Here is an imperfect proxy. It might also make sense to consider new business formation in certain sectors, e.g. technology.

Source: Census.gov

More Options

Here are some additional metrics I would consider:

• Leading indicators of progress. More and better long-term metrics and leading indicators of progress. How do we know we are accelerating it?
• Big long-term bets. Additional long-term metrics focused on risks/opportunities, such as automation and AI, climate (e.g. CO2 in the atmosphere), etc.
• Government efficiency. Can we track and regularly report on the total spend and number of people employed by various agencies? Can we ensure both go down due to increased efficiency, collaboration with the private sector, and automation?
• Well-being. Additional well-being metrics, such as healthcare, preventable deaths, poverty, equality, social trust levels, etc.
• Economy. Additional economic metrics, such as intergenerational economic mobility, unemployment, labor participation, etc.
• Individual rights and freedoms. Economic freedom, personal freedom, press freedom, etc.
• Social institutions. Independent courts, fair elections, corruption, etc.
• Sectors. Top 2–3 metrics for each major sector of the economy. For example, the average train speed and deaths/injuries per distance traveled can be used to evaluate progress in the transportation sector.
• States and counties. Top priority metrics on the state and county level.

Global Focus

This post was in a way US-centric. I believe developing a framework for global reporting might be even more important than driving accountability and focus on the national level. This is because so many countries are still lagging in terms of many of these metrics, but also in terms of governments’ accountability. Sustainable Development Goals are great but they focus more on the current problems and may miss long-term opportunities.

Beyond this, we should also consider system-level metrics in addition to national ones. For example, global mobility, international trade, global trust, and other ways to measure global collaboration. 

But let’s get back to the idea of the national dashboard…

Open Questions

What else would you like the government to regularly report on and present plans for?

Where would this approach fail?

How would you improve it?

If you like my articles, please consider signing up for radically infrequent email updates. And say hi on Twitter!

Notes

[1] And even when there is a continuous focus on a set of numbers, it’s often on seemingly arbitrary numbers. For instance, even the Wall Street Journal website still seems to suggest that performance of 30 arbitrarily selected companies is the most important metric to focus on, and that the most meaningful time interval to consider is 24 hours. ↺

[2] There is also amazing Our World in Data founded by Max Roser, a website that aggregates important metrics from multiple sources and shows them in a user-friendly way. And there are many government agencies and independent organizations doing analysis and reporting on all kinds of metrics. ↺

Cross-posted to LessWrong.com.

The first response that usually comes to everyone’s mind is “of course, the second decision was better”. But this is not necessarily the case. Can you think of a scenario where your response would be the opposite?

I’ve been thinking and reading about decision-making for many years while trying to put everything I learned into practice. Below, I summarize the strategies and mental models that I personally found most useful.

The Process vs. The Outcome

We all have a natural tendency to judge decisions based on their outcomes. This is not the worst heuristic, as there is a correlation between the quality of decisions and outcomes. But this heuristic has a major flaw — it doesn’t account for luck and incomplete information.

An individual can make a bad impulse decision yet get lucky and get a positive outcome. An extreme example could be buying a lottery ticket and winning. A less extreme example could be taking a high-stakes risky business decision without putting much thought into it. Should those decisions still be considered good?

On the other hand, another individual might dedicate the appropriate time to making a business decision, follow a thorough process, analyze all the available information, make the reasonable conclusions, and make an educated long-term bet. Now imagine that the business environment transforms in the upcoming years in a way that could not have been predicted. So the decision leads to a negative outcome. For example, a product launch fails. Does it mean the decision was bad?

Arguably, the first individual made a bad decision, and the second one made a good one. This might contradict conventional wisdom, but if you really think about it, the first person simply got lucky after following a bad process. And the second person did everything they could to make the optimal decision and took a calculated risk. That the outcome wasn’t positive in this particular case doesn’t tell us much about the person’s decision-making ability or judgment.

One reason it’s so hard to evaluate the process instead of the outcome in the business context is that it flies in the face of many other best practices. Hiring managers believe that past performance is the best predictor of success. Leaders extol results-driven culture. OKRs are set around outcomes — not processes.

But a deeper reason is that it’s more difficult to evaluate and establish processes. It’s much easier to evaluate results. Evaluating processes would require a deeper understanding of the context and how that context has evolved over time. It would also require disentangling various drivers of the outcome. You might be thinking… Who’s got time for that?

Realistically, I don’t see the outcome-based evaluation of decisions going away completely any time soon. But it doesn’t mean we cannot iteratively improve the process so we get better over time.

So what can we do to improve the process?

Decision Journal

A simple yet powerful idea is to document big decisions. Just like venture capitalists prepare detailed investment memos before investing in startups where they weigh all the pros and cons, we can also write summaries for big decisions in business and life.

In addition to outlining all the available options and considering their pros and cons, here are some additional things to consider.

The Main Goal and Frameworks

Asking “what are we optimizing for?” might sound obvious but clarifying what you’re optimizing for can simplify the entire process.

Asking this question is particularly useful when a group of people is making a decision. You might assume that everyone has the same goal in mind, but once you ask, you might realize that people might have different opinions about the most important goal or metric to optimize for.

And if you’re into spreadsheets, you can even come up with a list of criteria and assign weights to calculate the expected outcomes. For example:

Expected value = the probability of a certain outcome * the payoff in case of this outcome.

The payoff can either be expressed in monetary value or subjectively perceived value — for example, how happy you would be on the 10-point scale.

Developing a framework like this can be useful for thinking and clarifying what really matters to you and aligning with others. But quantifying things can also give you a false sense of confidence so I’d use this approach with caution.

Want to take it up a notch? Look into decision trees.

Expected Value vs. the Worst Possible Downside

Besides considering the most likely outcome or the expected value, I find it very useful to consider the worst possible downside.

Imagine you have two investment opportunities. One is expected to grow 7% annually on average in the long-term and the other is expected to grow 10%. Based on this information alone, the second one looks like a better investment. At least until you account for volatility. The extreme case of volatility is that the investment goes to zero or close to zero.

I personally find it useful to think about the worst possible outcome I would be comfortable with in addition to considering the expected value.

What if I told you that the value of the first investment has less than a 0.01% chance of dropping to near zero, while the value of the second investment has a 20% chance of doing so, how would it change your decision?

A Portfolio of Bets

Your response to the previous question might have been “it depends”. It might depend on a number things — such as your risk aversion, personal situation, or whether this is the only investment or one of many.

The portfolio approach is often used in finance. It’s been shown that most people would be better off allocating most of their investments into low-fee portfolios that track the overall market rather than picking individual stocks. Similarly, venture capitalists always spread out their investments across many startups.

And a similar framework can be applied in business and life — it’s sometimes useful to ask yourself a question “am I making and testing multiple bets here or primarily committing to one thing?”

Optionality vs. Efficiency

It’s useful to consider the trade-off between future optionality and efficiency.

Some business decisions might bring in more revenue or reduce costs, but they reduce your future options. Many companies optimize for efficiency so they could grow faster or improve their bottom line. Generally, public markets and quarterly financial reporting also incentivize them to do so. But sometimes over-optimizing leads to reduced optionality in the future.

Nassim Taleb developed and popularized the idea of antifragility which is highly related to optionality. Wikipedia describes antifragility as “a property of systems that increase in capability to thrive as a result of stressors, shocks, volatility, noise, mistakes, faults, attacks, or failures”.

What’s interesting is that sometimes building a system like this or putting yourself in a situation like this requires a certain redundancy which is the opposite of efficiency.

The PPE shortage in the US during the 2020 COVID-19 pandemic is sometimes viewed as a recent example of optimizing for efficiency at the expense of optionality. If you focus on efficiency and optimize for the short-term demand, you might choose to store only a limited amount “just in case” and purchase the equipment from countries that can provide it at a lowest cost. However, if you want to optimize for optionality, resilience, and antifragility, you might choose to have more redundant storage and manufacture some equipment internally — so you don’t rely on other countries as much.

Similarly, certain choices in life and business create more options in the future than others.

Let People Form Their Own Opinions First

To quote Daniel Kahneman, a Nobel prize-winning psychologist and economist:

“Subjectively, it feels like you believe in something because you have the arguments for it. But it works the other way around. You believe in the conclusion, and then you create supporting arguments. That’s fundamental. Why do people believe in these conclusions? Partly because people we love and trust believe in the same conclusion.”

Our brains are naturally biased to pay attention to what others believe and generally follow them. So if one person in a group confidently announces their opinion first, many people will be likely to follow them. You can see this effect on steroids if that first-to-speak person is a boss.

This is why it’s usually better to ask people to form their opinions first and, ideally, write them down before discussing them with others. I made sure my team utilized this strategy when running customer research and focus groups with multiple people being interviewed at a time. Before discussing a topic as a group, we would present participants with a question and then ask them to write their opinions before discussing them. This always leads to a greater diversity of ideas.

First Principles Thinking

It’s also been shown that people are more likely to accept evidence that is supporting their pre-existing beliefs and more likely to challenge evidence that is contradicting them. Motivated reasoning and confirmation bias are well researched.

First-principles reasoning is the opposite of that. You don’t reason by analogy, assume that things can only be the way they are today, or copy what others believe.

Instead, you start with the underlying, well-established facts and build your beliefs and conclusions independently based on these facts.

Or you can start by considering commonly held beliefs and deconstructing them with the “five whys” method to see if they are justified or not.

Elon Musk explains first-principles reasoning in this amazing two-minute video:

Quantify Probabilities

Philip E. Tetlock, the co-creator of The Good Judgment Project and professor at the University of Pennsylvania, recommends assigning specific probabilities to outcomes instead of using vague words like “likely” or “possible”. Language is imprecise. There is a big difference between “possible” and “probable”.

And “probable” means different things to different people. For example, this chart shows what different people infer when they hear these vague estimates:

More context and charts can be found on the University of Illinois website and in this Github article.

I remember how our Strategy professor at UC Berkeley Haas School of Business, made us quantify probabilities of all kinds of events. This doesn’t come naturally and not a panacea but it usually elevates the conversation to another level.

Philip Tetlock also recommends considering many decisions when evaluating the accuracy of one’s predictions (see “portfolio of bets” above) in his book Superforecasting: The Art and Science of Prediction. That’s how “superforecasters” were evaluated. Once you know what probabilities that an individual assigned to certain events and whether the events actually took place, you can estimate their overall accuracy over time.

That being said, operating with specific numbers can give a false sense of security and make people overconfident in the face of unpredictable events.

Don’t Neglect the Emotions

The science shows that emotions play a critical role in our decision making.

For example, one study published by Jennifer Lerner, a Harvard Kennedy School professor, concluded that “fear increased risk estimates and plans for precautionary measures; anger did the opposite” and that “motions also predicted diverging public policy preferences”. Another study by the same professor found that sad people spend more on welfare recipients than angry people, as they’re less to likely blame poor people for their own misfortune. And this study concluded that “fearful people made pessimistic judgements of future events whereas angry people made optimistic judgements”. And yet another one found that “incidental emotions can influence decisions even when real money is at stake”. The list goes on.

So aiming at being mindful of emotions and accounting for their influence might be a wiser strategy than trying to be the real-life version of Star Trek’s Spock.

