When starting a business it is very easy to envision a compelling end-state. Functionality that will be really useful once everyone is using your site/service. However, it is much harder to come up with something that anyone cares about day-one. Great examples of this can be found in any plan that involves a marketplace, reviews, chat, dating, benchmarking, P2P etc… the first customer gets very little. It is important to know where you are going with a business, but it takes relatively little genius  to imagine a service if you assume you can start with several million users day one. The real cleverness, and where startups need to focus, is the stepping stones from A to B (and not B).

With this (not so original insight) in mind here is my analogy (somewhat revealed by the title of this post):-
Building a startup is a bit like building a bridge. Once complete, its value is clear, but along the way you need to be very clever to make people use it. How does one get early adopters for a half-built bridge?
Step 1) Your 1/4 built bridge is a marine observation platform - it gives people an excellent viewing station to observe fish
Step 2) Your 1/2 built bridge is an urban bungy jump -  it provides an excellent extreme sports thrill
Step 3) Your 3/4 built bridge is a ‘Dukes of Hazard’ leap -  allows anyone willing to pay to play action hero in their car
Step 4) It is a bridge! - everyone one can see it - it takes no explanation!

By way of a quick history re-cap, the Internet promised to remove ‘intermediaries’ from the global value chain, thus making everything cheaper as there are less mouthes to feed between the ‘manufacturer’ and you (the end consumer). At the moment, in complete contrast to this vision, we are seeing the emergence of ‘Super Intermediaries’ in the form of Apple, Google and (to a lesser extent) Microsoft. These Super Intermediaries pose a macro threat to the current crop of ‘Web Portals’ (e.g. Expedia and Amazon), but don’t represent a fundamental collapse in the value claimed/charged in getting product to consumers.

The Semantic Web concept, put very simply and perhaps a bit incorrectly, pointed to a vision of ‘my robot’ chatting to ‘your robot’ in order to find things I’m interested in. In this world, the idea of a shop-window doesn’t exist, you buy direct from the ‘manufacturer’ and the intermediary role is displaced by an algorithmic ‘negotiation’. Brand/’Manufacturer’ advertising/promotion remains very important in creating desires for the products I search for in the first place. ‘Manufacturers’/Brands still need a logistics partner to get product to you, but the idea of Retailer brands and ’shopping malls’ declines. Social Recommendation and exhaustive algorithmic search displaces the art of shelf-placement and product buyers. Clearly this vision has not happened yet. Certain elements of it have, but these have simply been presented as new forms of Portal with a functional speciality (e.g. Kelkoo, Priceline)

So what do I mean by ‘Twitter for Industry’. Twitter is many things, but viewed from one angle it is a ‘message bus’, filled with noise that you extract points of interest from based on filters. You could view it a bit like Enterprise Application Integration (EAI) ware (if you are a deep SI geek). In the Semantic web-vision, ‘Manufacturers’ would ‘publish’ their products to a ‘Twitter’ like message bus (a bit like rich-snippets being published on the web). Robots, or ’shopping apps’ would then trawl this Twitter-sphere, using the product pointers identified in the stream to initiate ‘negotiations’ with the publisher. Publishers who lie to attract unwarranted attention would be ‘punished’ by having their ‘reputation’ marked down. A third party (either an entity of agreed code-base) would exist to keep the system ‘honest’. A world of permission based advertising would open up, but with ‘robots’ acting as your first filter. You state what your interested in (either directly as product, or people who’s product taste you respect) and the world comes to you with click-to-buy suggestions. Given the open nature of the ‘message bus’ any shopping app could exploit it - it is in this way that the Portal role is ‘destroyed’ and displaced with client side code you ‘trust’ or ‘train’. Retailer/Portal brand equity (and the value in commands) is displaced by ‘open source algorithms’ operating on client-side software (whose deployment is now made easy in the form of Apps).