As a side-note, you might find it interesting that neuroscientists who conducted this study showed that even brain activation patterns differ based on the context and our perception of the decision. Here are the differences between “certain”, “risk”, and “ambiguous” decisions:

The overall conclusion is that certain brain areas activated more when people make “risky” and “ambiguous” decisions. For example, frontal pole of the prefrontal lobe was strongly activated in the “ambiguous” condition.

Don’t Sweat the Small Stuff

Very few decisions require thinking that deep. In fact, most decisions should be done quickly.

Sometimes we should stop ourselves from overthinking decisions that don’t matter much so we could save time and energy for decisions that really matter. Besides, speed is an enormous advantage in itself and moving quickly also generates additional information that, in turn, helps us make better decisions.

Additional Reading

• Book: Thinking, Fast and Slow by Daniel Kahneman
• Book: Superforecasting: The Art and Science of Prediction by Philip E. Tetlock
• Book: Thinking in Bets: Making Smarter Decisions When You Don’t Have All the Facts by Annie Duke

If you like my articles, please consider signing up for radically infrequent email updates. And say hi on Twitter!

I’ve published three new articles since April. Here is the summary:

Marketing/Growth Channel Coordinate System

Choosing, prioritizing, measuring, and optimizing different channels is a major part of any marketing/growth job. In this post, I’m sharing a simple visual guide to selecting the most promising starting channels for testing and understanding their inherent trade-offs.

Read the full post here:
https://max2c.com/marketing-growth-channel-coordinate-system/

Learning How to Learn (And 20+ Studies)

This blog post went semi-viral on Hacker News where it received 108 points.

I’ve been interested in cognitive science and effective learning methods for years. I’ve read multiple books and articles and put many ideas to test. So I thought I’d synthesize my notes into the blog post, add references to scientific studies, and share it with you.

Read the full post here: 
https://max2c.com/learning-how-to-learn/

VC Twitter and More Tech Lists You Might Like

You’ll like this one if you’re looking for some really cool people to follow on Twitter.

Read the full post here:
https://max2c.com/vc-twitter-and-more-tech-lists-you-might-like/

Cheers, I hope you all have an amazing day!

But there is no such thing as the perfect channel that would work for every company and in every situation.

Once in a while, a new channel might emerge and it might perform better than everything else for certain companies for a period of time. But most of the time there are trade-offs to consider.

In this post, I’m sharing a simple visual guide to selecting the most promising starting channels for testing and understanding their inherent trade-offs.

Predictability vs. Cost Trade-off

The performance of some channels is easier to predict before you invest any time or effort. It’s also easier to forecast how they are going to perform in the future. Generally — but not always — these channels tend to cost more as companies are usually willing to pay extra for certainty.

Targeting Precision and Consumer/SMB/Enterprise Focus

One of the most important considerations is how precisely you need to target your prospective customers.

Sometimes you know exactly who your ideal customers are. For example, your product might solve specific problems that only engineering leaders at large healthcare companies face. Or your product might only appeal to teenagers who are into punk rock.

And at other times you’re selling something that’s more like peanut butter. Almost anyone can try it. Well, except people with peanut allergy…

The second consideration is whether you’re in a B2C or a B2B business. And if you’re in a B2B business, it’s to distinguish between channels that are more likely to work for the SMB segment and channels that are more likely to work for the enterprise segments.

Customer Intent and Consumer/SMB/Enterprise Focus

Another important consideration is how aware your prospective customers are about your product category and how actively they are searching for solutions.

For example, a finance manager might actively look for an accounting software solution. Or a suburban parent might look for pizza delivery. These are well-established categories.

However, if you work at an innovative company, you might find yourself with a novel product or product category that customers don’t know they need. In this case, you usually need to educate them about the existence of your product type and what it can do for them.

Additional Axes to Consider

I’ll stop here for now, but here are some additional dimensions to consider:

• Scalability. Even if you can make this channel work, how big can it get?
• Measurability and attribution. How measurable is this channel, how easy is it to attribute new users, customer leads, and revenue to it?
• Product channel fit. Is this a channel where the type of product you’re making is typically researched or sold?
• Audience channel fit. Is this where your audience already is?

What’s Next

After forming a hypothesis about which channels are likely to work well in your situation, define success metrics, and run a series of experiments.

Disclaimers

• These are just general guidelines. Each company is unique. Each product is unique. One could easily come up with many exceptions, “refuting” the location of each channel on these maps. The goal is to outline the general patterns in broad strokes and provide a starting place for people who might not know where to start.
• There is no perfect “mutually exclusive and collectively exhaustive” framework for marketing/growth channels. They’ll overlap no matter how you slice and dice them. There’s also debate on whether some of these channels should be considered even channels or strategies that reinforce multiple channels. For instance, content can be used across multiple channels — from social and SEO to even outbound sales.
• In reality, each of these bubbles can be both wider and taller. But that would make for a messy chart. For example, take podcasts or PR. You can make them more or less targeted depending on specific tactics and partners you choose. For example, if anyone can be your customer, you might partner with podcasts that cover a wide range of topics and whose subscribers include a wide range of demographics. But if you are targeting a very specific decision-maker at a specific company, you might sponsor a niche podcast. It’s the same with PR. That being said, these channels will still on average allow for less targeting than SEM ads, LinkedIn ads or outbound emails.

I hope you found these coordinate system charts helpful. Please feel free to share with your colleagues or friends and let me know if you have feedback!

Primarily technology and startups related:

• VC (Venture Capitalists) and Angel Investors — less parody, more substance
• Tech and Startups
• PM: Product Management
• Marketing and Growth

And beyond:

• Science and Research
• Life Extension and Longevity
• Finance and Economics

Say hi on Twitter and tell me whom I’m missing:

I think you might some of these Twitter lists I’ve curated:

1. Tech and Startups https://t.co/ksgLQFiJpt
2. Product Management and Development https://t.co/nA21N1b8uw
3. Marketing and Growth https://t.co/yg64SgKrpE
4. VC and Angel Investors https://t.co/puxeif2yB5

More

— Max (@maximmm) June 26, 2020

Summary (TL;DR)

Effective Learning Strategies

• Distributed learning. Study less in each session but more frequently.
• Active recall. Actively test your knowledge and skills.
• Distributed recall. Space the tests in time and adjust the intervals based on performance.
• Interleaving. Practice multiple related yet different skills/concepts simultaneously.
• Elaborative interrogation (quiz-and-recall). Ask yourself questions and use the material you’ve learned to answer them.
• Self-explanation and the Feynman technique. Explain what you’ve just learned in simple terms.

Physiology and Brain’s Health

• Sleep
• Exercise
• Nutrition

Disclaimer and Introduction

I have no formal background in cognitive science or neuroscience, and this has been more of a side interest. My understanding is limited, and I still need to learn how to effectively and consistently apply all these ideas to practice.

That being said, I found some of the methods described in this article very useful. For example, I’ve used them to learn foreign languages, the basics of programming, and various disciplines covered during the two-year MBA program.

Effective Learning Strategies

Strategy #1: Distributed (Spaced) Learning Practice

In short, it’s better to distribute one’s practice over a period of time than cram it into one day.

In one study, elementary school students were asked to study in one of the three ways: massed, clumped, and spaced.

• Massed = four lessons at a time
• Clumped = two lessons on one day and two lessons on the next day
• Spaced = one lesson per day for four days

The “spaced” group performed best, followed by the “clumped” group:

Another study compared comprehension scores under three different conditions:

1. Read a text once (“single”)
2. Read a text twice (“massed”)
3. Read a text twice with a week-long gap (“distributed”)

When tested immediately, the second group performed best. But when tested with a delay of two days, the third group performed best.

This method is also superior for learning motor skills.

How to apply this in practice:

Create a learning schedule or find time to practice a little bit every day or every few days instead of cramming all your learning into one or just a few days.

If you’d like to learn more, read the Wikipedia article on distributed practice.

Strategy #2: Active Recall (Retrieval) Practice

It might be more effective to actively retrieve the information you’ve already learned than passively re-read or try to learn it once again.

One study that compared a method that emphasized study sessions with a method that emphasized tests and found the latter to be more effective for delayed recall.

• SSSS = four study sessions
• SSST = three study sessions, followed by one test
• STTT = one study session, followed by three tests

Even imagining that you might be tested on the material you’re learning might help improve the recall.

How to apply this in practice:

If a few days ago you learned how past tense works in the Spanish language, try to remember the rules or even test yourself on your knowledge — instead of simply re-reading the same material once again.

You can read more about the active recall practice on Wikipedia.

Strategy #3: Distributed (Spaced) Recall Practice

Distributed recall practice is basically a combination of the two ideas above. You test yourself frequently and modify the test intervals depending on how familiar you’re with the material or how strong your skill is.

How to apply this in practice:

Many apps simplify the process by tracking one’s performance and adjusting the intervals automatically. Anki is a free and popular app that can be used for learning foreign language vocabulary as well as other things.

Strategy #4: Interleaving

Unlike the methods listed above, this one is not as widely used yet. The idea is simple: practice multiple skills at the same time. This can make practice more difficult but it pays off at the end. Sometimes this strategy is also called “varied practice”.

A 2010 study found that interleaved practice better prepares children for the complex tasks (full problems) — the ones you’re more likely to encounter in the real world — even though they make more mistakes during the practice.

A 2007 study showed that students who mixed different tasks performed worse during the practice session but better in the subsequent test of their skills.

These two studies focused on math tasks — finding the volume of various geometric solids — but similar effects were found in other domains, such as badminton, basketball, and medicine.

How to apply this in practice:

• Interleave skills that are related — for example, different chord progressions
• Mix old and new skills to integrate them

You can learn more about interleaving by reading the Scientific American article and the Wikipedia article.

Strategy #5: Elaborative Interrogation (Quiz-And-Recall)

Elaborative interrogation is a fancy term for asking yourself questions while studying. Sometimes it’s also called “quiz-and-recall”.

One study found that asking questions is useful for acquiring science facts. Another one showed that this method is particularly effective when you already have certain prior knowledge about the subject. But even if you don’t, elaborative interrogation still appears to be more effective than simply reading to understand.

How to apply this in practice:

• Ask yourself a question that starts with “Why” or “How” and try to answer it based on the material you’ve learned.

Strategy #6: Self-Explanation and The Feynman Technique

This strategy is built on a similar idea. After learning something new, try to explain it in your own words. If you have a partner or anyone willing to listen to your explanations, you can practice with them but explaining things to yourself works too.

One study evaluated the percentage of correct answers and found this method effective:

As some professors and teachers note, they not only teach to share knowledge but also to understand the subject better.

The method is similar to the so-called Feynman Technique attributed to the Nobel prize-winning physicist Richard Feynman. In addition to self-explanation, Feynman also emphasized that one should explain things simply so that even a sixth-grade student would understand them.

How to apply this in practice:

• Choose the subject you’d like to learn
• Pretend you’re teaching it to a sixth-grade student
• Notice the gaps in your understanding and use them to guide your further learning

A simple heuristic from my own experience:

I found that if whatever I learn can be practiced, then practice is superior to recall/explanations which are themselves superior to passive learning.

• For example with programming: write code to solve problems > explain how something works > learn passively.
• Similarly, with foreign languages: speak or write > remember/explain what you learned > passive learning.

Other Learning Strategies

• Deep work: turn off all distractions.
• Deliberate practice and chunking: break the big subject down into smaller manageable “chunks” that need to be learned.
• Learn in 25–90-minute blocks. See the Pomodoro technique.
• Take breaks and let your mind wander — both “focused” and “diffused” modes are important.

Biohacking Brain’s Health and Physiology

I have mixed feelings about the term “biohacking’ but I’ll use it here for the lack of a better word. The real question is whether we can do anything to make learning easier and more effective by changing our physiology.

Not surprisingly, the most studied, the safest, and the most effective “biohacks” are reminiscent of the advice some of us received from our grandmas.

Biohack #1: Sleep Well

This is hardly news to anyone. There are dozens if not hundreds of studies on the effects sleep deprivation has on cognitive performance and memory. And they all point in the same direction. Sleep takes time but it’s a secret weapon.

One study examined the effects of sleep deprivation on attention, working memory, executive function, speed of processing, and mood. You can probably guess what happened but these charts really bring the conclusions home:

You can learn more about the effects of sleep and deep deprivation on cognition and health in the podcast episode with Matthew Walker, a professor at UC Berkeley.

Biohack #2: Exercise

After sleep, there is one more secret weapon that is free, very effective, and thoroughly studied. Yes, exercise.

The study published in Nature is only one of many that concluded that exercise can improve cognition across multiple task types.

Even walking can be effective. Here is how walking affects creativity:

Exercise also increases brain-derived neurotrophic factor (BDNF) which is important for learning:

Finally, exercise has neuroprotective benefits and might even help prevent Alzheimer’s disease.

Different types of exercise have different effects so that it’s better to do a nice mix of resistance training (weight-training), high-intensity interval cardio training (HIIT), and low-intensity steady-state cardio training (LISS).

If you’d like to learn more about the effects of exercise on cognitive performance, check out the Wikipedia article on neurobiological effects of physical exercise.

Biohack #3: Eat Healthily

Just like sleep and exercise, nutrition might sound boring but it’s incredibly effective. Reviewing the principles of healthy nutrition is beyond the scope of this article but here are a few studies that might get you more excited about eating healthy.

Entire dietary approaches and individual foods have been studied in terms of their effects on cognition.

Foods that and nutrients that have been associated with better cognitive performance include green tea, fish, blueberries, nuts, olive oil, and many others.

Here are just a few study excerpts to spark your interest.

Blueberries have been shown to improve the reaction time and memory:

Omega-3 in fish oil — and in particular docosahexaenoic acid (DHA) — might improve cognition and delay age-related cognitive decline. And this randomized, controlled trial found beneficial effects of DHA on working memory scores:

To learn more about the link between nutrition and cognition, check out the Wikipedia article on nutrition and cognition, the summary of the MIND diet, Chris Kresser’s article, and Rhonda Patrick’s podcast.

Other Brain Biohacks

• Meditate. Mindfulness meditation in particular has been shown to improve cognition.
• Drink coffee and green tea — they contain caffeine and l-theanine that have nootropic effects.
• Ensure you’re breathing clean air by opening windows and/or using an air purifier. CO2 and Particulate Matter (PM2.5) have been shown to reduce cognitive performance.
• Consider researching nootropic supplements but err on the side of caution — the methods outlined here are much more powerful and safer.

What Doesn’t Work

This post wouldn’t be complete without the list of popular yet ineffective methods. So here is a non-exhaustive list:

• Multi-tasking. Focus on one thing at a time instead.
• Passive re-reading and highlighting. Practice active recall instead.
• Finding your unique “learning style” — the theory behind Visual, Auditory, or Kinesthetic learning isn’t backed up by strong evidence.

More Resources

• Online course: Learning How to Learn
• Book: Your Brain at Work
• Book: Why We Sleep
• Book: Range: Why Generalists Triumph in a Specialized World
• Book: Brain Rules
• Book: The Talent Code
• Book: Deep Work

If you like my articles, please consider signing up for radically infrequent email updates. And say hi on Twitter!

A Happy Job is a team-based and culture-centric recruiting platform. We help job-seekers — initially engineers — find teams that are actively hiring and fit their preferences. Search by product type, tech stack, team set-up, culture, values, working styles, or perks.

• See it in action!
• Say hi and support us on ProductHunt today
• Sign-up for email updates about new companies or to get notified when we launch the support for non-engineering jobs
• If you’re an engineering leader or recruiter, we invite you to showcase your team on the platform to attract and retain top engineers who care about your team’s values

If you’d like to learn more, keep reading.

What We Do

A Happy Job is a platform where engineers can find teams that are actively hiring and fit their preferences. You can search by product type, tech stack, team culture, remote-friendliness, or perks — no need to sign up.

The information about companies is a blend of answers provided directly by companies and thousands of public data points analyzed with machine learning.

If you’d like, you can answer a short questionnaire and get matched with the best engineering teams. Hiring managers and teams are matched with engineers based on intangible factors that really matter, such as culture, working styles, team set-up, etc. 

Why We’re Working on This

We believe that the future of recruiting is team-based and culture-centric.

Only 48% of employees are completely satisfied with their jobs. Around 50% of new startup hires are gone in two years. Companies in the SF Bay Area pay 20–30% of the new hire salary to recruiters and still struggle to attract, convert, and retain top talent.

At least part of the reason is that people apply to, interview with, and join the wrong companies and teams that were not a good fit for them to begin with.

Why Job-Seekers Should Care

Think about the last time you looked at the list of job openings, wondering if you should apply to any. 

If you ever considered working for startups, you likely stared at a long list of company logos and names that you likely had never heard of. But even when it came to companies with familiar names, you still had no idea what working there would be like… 

What’s the team like? Who is the hiring manager? Will it be a risky move that will stall or accelerate your career? Do they work crazy or normal hours? Are they all about competition or collaboration? Are there early morning calls with coworkers in different time zones?

Career moves are among the most impactful decisions we will make in our lives. Yet we often base them on very limited information. Several anonymous reviews you cannot validate. A couple of questions you had time to ask during those 30-min interviews. And the gut feeling. Oh, the gut feeling. 

Perhaps, if you’re diligent and well-connected, you speak with someone who works for the same company. And you make the jump, hoping for the best.

Or sometimes you don’t and miss a good opportunity because you don’t have enough information. And the hiring manager sees another rejected offer.

Why Hiring Managers Should Care

If you ever hired people, you might have experienced the opposite side of the same problem. 

First, you struggle to attract top talent. Then, after rounds of interviews, you’re still not 100% sure whether the person would fit in, become a stellar performer, and stay with the company in the long-term. 

Perhaps a few candidates you liked even rejected your offers, or even worse… Left or were asked to leave after working for the company for only a couple of months. According to Carta, 25% of startup employees are no longer with the company after 15 months and only around 50% stay over 2 years.

It doesn’t have to be this way. 

Imagine getting qualified and motivated candidates who self-select to apply because they share your team’s priorities, working styles, and culture. Get those offers accepted. And get people to stay for years.

Why Focus on Teams

We believe that the future of recruiting is team-based.

Job-seekers apply to and interview with companies but they work on teams. Most people spend most of their time interacting with their peers. 

Why Focus on Culture

Why is culture important? 

The primary reason people leave their jobs is because of their boss or their peers. Often, there is a misalignment on values, working styles or other “intangible” preferences.

Intuitively, we all understand that being aligned on these things can be a major driver of job satisfaction. Yet, hiring managers primarily focus on experience or skills and job seekers often focus on the title and compensation, overlooking other important things.

Culture is not about claiming a long list of generically sounding values. And it’s not about perks. Culture is about trade-offs and making hard choices. Very few companies can simultaneously be “adaptive” and “stable”. Or “meet aggressive deadlines and perform under pressure” while “having a real work-life balance and family-friendly culture”.

Yet job-seekers can only find generic values and perks on company pages when they do their research. That’s why we ask companies and teams to make hard choices and force them to prioritize so that we get an accurate picture. That’s why we also leverage multiple sources of data to validate the information.

What’s Next

Today we’re starting with engineers and limited geography. And we plan to expand to other roles, industries, and countries later. Design, product management, marketing are likely to be next.

As we grow, we’ll be adding more information about individual hiring teams. But today we mostly focus on company-level information that is hard to research.

Our Vision

We’ve built “A Happy Job” to help job-seekers find teams they would love. 

Job-seekers make better decisions when they get access to more information, such as — the team mission, set-up, culture, working styles, and other factors.

Even better if this information is available on the team level as opposed to the company level. Imagine being able to find the best opportunities for you in just a few clicks. 

If we could solve this problem and help people make conscious choices, the impact would have no limits. We could make millions of people happier and millions of organizations more productive.

Please check it out, let us know what you think, and spread the word!

Books

• Enlightenment Now: The Case for Reason, Science, Humanism, and Progress
• The Mind Illuminated
• Why Buddhism is True
• Becoming
• Culture Code
• Factfulness
• Refactoring UI
• The Brain: The Story of You
• The Story of Your Life and Others
• The Exhalation: Stories
• The Paper Menagerie and Other Stories

Music

• Mostly my favorite electronic and sometimes jazz music for focus on Spotify
• Top played: Tycho, Blackmill, Jon Hopkins, Odesza, Phaeleh, Chopin, Bach, Jimmy Smith, Aphex Twin, Grant Green, Bonobo, Brian Eno, Bill Evans, Youandewan
• New discoveries: East Forest, Tourist, Monolink, Jan Blomqvist, Jonas Saalbach, Laraaji, and Khruangbin

Documentaries and Movies

• The Mind Explained by Vox on Netflix
• General Magic
• It’s a Wonderful Life
• One Strange Rock
• Green Book

Podcasts

• This Week In Startups
• How I Built This
• Rhonda Patrick’s Foundmyfitness
• Selectively: Making Sense, Peter Attia’s Drive, Tim Ferriss, Startup Straight Talk by Atrium, YC, a16z, Below the Line

Or, if you want to go deeper, check out this book called Refactoring UI.

What this post is not: it’s not another “you should meditate” article. Even though I’ve seen quite a few valuable benefits, my practice is still semi-regular. And l debate how much time I’m willing to spend on meditation, given the opportunity cost. 

Instead, I’d like to share a few observations and invite you to share yours. If you have never meditated, some of these ideas might intrigue you and spark an interest in trying. And if you have experience with meditation, I’d love to compare notes!

Here are some realizations I had over time:

1. Our thoughts and emotions largely determine our destiny. And unlike other factors, they are not entirely outside of our control. Of course, they also define our subjective experience.

2. We probably shouldn’t believe all our thoughts. 

Our thoughts are quite random and are not necessarily representative of what we really think or want to think, especially the instinctive thoughts that come first. When you sit down to meditate, you realize how much stuff just pops up out of nowhere. 

Here is a thought experiment. Once you notice yourself making a strong judgment in your head about something, pause for a second and wonder if you actually fully agree with this judgment. It’s also interesting to wonder whether this judgment could have been different on another day—when you had more sleep, ate better food, had different levels of hormones and neurotransmitters or simply were in a more cheerful mood. 

3. We have some degree of freedom in choosing our internal experience—interpretations, opinions, and emotions—and external actions. 