Fancy (www.thefancy.com)  is an example of someone potentially exploring this vision, but still very far from what I’ve described. All of these visions require ‘Manfacturers’ to be willing to sell direct and upset their existing channels (many of which do so already, as Portals are actually only ’standing in front of them’ on the web). It also requires (for physical goods) the existence of logistics companies to provide product delivery - something the mailing giants are already moving into. The big question in my mind is, who has enough money and power to bring this concept to life, who doesn’t already profit from maintaining the status quo?

First off, payments is a tough space. Consumers like innovation a lot less than people think, and in the area of payments and personal finance people are even more risk averse. Often the perceived risk of trial vastly outweighs the immediate benefits (e.g. I’ll pay 5 seconds faster, but may have all my money stolen).

Second off, payments is a heavily regulated industry dominated by a large oligopoly. You have to ‘dance with elephants’ and may easily get crushed if you choose the wrong part of the value chain. In general there are very few startups that have succeeded in payments - the odds seem worse than other industry areas.

Thirdly, payments is less interesting than you may think! At least the handling of the money is. People look at high percentage fees and instantly assume this is valueless profiteering by a monopoly system. I’m not saying it is fairly priced, but there are hidden costs in payments which drive/require fees (e.g. fraud/insurance).

So, all that said, what would I invest in. I see a few themes, each with different merits:-

1) Increase consumer Point-of-Sale conversion
- 1.1 Increase the addressable market to those without debit/credit cards (unbanked, youth …)
- 1.2 Reduce PoS hurdles by reducing the time-to-pay (stored details, NFC)
- 1.3 Reduce the effective price (by channeling back payment fee savings)
- 1.4 Offer ‘easy’ credit at the Point-of-Sale (storecards, BillMeLater)

2) Deliver better merchant service levels
- 2.1 Reduce barriers to acquiring an initial MID (the first step to accepting card payments)
- 2.2 Deliver ‘betters. faster, cheaper’ payments processing (a matter of scale and attitude)

3) Payments as Identity based plays
- 3.1 Customer identification and behavioural profiling (with possible linkage to PoS data)
- 3.2 Consumer facing loyalty and discounting schemes

Of the themes above I would say that 2) is best suited to growth capital/private equity. You need a modicum of technology, but you mainly need funding to build and buy a large merchant customer base. Number 1) is increasingly difficult, but if I were to place a bet it would be on the PoS credit concepts - big plays like Klarna and BillMeLater can still be improved upon and rolled-out wider. Number 3) is where the garage start-up may yet unsettled the status quo. Depending on how rules & regulations are amended, it may be possible for the treasure trove of payments data to be exploited by players outside the existing ecosystem.

In general, I think payments will trend towards enabling big companies to do what you see happening everyday (in a very low tech way) in your average ‘cafe’:-

1) You’ll pay after you’ve consumed/seen your goods (* but based on a ’soft’ credit check you may be asked to pay up-front if you look a little rough around the edges - the check won’t ask any questions!)

2) If you’re a regular you may get a secret discount at the till if the owner thinks its important to you (* so as not to upset other customers who pay full price, nor embrassas you with vouchers etc…)

3) When you walk in they know you, and know what you like - you’re coffee may even be waiting for you (* over time the staff pay attention to what you like, and if they haven’t seen you in a while - they make you feel extra special [and a little guilty])

4) If you’re a really good customer, you’ll be allowed to ’settle up tomorrow’ - not because you need the credit, but because they know it makes you feel special - and makes you come back tomorrow!

All of the above are being worked on, with various degrees of buzz-wording around NFC, Mobile, SMB, Card Readers etc…. I think the winner is likely to be the company that attaches itself least to the new technologies, and makes the maximum use of existing data and existing tech.

Topic continued over a coffee/beer …

It should not go unnoticed that mobile telcos already have a directory of always on identities, which they are able to observe the behaviour of (they’re your ISP when you’re not on WiFi). They have not however managed to leverage this great asset from the world of voice and SMS into the internet value chain. There have been passed attempts, most notably Phorm, which failed to respect consumers and the existing internet ecosystem players. More recent forays in the space seem to lack ambition, whilst the OTT players are gaining in speed and scope (see X.Commerce (https://www.x.com/) from eBay and Facebook Credits rolling out across 3rd party sites).