It’s not like we can simply re-write the entire code of our operating system. But by repeatedly making different choices we can nudge ourselves into thinking and behaving differently.

We can, of course, go into a rabbit hall and argue about whether the thoughts that come after more deliberation are any different. But other things being equal, it’s safe to say that some reflection is better than no reflection and that a bit more involvement of frontal cortex and a bit less involvement of the limbic system is probably better—at least when it comes to making complex and high-impact decisions about the modern world. 

In theory, even without ever meditating one can: 

(a) understand these three ideas rationally by reading and thinking about them, 
(b) see them as true via having a first-hand experience,
(c) and even apply them in practice.

It’s just that when you sit in silence with nothing but your thoughts, you are more likely to get that first-hand understanding. And that first-hand experience is more likely to motivate you to experiment with putting these ideas into practice.

I do wonder whether the repeated formal practice is needed to strengthen or maintain this understanding and motivation—and how much is needed. An appealing alternative could be integrating moments of heightened awareness into regular life at appropriate—or all—moments. The goal would be to eventually have fewer automatic reactions or have automatic reactions that are more in line with the vision we have for the type of person we want to be. A possible intermediate and time-saving approach I’m experimenting with and having mixed results with is meditating during other activities, e.g. when going on long slow runs. Overall, I still feel like I only scratched the surface and hardly put anything in practice. 

Here are some books on this topic that I liked:

1. Mind Illuminated
2. Waking Up (plus, the app)

Amsterdam is beautiful. One word that was stuck in my head was “decorations”. Everything seems to be impeccably decorated. On the macro scale, the city itself is decorated with canals. On the medium scale, roads and buildings are decorated with trees and flowers. And on the micro-scale, there are thousands of tiny details—from exquisite architectural sculptures to art on building walls.

Amsterdam is lively. You see a lot of people walking on the streets—as the city was built for pedestrians and bicycles, not for cars. You also see a lot of people sitting outside, talking, and having fun. This creates the impression that people know how to enjoy life. And it also makes the city feel a little bit more like a community of people.

Amsterdam is cosmopolitan and open-minded. Everyone speaks perfect English. The city seemingly embraced globalization and adopted the best things from other countries. But it also definitely kept its unique character. And, of course, the Dutch government doesn’t seem to be in as much hurry as other governments to ban anything.

If you’re going yourself and looking for recommendations, here are some:

• We stayed at an AirBnB in De Pijp neighborhood and loved it.
• Coffee & Coconuts has great coffee and a very lively and cozy ambiance. 
• If you like fish, Bar Fisk would be a nice dinner place with slightly fancier food but reasonable prices.
• And, obviously: walk the canals, eat at random places you like, and enjoy your trip! 

Enlightenment Now by Steven Pinker

“What is progress? You might think that the question is so subjective and culturally relative as to be forever unanswerable. In fact, it’s one of the easier questions to answer. Most people agree that life is better than death. Health is better than sickness. Sustenance is better than hunger. Abundance is better than poverty. Peace is better than war. Safety is better than danger. Freedom is better than tyranny. Equal rights are better than bigotry and discrimination. Literacy is better than illiteracy. Knowledge is better than ignorance. Intelligence is better than dull-wittedness. Happiness is better than misery. Opportunities to enjoy family, friends, culture, and nature are better than drudgery and monotony. All these things can be measured. If they have increased over time, that is progress.”

21 Lessons for the 21st Century by Yuval Harari

“Homo sapiens is just not built for satisfaction. Human happiness depends less on objective condition and more on our own expectations. Expectations, however, tend to adapt to conditions, including to the condition of other people. When things improve, expectations balloon, and consequently even dramatic improvement in conditions might leave us as dissatisfied as before.”

Culture Code by Daniel Coyle

“Give a good idea to a mediocre team, and they’ll find a way to screw it up. Give a mediocre idea to a good team, and they’ll find a way to make it better. The goal needs to be to get the team right, get them moving in the right direction, and get them to see where they are making mistakes and where they are succeeding.”

Life 3.0 by Max Tegmark

“I’m encouraging mine to go into professions that machines are currently bad at, and therefore seem unlikely to get automated in the near future. Recent forecasts for when various jobs will get taken over by machines identify several useful questions to ask about a career before deciding to educate oneself for it. For example: Does it require interacting with people and using social intelligence? Does it involve creativity and coming up with clever solutions? Does it require working in an unpredictable environment?”

Doing Good Better by Will MacAskill

“Effective altruism is about asking “How can I make the biggest difference I can?” and using evidence and careful reasoning to try to find an answer. It takes a scientific approach to doing good. Just as science consists of the honest and impartial attempt to work out what’s true, and a commitment to believe the truth whatever that turns out to be. As the phrase suggests, effective altruism consists of the honest and impartial attempt to work out what’s best for the world, and a commitment to do what’s best, whatever that turns out to be.”

Guns, Germs, and Steel by Jared Diamond

“In short, Europe’s colonization of Africa had nothing to do with differences between European and African peoples themselves, as white racists assume. Rather, it was due to accidents of geography and biogeography—in particular, to the continents’ different areas, axes, and suites of wild plant and animal species. That is, the different historical trajectories of Africa and Europe stem ultimately from differences in real estate.”

The Blank Slate by Steven Pinker

“Equality is not the empirical claim that all groups of humans are interchangeable; it is the moral principle that individuals should not be judged or constrained by the average properties of their group.”

Bad Blood by John Carreyrou

“A sociopath is often described as someone with little or no conscience. I’ll leave it to the psychologists to decide whether Holmes fits the clinical profile, but there’s no question that her moral compass was badly askew. I’m fairly certain she didn’t initially set out to defraud investors and put patients in harm’s way when she dropped out of Stanford fifteen years ago. By all accounts, she had a vision that she genuinely believed in and threw herself into realizing. But in her all-consuming quest to be the second coming of Steve Jobs amid the gold rush of the “unicorn” boom, there came a point when she stopped listening to sound advice and began to cut corners. Her ambition was voracious and it brooked no interference. If there was collateral damage on her way to riches and fame, so be it.”

Lost and Founder by Rand Fishkin

“The problem with MVPs, and with the “something > nothing” model, is that if you launch to a large customer base or a broad community, you build brand association with that first version. […] Bing’s gotten much better since then, and is now as good as or better than Google on most queries, but that MVP hangover has stuck with the brand for years and, in my opinion, continues to dampen the prospects of what should be a very decent option for web searchers.”

Angel by Jason Calacanis

“If a startup called the Delta Corporation is making $10,000 a month selling enterprise software and they have five FTEs here in Silicon Valley, I simply calculate the FTEs by $120,000 each all in—or $10,000 a month—because they might have non-engineers getting $70,000 sitting next to developers making $150,000. […] So, Delta Corp is spending $50,000 on headcount and probably has $10,000 in miscellaneous expenses a month, for a total spend of $60,000 a month, which means they have a burn of $50,000. I’ve also asked at some point during the conversation how much they’ve raised. Let’s say they raised $1 million a year ago. […] They are burning $50,000 a month, so they have eight months left. I’ll say something to the founder like, “So, you’re burning about fifty thousand dollars a month and have six to eight months of runway left? Like four hundred thousand dollars in the bank?” They look at me with shock. How do I know that?”

How to Change Your Mind by Michael Pollan

“Habits are undeniably useful tools, relieving us of the need to run a complex mental operation every time we’re confronted with a new task or situation. Yet they also relieve us of the need to stay awake to the world: to attend, feel, think, and then act in a deliberate manner. (That is, from freedom rather than compulsion.) If you need to be reminded how completely mental habit blinds us to experience, just take a trip to an unfamiliar country. Suddenly you wake up! And the algorithms of everyday life all but start over, as if from scratch. This is why the various travel metaphors for the psychedelic experience are so apt. The efficiencies of the adult mind, useful as they are, blind us to the present moment.”

Stories of Your Life and Others by Ted Chiang

“Freedom isn’t an illusion; it’s perfectly real in the context of sequential consciousness. Within the context of simultaneous consciousness, freedom is not meaningful, but neither is coercion; it’s simply a different context, no more or less valid than the other. It’s like that famous optical illusion, the drawing of either an elegant young woman, face turned away from the viewer, or a wart-nosed crone, chin tucked down on her chest. There’s no “correct” interpretation; both are equally valid. But you can’t see both at the same time. Similarly, knowledge of the future was incompatible with free will. What made it possible for me to exercise freedom of choice also made it impossible for me to know the future. Conversely, now that I know the future, I would never act contrary to that future, including telling others what I know: those who know the future don’t talk about it.”

Never Split The Difference by Christopher Voss

“Negotiate in their world. Persuasion is not about how bright or smooth or forceful you are. It’s about the other party convincing themselves that the solution you want is their own idea. So don’t beat them with logic or brute force. Ask them questions that open paths to your goals. It’s not about you.”

The Spring Chicken by Bill Gifford

“If you could put the benefits of exercise in a pill, it would be an astonishing pill,” says Simon Melov, a researcher at the Buck Institute who has studied exercise extensively. ‘The data is now coming out on the effects of chronic exercise, and it is astonishing in terms of its ability to prevent all sorts of age-related disease, everything from cancer through to neurodegenerative disease to heart disease, even arthritis. All of these things have vastly lowered risk in people who exercise regularly—and if that was in a pill, it would be insane.’”

Chasing Excellence by Ben Bergeron

“What is grit, really? It’s a word that’s been used to describe everything under the sun, but it means something specific: when things get hard, you push harder; when you fail, you get back up stronger; when you don’t see results, you don’t get discouraged, but you just continue to pound away day, after day, after day, with relentlessness, consistency, heart, and passion — that’s grit.”

What have you enjoyed recently?

I hope you enjoy the read with your morning cup of coffee. Let me know what you think! ~Max

Technology and Startups

• Fake videos of real people — and how to spot them. (TED video)
• The Cloud 100. The top private companies in cloud computing 2018.
• Magic Leap One first look: worth the hype? Spoiler: not quite. (TechnologyReview)
• The Impossible Job: Inside Facebook’s Struggle to Moderate Two Billion People. The struggle is real. (VICE)
• Tesla, software and disruption. A deeper look at Tesla’s business strategy and future. Is Tesla going to completely disrupt the car industry? Or will the incumbents learn to build software faster than Tesla can learn to make cars at scale? (Ben Evans, a16z)
• Everybody Dance Now. “Do as I do” motion transfer facilitated by machine learning: given a source video of a person dancing, that performance can be transferred to a novel (amateur) target after only a few minutes of the target subject performing standard moves. (3-min video)
• WhatsApp Cofounder Brian Acton Gives The Inside Story On #DeleteFacebook And Why He Left $850 Million Behind. The story behind the conflict between WhatsApp founders and Facebook business strategy. Bonus: promising messaging apps that prioritize privacy, such as Signal and AnchorFree. (Forbes)

Growth and Marketing

• Slack Copywriting: What They Say to 9.6 Million Pageviews Every Month. Sometimes talking about specific product features can be more effective than talking about customer benefits.
• All Things Sales! 16 Mini-Lessons for Startup Founders. (a16z)

Product Management and Design

• The Lever. New Google blog about the way machine learning can be integrated into various tech products.

Leadership and Management

• The Two Traits of the Best Problem-Solving Teams. Build teams composed of people with diverse cognitive skills and create psychological safety. (HBR)

Economy and Global Development

• We’re Measuring the Economy All Wrong. Averages vs. the distribution. Stock prices vs. net worth. GDP vs. income. (NYT)

Books

• The Mom Test: How to talk to customers & learn if your business is a good idea when everyone is lying to you. An easy read that explains exactly how to conduct customer development interviews.

Podcasts and Documentaries

• Borders Series by Vox. A fascinating documentary about borders and their impact on countries, people, and the world.
• Elon Musk on Joe Rogan’s podcast. In case you’ve somehow missed this one.
• Scott Belsky and Drew Houston on Tim Ferriss’ podcast. Entrepreneurial stories and lessons learned from building Behance and Dropbox.

Cognition and Psychology

• How to read. You might want to take notes or even create flashcards and review them if you want to remember what you read.

Smarter Living

• Habits vs. Goals: A Look at the Benefits of a Systematic Approach to Life. “Exercise for one hour after work every Monday, Wednesday, and Friday” might be more effective than “lose 10 pounds by the end of the year”.

Science and Research

• The Complete Guide to the Science of Meditation. (Endpoints by Elysium Health)
• Workplace Wellness Programs Don’t Work Well. Why Some Studies Show Otherwise. Observational studies vs. randomized controlled trial. The former shows a strong effect, the latter shows almost none. Spoiler: healthy and motivated people are more likely to take advantage of these programs. (NYT)

Numbers

• $13T GDP could be delivered to the global economy by AI by 2030. (McKinsey)
• $4.15B is the valuation of Peloton, a connected training bike system. It’s impressive how the company managed to create a habit.

My Essays

• Careers in Tech Marketing For MBA Graduates
• Do You Need to Be Technical to Do Marketing in Tech?

Contemplation Break

“The way you train reflects the way you fight. People say I’m not going to train too hard, I’m going to do this in training, but when it’s time to fight I’m going to step up. There is no step up. You’re just going to do what you did every day.”
— Georges St. Pierre

Previous Issues

Issue 1, Issue 2, Issue 3, Issue 4, Issue 5.

Like blog posts like this one? Subscribe to my monthly email newsletter!

I hope you enjoy the read with your morning cup of coffee. Let me know what you think! ~Max

Technology and Startups

• The State of Artificial Intelligence in 2018: A Good Old Fashioned Report.
• Learning Dexterity (3-min video). Open AI trained a robot to manipulate physical objects. Plus, the NYT article on the same subject.
• How Old Are Successful Tech Entrepreneurs? A person who is 40 is 2.1x as likely to found a successful startup as a person who is 25. The average founder who had a successful exit through an IPO or an acquisition is 46.7 years old. (Kellogg)
• Amazon’s Facial Recognition Wrongly Identifies 28 Lawmakers, A.C.L.U. Says. ACLU used Amazon face recognition software, trained the model on a mugshot database and applied it to pictures of US congressmen. This led to some false positives — 28 members of Congress were identified as people who were arrested. The high false positive rate was mostly due to the low confidence threshold (80%) and the biased training set but the point that this technology can be misused is well taken. (NYT)

Growth and Marketing

• 19 Tactics to Solve the Chicken-or-Egg Problem and Grow Your Marketplace. A short read and practical advice.
• Intercom on growth. The new book by Intercom that provides a high-level overview of all key growth drivers — from acquisition to retention. (Intercom)
• On writing for deprecation. How to announce feature deprecation: be clear, be compassionate, communicate early and provide a path forward. (Slack)
• DAU/MAU is an important metric to measure engagement, but here’s where it fails. Andrew Chen on when to focus on daily/monthly active users and when not to.

Product Management and Design

• How to build a better product with UX writing. Understand your users, be empathetic and clear. See the visual examples in the article. (IndieHackers)

Leadership and Management

• What great listeners actually do. Create a safe environment, reduce distractions, understand the substance, pay attention to emotions and nonverbal cues, and ask clarifying questions. (HBR)

Economy and Global Development

• Memorizing these three statistics will help you understand the world. Child deaths 20M a year –> 6M since 1960. The fertility rate ↓ 50% since 1960. 137K people escaped extreme poverty every day since 1990. (Max Roser, gatesnotes.com)

Books

• Doing Good Better. I think every high-schooler and everyone interested in making a positive impact on the world should read the book. It lays out a framework for evaluating the social impact of non-profits and individuals in a very logical yet uncommon way. Also see: effectivealtruism.org

Podcasts and Documentaries

• Vitalik Buterin on Cryptoeconomics and Markets in Everything. Vitalik and Tyler Cowen talk about cryptocurrencies, economics, decentralization, and other topics.
• How Social Media and AI Hijack Your Brain. Tristan Harris, the founder of the Center for Humane Technology, talks about building ethical tech products and the impact of social media on the world and individuals.

Cognition and Psychology

• “If You Say Something Is “Likely,” How Likely Do People Think It Is?”. This reminds me of the Strategy class at  Berkeley-Haas where we talked about assigning specific numeric probabilities to various business outcomes instead of using vague words like “likely”. (HBR)

Smarter Living

• The One Routine Common to Billionaires, Icons and World-Class Performers. Spoiler: it’s meditation. Perhaps consider picking up this habit if you haven’t yet. (Tim Ferriss)

Science and Research

• The Lifespan of a Lie. It turns out that one of the most famous psychology experiments in the history — the Stanford prison experiment — had serious methodological flaws. Plus, Philip Zimbardo’s defense.

Numbers

• 10B rides were provided cumulatively by Uber.
• 8M miles is the total distance that was covered by Waymo cars.

My Blog Posts

• Best Tools For Competitive Intelligence. Wouldn’t it be nice if you could always be on top of what your competition is up to without making this your full-time job?

Contemplation Break

“Men who believe that they are accomplishing something by speaking speak in a different way from men who believe that speaking is a waste of time.”
— Neal Stephenson, Cryptonomicon

Previous Issues

Issue 1, Issue 2, Issue 3, Issue 4.

Like blog posts like this one? Subscribe to my monthly email newsletter!

There are many online tools available in 2018 that can help you achieve exactly that.

For the purposes of this post, I will omit all the context about competitive intelligence — what it is, why it’s important, how to do it systematically, and how to make it actionable. Instead, I will simply list useful tools and apps and provide brief descriptions.

Let’s dive right in!

Listen to Their Customers

• B2B software companies: Capterra and G2Crowd reviews
• E-commerce companies: Amazon reviews and Google Shopping reviews
• Mobile apps: reviews on iTunes and on Google Play marketplaces
• Reviews on social media platforms: read comments on ProductHunt, search Twitter for the responses to the company, look at comments on their Facebook page, and search for “company name Reddit”
• Other categories: try searching directly for “company name reviews” or “product category reviews”

Listen to Their Employees

• Do people enjoy working there or are they leaving in droves — Glassdoor.com
• See how their headcount has been growing over time on LinkedIn

Review Their Funding and Investors

• Crunchbase.com
• For startups: AngelList

See Whom They Are Hiring And Where

• LinkedIn Jobs
• Glassdoor Jobs
• For startups: AngelList Jobs

Automatically Receive News About Your Competitors

• Google Alerts

Stay on Top of Company Announcements

• Subscribe to their email newsletter(s)
• Subscribe to their blogs and press releases

Detailed Analysis

• Watch, read, or listen to interviews with their CEOs or key executives.
• If your competitors are public, review their S-10, 10-Q or 10-K reports on sec.gov. Insights to expect: financial metrics, market intelligence, future plans, etc.
• “Mystery shopping”. Sign up for product trials or product demos to learn more about their products. Visit their stores and talk to their salespeople.
• Do your own qualitative research. Not only talk to your customers but to customers of competing products too — so that you could learn about their strengths and weaknesses.

Other Tools

• Review their traffic — Similarweb
• Review company profiles — Owler
• Understand their content strategy — BuzzSumo
• Review their SEO rankings and PPC strategy — SpyFu
• Get notified when their website is updated — Visualping

Consider Purchasing Research Reports

Browse reports published by these companies:

• McKinsey & Company
• Boston Consulting Group
• Bain & Company
• Gartner, Inc., especially their Magic Quadrant reports
• Forrester, especially their Wave reports
• IDC

Obvious But Sometimes Overlooked

• Corporate website and blog.
• Social media: LinkedIn, Twitter, Facebook, etc.

Leverage Your Colleagues

• Create a slack channel for competitive intelligence. Post big news and encourage all customer-facing teams to do the same, so nothing important goes unnoticed.

Let me know if I left anything out!

Hope, you enjoy the read with your morning cup of coffee. Let me know what you think! ~Max

Technology and Startups

• Deep Learning: Application Landscape – March 2018. An informative and very visual overview. (106-slide deck)
• F8 – Keynote Day 2. Facebook talked about ML applications to video analysis, social VR, AI ethics, etc. (FB Video)
• Top AI research researcher can be paid $1M. Some argue that the majority will be replaced by their own code. (NYT)
• Iran and Russia have been trying to ban Telegram. Telegram’s end-to-end encryption made several governments uneasy. (NYT)
• Attack of the micro-brands. Dropshipping, Instagram advertising, and other trends that open the way for niche consumer brands. (Scott Belsky)
• How merchants use Facebook to flood Amazon with fake reviews. Amazon has a fake review problem (duh!) and is unwilling to do much to fix it. Fakespot and ReviewMeta to the rescue! (WaPo)
• Mary Meeker’s 2018 internet trends report. Mobile growth stalled; people are spending more time online; voice interfaces are taking off. 11 of the world’s 20 biggest Internet companies are in the US. The rest are in China. (Recode)
• Google I/O 2018 Keynote in 14-min. The announcement of Google Duplex and Smart Compose probably generated most fears excitement. Duplex allows Google Assistant to make calls on your behalf. Smart Compose auto-suggests email responses – paving the way for Yuval Harari’s dystopian vision.
• Labor 2030: The Collision of Demographics, Automation, and Inequality. A very detailed report on the impact that automation might have on labor markets. 40M jobs might be eliminated in the US alone. The report offers no insight into the degree to which this loss will be (over)compensated by the creation of new job categories though. (Bain & Company)

Growth and Marketing

• The top marketing statistics for 2018 and why they matter. Trends in content marketing, video, email, paid ads, messengers, and social media. (Intercom)
• How startups die from their addiction to paid marketing. Instead of relying on paid marketing to grow, try to create a product-based viral loop. If you cannot, carefully measure LTV/CAC by channel. (Andrew Chen)

Product Management and Design

• 7 Practical Tips for Cheating at Design. Tiny design changes that can make products look sleeker.

Leadership and Management

• The Power of Listening in Helping People Change. Listening might help lower social anxiety and increase motivation. (HBR)

Books

• Crucial Conversations: Tools for Talking When Stakes Are High. Key Practices: make everyone comfortable and find/invent a common goal.