If Telcos do not act, by necessity over the top players will be forced to re-create or work-around their assets - at the requests of both merchants and consumers. Complicating things further for Telcos, device association to network connectivity (for data) is now as low as 20-30%, forcing telcos to innovate their own Over-The-Top solution to let them span these ‘off-net’ experiences. All this can be done, the question is to they have the ambition to do it? Signs are we may see a resurgence of ‘operator portals’ (dressed up as App Stores or Deal Sites) but little engagement with the wider world. This would be a tragedy for the industry, as we all watch more and more of the ecosystem drift State-side.

I want a series of folders, each shareable with a difference group of people. The folders should be accessible via my naive drag-and-drop GUI in Windows and OS-X. They should also be accessible via my mobile and tablet. The folders should be automatically sync’ed over the internet (being hosted either locally or in the cloud).

The service would be a bit like drop-box, except it would allow multiple different ‘drop boxes’ per machine (with different sharing profiles). It also would have full integration with the host OS (for drag and drop in standard GUI). The fact that it could leverage local storage instead of cloud resources would help to keep the service costs (and therefore fees) nice and low.

I’d use the service to share photos with relatives (the kind who don’t use Facebook etc…) and documents amongst co-workers and across machines. It’s not much of an ask, but nothing seems to offer this setup in an easy to use way straight out of the (drop) box.

If today’s operators had to invent voice services, what approach would they take?

They could go for a horizontal infrastructure play, focusing on connecting individuals via a generic mechanism. Businesses and consumers would use the network for any and all types of conversation. It would start out as simple telegram displacement service and would find its way to becoming the everyday chit-chat mechanism we know today.

Or, they could decide to launch a vertically integrated approach. They’d sign deals with leading voice artists and have conversational themes people could sign-up for in bundles. They’d launch with something like weather chat and expand over time. The service could be merged with radio partners and lead to a 2 way interactive audio play in the fullness of time.

The point of the above example is to highlight what has become a widespread fear amongst telcos - that if they expose ‘infrastructure’ functionality to the wider world, without wrapping it up in a ‘higher value’ service, they will fail to capture enough value. This is of course the exact opposite to how internet players work today - they seek first to control the infrastructure layer, and then monetize from their position of strength (see Google, Facebook, even Twitter [but not yet]).

I’d like my calendar alarms to not only be a fixed time in advance, but rather vary the warning time given based on my distance from the meeting location. I’d set a standard 10 minute warning, but if he Calendar detected I was more than 30 minutes from my destination then the warning would be brought forward. Exactly when the alarm would go off would be a mix of calculated distance and guessed mode of transport (details , details …)!

I’d like a soft-phone, which would work over my home wi-fi, loaded onto my iPhone. It would only go over WiFi as far as the router, and then would convert back to regular fixed telephony. I don’t want a Pico-Cell, I don’t want a VoIP phone. I want a sort of DECT via WiFi phone. I’d get landline voice quality, would save the cost of a physical DECT phone, and BT would get more call revenue. Ideally they could build the software into their existing HomeHub rather than needing a new adapter. If BT are listening … I can wait for 6 months for a solution!

As a general theme, new mobile technologies and service that reduce the need to input, or simplify the input process, have got to be a good bet.

It is somewhat ironic that the new voice services require (I believe) more meaty processing power on the device. Given the devices are/were built in the first place to communicate voice back to a giant cloud/exchange, it is strange that voice recognition is not a prime candidate for centralised processing power (please correct me if Siri is using this approach).

[Update] I believe Siri has to have an internet connection to work - not sure how it shares the load between local and cloud however

On a different note, the use of open WiFi (as the easiest way to let users ‘roam’ onto WiFi) may accelerate the use of end-to-end encryption, further damaging the value potential of Operators. Innovations like Firesheep will quickly either reduce the use of open WiFi networks or encourage greater HTTPS usage by major websites.

* The offloading is unmanaged in the sense that operators are not involved in the authentication process, nor are they able to see the traffic being delivered over the access network.