Podcasts and Documentaries

• Mark Zuckerberg talked to Ezra Klein about Facebook vision, fake news, electoral interference, privacy violations, etc. (Vox)
• JUMP Bikes founder Ryan Rzepecki talked to Jason Calacanis about the ways in which multimodal transit solutions can transform cities, the future of commuting versus denser cities, etc. (This Week in Startups)
• Airships and the Future that Never Was – an educational episode on the history of airships and the hype cycle they went through. (99% Invisible)

Cognition and Psychology

• Mental Models: The Best Way to Make Intelligent Decisions (113 Models Explained). (Farnam Street Blog)

Products

• Bubble.is is trying to make visual programming work. If successful, it could enable a whole new wave of makers to create apps without learning to code.

Smarter Living

• Safest seats on a plane are middle seats in the rear section.
• Cook at home instead of eating out to minimize your exposure to phthalates from plastic. (Guardian)

Science and Research

• The Non-Scientist’s Guide to Genetic Testing. (Elysium Health)

Photography and Art

• Earth’s Wonders Like You’ve Never Seen Them Before. Beautiful satellite pictures, taken at an angle.

Numbers

• 70M MAU on bike sharing platforms in China.
• 35% is the claimed Lyft’s market share of the ride-hailing market.
• 121 is the number of emails the average worker receives each day.
• 40M jobs might be eliminated by automation by 2030 in the US alone.
• 100M customers subscribed to Amazon Prime and bought 5B items in 2017.

My Essays

• How to Find Your Ideal Customers

Contemplation Break

“Money is like gasoline during a road trip. You don’t want to run out of gas on your trip, but you’re not doing a tour of gas stations.”
– Tim O’Reilly

Previous Issues

Issue 1, Issue 2, Issue 3.

Like blog posts like this one? Subscribe to my monthly email newsletter!

Any Customers Vs. Ideal Customers

A new startup might try anything to get its first customers. At this exploration stage, you cast the net wide. You approach different individuals and companies and do whatever it takes. Any customer is a great customer.

But as the company grows, it learns that not all customers are created equal. Some are easier to acquire. Others benefit more from the product and are easier to retain. Yet others are easier to expand and upsell to.

This raises a question. How do you know who is your primary target audience?

At an early stage, you “just know” your ideal customer. Your ideal customer profile is whoever was willing to buy your first product or sign an agreement before you even developed a prototype.

Of course, you can always stop here.

But if you’re planning to become a world-class company, you might need to take it to another level and make sure that your marketing is also world-class. All top technology–, and more narrowly, Software as a Service (SaaS) companies have a very clear understanding of their ideal customer profile. Even though this post will be particularly relevant to this type of company, the general principles can be applied to any industry.

A side-note on terminology: terms such as “ideal customer profile” and “target audience” are often used interchangeably. “Buyer Persona” is another similar term that emphasizes the focus on individuals – as opposed to companies.

Why Define Ideal Customer Profile (ICP)

An ideal customer is a customer that is more likely to like your product and pay for it. Knowing your ICP should help you grow revenue while investing less. This higher return on investment is achieved through better targeting among other things.

Imagine, you sell software that helps cities analyze traffic patterns and make better urban planning decisions. You could place a TV commercial and hope that the right people will see it. This would be utterly inefficient.

Alternatively, you could identify a subset of cities that are most likely to adopt this type of software, then identify the right officials who are most likely to be decision-makers and target them with a proposal that addresses their specific pains and explains how your product can alleviate them. Needless to say, this would be infinitely more effective.

And, by the way, customers will also appreciate this targeted approach. See “Principles of Ethical Marketing” for more on this.

If you need more reasons:

How to Develop ICP

The approach will depend on a number of things:

• Company size and stage
• Available resources
• Available data

You can find an example below that you can use a starting place and tailor it to the needs of your company if needed.

What Insights Do You Need To Get

Here are some specific questions you’d need to answer:

1. What companies should we target and why? – assuming a B2B product
2. What group of individuals should we target and why?
3. How to target them most effectively?

Here is another graphic that illustrates the types of insights you might want to obtain: Let’s get more specific. What do you need to know about companies and individuals to decide whether you want to target them and to do it effectively?

Let’s start with the end in mind.

Final Result: ICP / Buyer Persona Template

The final result should be very easy to understand. Creating a visually-appealing design for communication purposes is a good idea!

Here is one example, focused on company data and financial metrics:

Here is another example, focused on the buyer’s personas:

If you’d like, you can request a free copy of this ICP template in Google Slides or pptx format.

So, what you need to decide is what companies and personas to target and how to do it? Let’s start with companies.

Target Companies

First, you might want to decide what metrics you care most about.

Examples:

• LTV (customer lifetime value)
• CAC (customer acquisition cost)
• MRR (monthly recurring revenue)
• NPS (net promoter score)
• Retention rate

Second, you can identify attributes that can help you in the future.

Examples:

• Industry or vertical
• Company size
• Geography

Third, you might want to correlate these attributes with metrics of interest to find your sweet spot.

Example:

• Priority 1. Tech companies in the US, size 50-200. Average LTV: $8456.
• Priority 2. Manufacturing and construction companies in the US, size 100-500. Average LTV: $7863.
• Priority 3. Marketing agencies in the US and EMEA, size 20-100. Average LTV: $3753.

A side-note: to learn more about key business metrics commonly used at technology companies, read this blog post.

Now, let’s talk about buyer personas.

Buyer Personas

First, understand who is most likely to benefit from your product.

Additional questions:

• Who is most likely to use it?
• Who is most likely to champion it within an organization?
• Who is most likely to pay for it?

Second,  figure out the key distinguishing characteristics of these personas.

Often these characteristics are titles. But sometimes these characteristics can be demographic, behavioral, or technographic.

Here are examples of each of the above:

• Marketing managers who are responsible for sales enablement
• Asian men who are 40-50 years old, have one or two kids and travel to Asia at least once a year
• Tech-savvy individuals who currently use Trello, Google Spreadsheets, and Dropbox

Third, you need to get into these people’s heads.

Examples of things you might want to understand:

• Above-mentioned demographics, psychographics, and technographics.
• Professional experience and background.
• Day-to-day work: goals and KPIs, challenges and pains, cross-functional collaboration, existing workflows, and use cases, etc.
• Everything about the problem you’re trying to solve: how prevalent it is, current solutions, what works about current solutions, what doesn’t work, what they wished worked better, and “whys” behind all those answers.
• How vendor choices are made: what triggers a decision, who makes it, what’s taken into account, as well as price sensitivity and the vetting process.
• Deep-dive into solutions: what competitors customers know about, what business benefits they expect, what features they care about, what other considerations they have in mind; “whys” behind all these.
• Media preferences: video vs. text; blogs vs. email; LinkedIn groups vs Facebook groups, etc.

Types of Research and Data

These are the main types of research that are most commonly used to obtain insights about the buyer personas:

1. Interviews
   1. Interview your internal customer-facing teams to leverage the existing internal knowledge
   2. Interview your existing customers
   3. Interview you free trialists and/or customers who churned
   4. Interview prospects (individuals in your target audience and who have NOT yet used or even learned about your product) *
2. Surveys
   1. Survey your customer-facing teams to leverage internal knowledge
   2. Survey your existing customer base
   3. Survey you free trialists and/or customers who churned
3. Analysis of customer reviews
   1. B2B reviews: G2Crowd and Capterra
   2. B2C: Amazon, Google Play, iTunes store or other platforms used for distribution
4. Financial analysis – collaborate with Finance, BizOps and/or Data Science team
5. Product usage analysis – collaborate with the Product team
6. Secondary: competitive intelligence

* This is arguably the most important source of information – watch this lecture by the Twitch.tv founder at the YC Startup School at Stanford to learn why.

If the list seems overwhelming or you’re resource-constrained start with interviews. Focus primarily on the existing customers and on your presumed target audience that never heard about your product. The reason is that interviews come first is that they are great for digging deep into specific topics and uncovering non-obvious insights. Examples:

• If a customer says “we just want something that integrates with our CRM”, you can ask “why?” and hear “because we need to get started quickly”.
• Then, you ask “why?” again and hear “because there is a high turnover on the team and we have no bandwidth to learn new tools”.
• Now, you know the pain.

After you are done with interviews, you can use quantitative analysis to validate hypotheses generated during interviews. For example, you can look into your product usage data to see how many customers actually enabled a CRM integration. Or you can survey the broad customer base to see if the problems uncovered during the interviews are universal and shared by many customers.

Now, you have hard data. A side-note: you might choose to partner with a research agency if resources permit and/or if you lack in-house expertise. This partnership can be particularly useful when it comes to interviews. There are agencies that specialize in qualitative research and have professional interviewers who can help you get the insights you need.

What’s Next

Ideal customer profiles are useless unless they are effectively communicated to the whole company and all key stakeholders are aligned. So, here is what to do next:

1. Make sure that you get inputs and feedback from all key stakeholders. Usually, these include the Executive team, other Marketing teams, Sales, Customer Success, Product, and Finance.
2. Finalize the company profiles and buyer personas, decide what insights are most important and need to be used and communicated.
3. Update value proposition and write positioning statements, tailored to each persona, to appeal to their specific pains and purchase drivers.
4. Work with the design team to create visually-appealing profiles that can be printed out, hang on the walls, and included in the onboarding materials.
5. Communicate to key teams or to the entire company.
6. Implement all the learnings. Marketing: better targeting and content. Product: better product decisions. Sales: better demos.
7. Continue updating your understanding of ICP as the company, products, and the market all evolve.

Let me know if you have any questions!

Related Posts

1. How to Develop and Validate a Value Proposition
2. Key Growth Metrics, Part 1: Churn/Retention, and Virality
3. Key Growth Metrics, Part 2: CLV/LTV, and CAC

Now, get a cup of coffee and enjoy!

Technology and Startups

• Latest YC Batch: 64 Startups. (Techcrunch)
• Crypto Canon. A list of crypto learning resources. (A16z)
• A visualization of how self-driving cars see the world. (Waymo on Medium)
• Live video of the Falcon Heavy launch. This was just epic and worth a re-watch. (SpaceX)
• B2B Buying Guide – AI, HR, legal, finance, marketing, productivity, and other products. (Y Combinator)
• Cora. The electric aircraft company that Larry Page invested in unveiled its first “self-piloting” product.
• The Building Blocks of Interpretability – visualization of image recognition machine learning algorithms. (Distill)
• Here Come the Fake Videos, Too. It gets increasingly easy to develop fake videos using machine learning. (NYT)
• YouTube, the Great Radicalizer. YouTube’s recommendations might lead you to increasingly radical content – no matter the topic. (NYT, Zeynep Tufekci)
• 10 Breakthrough Technologies of 2018. Sidewalk Labs’ “sensing city” and “zero-knowledge” proofs facilitated by Zcash cryptocurrency are interesting. (TechnologyReview)

Growth and Marketing

• What 10 Studies Say About The Best Time To Send Email. Best options: Tue at 10 am, Thu at 8 pm, Wed at 2 pm, Tue at 6 am, and Thu at 10 am.
• Analysis of the Dropbox financial model based on the S-1 filing. 1.1B revenue; 35% YoY; 7.97 multiple; 67% gross margin; -0.1% net income margin; 2% Free–>Paid conversion; 93% of customers are self-serve. (Tom Tunguz, VC)
• Lessons from Spotify. Atypically for tech, Spotify’s marginal cost is not zero as it pays royalties. Since it’s hard to negotiate better rates with the music industry, Spotify will either need to tightly control operational costs or cut out the labels altogether – not a likely option. (Stratechery)

Product Management and Design

• Making Good Decisions as a Product Manager. Come up with a framework to evaluate decisions. Possible criteria: investment of resources, the impact of a positive outcome, the impact of the negative outcome, and overall importance. (Shopify Product Director)

Leadership and Management

• To show recognition, try speaking a different language. How to give positive feedback: encouraging words, tangible rewards, and focused time. (VP of Design at Lyft)
• Ahead of the curve: The future of performance management. Real-time feedback, decoupling compensation from performance evaluations, and future-focused coaching. Overall, more question marks than answers. (HBR)

Books

• The Moral Epistles: 124 Letters to Lucilius. I’ve been re-reading this classic and taking more time to think about stoic ideas. To me, the main recurring themes are non-attachment, perspective, and contentment.
• Rise: 3 Practical Steps for Advancing Your Career, Standing Out as a Leader, and Liking Your Life. Some practical and honest advice on how to become more effective as a professional and as a leader.

Podcasts and Documentaries

• Jimmy Wales on Wikipedia. (How I Built This by NPR)
• Robin Hanson on Signaling and Self-Deception. (Conversations with Tyler)
• Mental Models for Understanding Crypto Tokens. (A16z)

Cognition and Psychology

• How to Remember What You Read. Read books that are relevant to your situation, have a specific reason for reading them, take notes, build a vivid mental picture, find connections with familiar topics, and teach what you learned. (Farnam Street)

Products

• Nuzzel sends you a digest of links most frequently shared by people you follow. I’m using it as a social filter for potentially interesting content.

Numbers

• $1.9B is 2017 total robot sales.
• $1B is processed via mobile each day.
• 93% of professional developers are male. (Stack Overflow survey)
• $8.55–$10/hr is the median profit from driving for Uber/Lyft. (MIT, initially estimated at $3.4/hr)
• 1B hours are spent on Facebook each day — ½ hour per active user. (Facebook investor call)
• $8B is how much Netflix is planning to spend on original content in 2018.
• 55% of adults in the US seem concerned about government not doing enough to regulate big tech. (Online survey by Axios)

Smarter Living

• World Happiness Report 2018. The happiest countries: Finland, Norway, Denmark, Iceland, Switzerland, Netherlands, Canada, New Zealand, Sweden, and Australia.
• Expat Insider 2017. The ranking of the most expat-friendly countries. Top 10: Bahrain, Costa Rica, Mexico, Taiwan, Portugal, New Zealand, Malta, Colombia, Singapore, and Spain. (Internations)

Science and Research

• A summary of the 800-page report on longevity research covers the state of affairs and the industry landscape. (SingularityHub)
• Studying Studies: Part V – power and significance. Determinants of statistical significance and the difference between statistical significance and practical/clinical significance. (Peter Attia)

Music

• These Spotify playlists. Especially #1 and #2. Also, #4.

My Essays

• Principles of Ethical Marketing.

Contemplation Break

“No good thing renders its possessor happy unless his mind is reconciled to the possibility of loss.”
– Seneca, Moral Letters to Lucilius

Previous Issues

Issue 1, Issue 2.

Like blog posts like this one? Subscribe to my monthly email newsletter!

Now, get a cup of coffee and enjoy!

Technology and Startups

• The best products launched in 2017 by ProductHunt.
• The list of 26 companies started by 1st or 2nd generation immigrants.
• CES 2018 summary by Steven Sinofsky. Big trends: voice interfaces, home electronics, robots, cars, and AR/VR headsets.
• AI and Deep Learning in 2017 – A Year in Review. A great summary of progress and links to online courses by UC Berkeley, Stanford, and others.
• Amazon opened a store with no checkout lines (NYT). Customers are charged automatically, thanks to machine learning and hundreds of cameras.
• Blockchain Governance: Programming Our Future by Fred Ehrsam. Setting up the right mechanisms for coordination and incentive structures can be critical for new blockchains.
• Beyond the Bitcoin Bubble (NYT). All the hype and crypto speculation aside, blockchains hold a promise of a truly open and decentralized Internet – the way it was meant to be.
• How an A.I. ‘Cat-and-Mouse Game’ Generates Believable Fake Photos (NYT). Generative adversarial networks at work. Getting AI to classify content is important but getting AI to create content is even more exciting.
• Twitter’s investigation of Russian interference in the US election. Twitter removed Russian government-linked accounts and emailed 678K people in the US who followed, retweeted or liked a tweet from these accounts during the election period. I’m glad social media companies start to accept more responsibility for how their platforms are used and abused.
• Facebook will show more content from friends and family and less content from publishers. Overall, it’s good that Tristan Harris’ “Time Well Spent” ideas are getting traction. I wonder what would happen if FB let users adjust defaults at any time by choosing % of content from friends vs. % of content from publishers; long-form vs. short-form; education vs. entertainment; text vs. video; new vs. old; from currently followed sources vs. discovery; as well as mute specific topics/keywords.

Growth and Marketing

• Intercom on Marketing (ebook). A good intro to marketing and, in particular, product marketing that will be interesting to those who are relatively inexperienced.
• How to Design Marketing Campaigns. Basics of marketing segmentation, messaging hierarchy, and campaign management – this article will be useful to those who’re new to marketing or looking for a refresher.
• HubSpot’s Pricing Page Redesign → MQL Conversions 165%↑ & Free Sign-Ups 89%↑. How: research first – usability testing, internal feedback, and customer intelligence; then design based on insights and A/B test.
• “What’s next in growth?” (video) talk by Andrew Chen who leads the rider growth at Uber. Andrew recommends you ignore “growth hacks” and focus on fundamentals that worked for decades. E.g. user referrals, shareable content, and using discounts to jumpstart demand for new products.

Product Management and Design

• Facebook tried flagging fake news. Looks like the main challenge here is human psychology and confirmation bias: people who want something to be true aren’t usually very open to counterevidence.

Leadership and Management

• The surprising thing Google learned about its employees (WaPo). “Among the eight most important qualities of Google’s top employees, STEM expertise comes in dead last. The seven top characteristics of success at Google are all soft skills.”
• Science-based interviewing by Don Moore, a professor at BerkeleyHaas (video). Be humble about your abilities and use structured interviews. Surprisingly, intelligence– and integrity tests might be relatively predictive too. Cultural fit, if not defined, leads to “hire people who are like me” bias.

Books

• The Gene: An Intimate History by Siddhartha Mukherjee tells the story of genetics from early days of Aristotle all the way to CRISPR/Cas9.

Podcasts and Documentaries

• This Week in Startups by Jason Calacanis. Try Coinlist or Finless Foods episodes.
• Sam Harris’ Podcast. Try recent The Science of Meditation episode.
• AlphaGo documentary on Netflix. The storytelling is great but it’s short on technical details.
• The Daily by The New York Times. Whether consuming news daily is a good use of one’s time is still an open question to me. But there are better and worse ways of doing it. The Daily appears to be a better one. Try “The Best of 2017” episodes first.

Cognition and Psychology

• Cognitive Bias Cheat Sheet. A very comprehensive list of 175 bugs that our brains have.
• Ten Commandments for Aspiring Superforecasters. These takeaways from Philip Tetlock’s book “Superforecasting: The Art and Science of Prediction” can serve as guidelines for predicting future events.

Products

• TaoTronics Earbuds – my favorite $26 Bluetooth earphones that are perfect for listening to podcasts and audiobooks when running and commuting.

Numbers

• $16K is the price of a laundry-folding robot.
• 8.3M subscribers were added by Netflix in Q4.
• $380B is the iOS economy revenue in 2017, including hardware.
• 39M Americans own a smart speaker, 6M of them sold by Google in 2017.
• 1.4K electronics products were ordered per second on Amazon via mobile this holiday season.
• Polls show no decline in American people’s impressions of tech companies despite all the talk about the “tech backlash”.

Smarter Living

• Write a to-do list for the next day to fall asleep faster.
• How to be “good at self-control”: create habits, learn to enjoy the activities, and eliminate temptations.
• Strava published a global heatmap of all fitness activities. Aside from being aesthetically pleasing, it can be useful for finding places to run, bike or swim.

Science and Research

• Longevity FAQ by Laura Deming. Scientific ideas that have been linked to longevity in mice or human: caloric restriction, insulin/IGF, parabiosis, senescence, autophagy, sirtuins, etc.
• Studying Studies: Part I by Peter Attia. The first article in series that explain how to read scientific studies. This one explains the difference between relative risk reduction often reported by the media and absolute risk reduction that sometimes might be more relevant.

My Essays

• A Long Read of My Dream
• Principles of Ethical Marketing
• How to Develop and Validate a Value Proposition

Contemplation Break

“Results are obtained by exploiting opportunities, not by solving problems.” – Peter F. Drucker

Previous Issues

Issue 1.

Like blog posts like this one? Subscribe to my monthly email newsletter!

• What does your company do?
• What product does your company sell, and how is it different from the competition?
• Who is this product for?

Seriously, stop reading right now and try to answer these questions aloud or in your head.

Can you answer these questions in 1-3 sentences and in a clear way that wouldn’t invite follow-up questions from an average customer?

Can everyone in your organization?

What is a Value Proposition?

A value proposition is simply a way to connect a product offering with customer needs in a clear way.

It should resonate with customers and would motivate them to try or buy your product. Essentially, the value proposition is synonymous with the unique selling proposition and product positioning.

In order to find its product-market fit, startups need to come up with value proposition statements for its products. Initially, they begin with a hypothesis or a simple guess.

This hypothesis then needs to be validated with customers and refined as the product evolves and as the company acquires more knowledge about the market and customers.

A validated value proposition statement is the ultimate outcome of finding the product-market fit. This statement should encapsulate the key insights you’ve gathered about customers, as well as unique product strengths.

Do You Need One?

If you don’t have one, you’re shooting in the dark. If you have one but it’s not validated, you might be deluding yourself.

Here are some practical benefits of developing a value proposition:

1. A value proposition forces the company to focus. It’s impossible to write a value proposition without having difficult conversations and answering tough questions about priorities. For example, what does the company stand for and what product offerings will it commit to?
2. A value proposition informs all marketing campaigns. Once the target audience and benefits of a product are identified, the team can be much more effective at creating impactful content and making sure that this content is aligned with the main message.
3. A value proposition unites and aligns everyone in the company by clearly describing what the company is offering to the world. Product teams can use this information to make better product decisions and sales teams can use it to present the product in a way that resonates with customers.

So, how do you develop one?

Data Used to Develop a Value Proposition

1. Customer needs. The product or marketing team must understand customer needs, problems, and desires by doing market research, arranging usability studies, listening to sales calls, and getting feedback from paid customers.
2. Product and company strengths. A value proposition should reflect the company’s core strengths and priorities, as well as the product’s existing and future functionality.
3. Competition. The product or marketing team should conduct a competitive analysis to understand how other companies are positioning themselves to decide what segments to target and how to differentiate their product.

Now, let’s look at some templates and examples.

A Simple Value Proposition Template

Here is a simple and useful template that you can use to put a value proposition in writing:

To <our customers/prospects>, <name of the brand> is the brand of <frame of reference> that promises <relevant, superior benefits> based on <compelling rational or emotional proof points>.

Examples of Value Proposition Statements

Here is an example of a value proposition or positioning:

To savvy business and leisure air travelers (average age: 44), Virgin America is the brand of contemporary air travel that is making flying fun again by disrupting the domestic airline industry with distinctly designed aircraft, technologically advanced amenities, and world-class service.

Notice how this format forces company to focus. To include “technologically advanced amenities, and world-class service” in its value proposition, Virgin America also needs to make a decision not to compete on price.

By defining its target audience as savvy business and leisure travelers and by choosing to focus on the domestic market, Virgin America also decides not to cater to long-distance, cross-border travelers.

Here is how a hypothetical startup might use the same formula:

X is design software that teams of user interface designers use to streamline collaboration and reduce busy work. See everyone’s edits and comments in one place in real time. X won a “Breakthrough of the year” award by The New Tech Times and is used by Tesla and Uber. Our clients report completing 30% more designs and reducing the number of creative reviews.

Let’s break it up:

• Customers/prospects: teams of user interface designers.
• Name of the brand/product: in this case, “X”.
• Product category or frame of reference: design software. Notice the deliberate choice of words: we are not describing the product as “a design app” or “a design platform“.
• Relevant, superior benefits: streamline collaboration and reduce busywork; see everyone’s edits and comments in one place in real-time.
• Compelling emotional and rational proof points: “see everyone’s edits and comments in one place in real-time” — this is both a benefit and a proof point that explains how customers are going to get benefits streamlined collaboration and reduced busy work.
• Compelling rational proof points and social proof: won a “Breakthrough of the year” award; used by well-known companies.

Validate a Value Proposition with Market Research

How do you know that you have actually found your value proposition and your product-market fit?

Once a value proposition is outlined, it needs to be validated. This section might be its own blog post and even its own book, but I will try to be brief.

Here is a non-exhaustive list of methods a startup or a company can use to validate a value proposition:

• Product analytics. Do customers use the features you deem most valuable?
• Usability studies. Is there a disconnect between what customers expect and get out of a website and a product?
• Customer interviews. An in-depth conversation about customer needs and about their reaction to the value proposition.
• Search volume analysis. Are you using the same words as your customers, are they searching for the terms you use?
• Competitive intelligence. How do competitors describe their products, are you clearly differentiating your product?
• Internal employee surveys and interviews. What insights do your customer-facing teams – customer support, sales, and research – have about customer needs and what ways of positioning the product worked best for them?
• Tests via paid advertising and outbound campaigns. What versions of the value proposition yield the best click-through ratios and response rates?
• Website heat-maps and A/B test results. What versions of the value proposition yield the best conversion rates and engagement?

Value Proposition Hierarchy

Most companies will also find it useful to have multiple versions of a value proposition, tailored to specific customer profiles.

Here is a simple hierarchy:

1. The universal value proposition should appeal to the broadest possible customer base. This version should be used in marketing and sales channels that don’t allow customization.
2. Custom value propositions should address the needs of specific audiences and customer segments. For example, they might highlight benefits relevant to specific roles within the company, relevant to specific verticals or specific company sizes.

Example of how different versions can be used:

• The main heading on the landing page can be based on the high-level proposition as our goal is to elicit interest from a wide range of website visitors.
• Other sections of the website can address the needs of specific audiences and should be based on custom value proposition tailored to specific audiences.

Another example:

• Sales managers can lead with a high-level value proposition first — before they get to know a lead’s specific needs and use cases.
• Later in the conversation, once they get more information about the lead, they can customize the way they sell the product and use a custom value proposition.

Related Posts

1. How to Find Your Ideal Customers
2. Key Growth Metrics, Part 1: Churn/Retention, and Virality
3. Key Growth Metrics, Part 2: CLV/LTV, and CAC

If you liked this article, read more about marketing and growth and subscribe to my email newsletter!

You probably heard people use marketing as a derogatory term.

Admit it, you probably did too.

Why?

You can probably think of many examples when marketing wasn’t practiced ethically. It’s true that one doesn’t need to go far to find them. Spam emails that lack “unsubscribe” buttons, misleading packaging, annoying pop-up, or video ads that you cannot skip.

Marketing can go wrong for all the same reasons business can go wrong: misaligned incentives, short-term focus, and tunnel vision.

But business is not inherently and inevitably net-negative to society – quite the contrary. So isn’t marketing.

The thing is not all marketing is created equal. It’s just that some people in marketing make the wrong decisions. Just like some engineers ignore safety standards, some doctors overprescribe medications, and some businesses pollute the environment.

Essentially, marketing is simply helping organizations serve more customers.

More specifically, marketing is about finding the right customers who actually need your product, using their feedback to make the product better, and consistently communicating useful information through the right channels and at scale.

With this in mind, I decided to outline a few high-level principles of ethical marketing.

Principles of Ethical Marketing

1. Create value before capturing value.
2. Long-term over short-term optimization.
3. Customer feedback over company politics.
4. Educate customers instead of misleading them.
5. Build products that people want instead of selling the ones you have.
6. Target customers who can benefit from the product instead of targeting everyone.
7. Measure NPS, retention, virality, and product usage in addition to ROI, revenue, and growth metrics.
8. Consider the impact of marketing decisions on employees, other companies, society, government, and the environment.

Here you go, a small manifesto of sorts.

How Marketing Can Be Helpful

Here are just some practical examples:

• Marketing can conduct research on customer needs, preferences, and purchase drivers. This, in turn, can help companies build the right products instead of wasting millions on the development of unwanted products.
• Marketing can target the right audience and help customers discover products they can actually be interested in. This, in turn, can save money otherwise wasted on advertising to prospects who’d never purchase a product.
• Marketing can help educate customers so that they make better decisions and learn to use complex products faster. This, in turn, can help companies retain customers and grow revenue.
• Marketing can create useful content that customers would like to read, watch, or listen to. This, in turn, can help companies build strong brands.
• Marketing can help automate, customize, and scale company communications. Thousands of customers can get product news via email and blog subscriptions. This, in turn, can save money otherwise spent on 1:1 communication done by customer support or sales teams.

Practical Takeaways

• If you are in a leadership position, create the right incentive structure at your company that motivates people to create long-term value.
• If you are on an engineering or product team, engage marketing early to get customer feedback, and build products that people actually want.
• If you are on a marketing team, be conscious of the broader implications of your decisions, and educate others on how marketing can help.

Will all companies and all teams be able to follow all the principles all the time? Of course, not. There are always constraints. You might be stuck with the product you have. You might be limited in terms of what decisions you can make as a marketing team. You might be given aggressive short-term goals, etc.

But even if we shift the balance a little bit, the world might become a slightly better place. And people might say “Wow, that was great marketing” a little bit more often.

If you liked this article, read more about marketing and growth or subscribe to my monthly email newsletter!

Technology and Startups

• Ten-year Futures – a presentation by A16z. New technologies enable new use cases. Seeing them, as well as non-obvious “second order” effects, is key. E.g. mobile enabled Instagram, Instacart, and ride-sharing.
• Decrypting Crypto – another presentation by A16z. Bitcoin is a combination of three old technologies: hashcash, public key cryptography, and distributed ledger. Value of cryptocurrencies goes beyond the traditional store of value and medium of exchange. E.g. tokens can help bootstrap new protocol-level innovation and incentivize developers, customers, and investors to contribute.
• AlphaGo Zero masters the game of Go from scratch. The ML algorithm learned the game without any pre-existing understanding of rules or strategies. Building a general or at least a-little-bit-less-narrow AI appears to be a big priority for DeepMind. Perhaps this can count as a small step in this direction?
• Delivering blood with drones in Rwanda – a TED talk by the founder of Zipline. What an amazing application of new technology and a case study in social entrepreneurship.
• Tacotron 2 is a new text-to-speech technology by Google that is (almost?) indistinguishable from a human voice. If Google manages to make it less computationally demanding and ship it as part of the Android OS, all kinds of interesting use cases will be made possible. I personally will listen to more of my Pocket articles in audio.
• Magic Leap is launching its SDK, shipping in 2018. AR/VR is already quite a saturated market. It’s not entirely clear yet how hyped Magic Leap technology will compare to Microsoft HoloLens, as well as to VR headsets: HTC Vive and Oculus Rift.

Growth and Marketing

• Product Marketing for New Products. Emphasize the fundamental shift in the way business is done and ensuing challenges and opportunity. Explain how your product will facilitate the transition and contrast it with legacy solutions on the market.
• Ship by ProductHunt. Might be a useful tool for early-stage startups.

Product Management and Design

• Intercom on Product Management ebook. The book lays out a framework for doing a feature audit and deciding which features to develop and which ones to kill.
• How Technology is Hijacking Your Mind. Tristan Harris, an ex-Googler, makes a persuasive case for ethical product design and for putting more thought into what metrics tech companies should choose to optimize for.

Books

• Life 3.0 by Max Tegmark jumps from AI to evolution to physics to philosophy in a mind-blowing way. It reads like science fiction that has a non-zero probability of turning into non-fiction.
• Angel by Jason Calacanis describes startup fundraising from an angel investor’s point of view. The book can be useful to both aspiring investors and entrepreneurs looking to raise money. I appreciated Jason’s transparency and focus on details.
• Behave by Robert Sapolsky is a very comprehensive and thought-provoking primer on evolutionary biology, neuroscience, and human science that gets a bit too dense and detailed at times.

Podcasts

• Constitutional by The Washington Post tells the story of America by focusing on one issue at a time. E.g. “more perfect union”, “establish justice”, “common defence”, etc. Get all 13 episodes that were released so far!
• The Kevin Rose Show by Kevin Rose. This one is similar to The Tim Ferriss Show. Most guests are from the technology world but some are bloggers, writers, and other interesting folks. I particularly enjoyed this episode: Bitcoin’s true potential, with Andreas M. Antonopoulos.

Products

• Oura Ring is a “quantified self” gadget that I use to track sleep. It’s been shown to be relatively accurate compared the underwhelming performance of competing gadgets. Oura also breaks the sleep into deep, REM, and light stages. A very detailed review.
• Yousician is a cross-platform app (MacOS and mobile) that can help you learn to play guitar or piano. It has helped me gamify the learning process and reduced the friction by showing a clear learning path.

Numbers

• 1,665%↑ is Bitcoin YoY growth
• 47 years is the average age of founders whose startups had successful exits
• 50% is iPhone’s share in Beijing
• 11%↓ people in extreme poverty in 2015 vs. 1987 (-41% in Africa)
• 2017 in Review by The World Bank

Smarter Living

• A NASA study identified indoor plants that are most effective at improving air quality.
• Dr. Bredesen’s protocol might help prevent Alzheimer’s and optimize cognitive function.

Science and Research

• 24 sessions of HIIT over 8 weeks → 12%↑ VO2 max.
• 4+ cups of coffee → 64%↓ risk of all-cause mortality.

My Blog Posts

• Key Growth Metrics, Part 1: Churn/Retention, and Virality.
• Key Growth Metrics, Part 2: CLV/LTV, and CAC.

Contemplation Break

“The happiness of your life depends upon the quality of your thoughts.” ~Marcus Aurelius

Like blog posts like this one? Subscribe to my monthly email newsletter!