Who are my visitors that purchased last month and have returned to the site this month, but not purchased?

The segment needed to answer this question is fairly complicated.  I recommend reading Kevin’s recent post on Discover AND/OR Segmentation, to better understand nesting containers, which we will use a lot.  To begin creating this segment you must launch the Segment Filter Builder.  Now, let’s breakdown the question above.

Visitors that purchased last month

1. In order to successfully analyze the group of visitors that purchased last month, we need all their visitor data, which means the first thing you will do is drag over the Visitor container onto the segmentation canvas.

2. The visitor must have had at least one visit last month, so you will drag over the Visit container and nest it within the Visitor container.

3. Next, “click here to define Visit”.  This is where we will define the time period that the visitor came to the site.

4. Within the rule list select the report “Date: Month” and select the month you want to define as last month.

5. Finally, since we are requiring the visitor to have purchased, you will drag over the Orders event container and nest it within the Visit container.  Your segment should now look like the following.

Visitors that returned to the site this month

1. Adding on to the segment above, we want a visitor to also have visited the site last month, which means we need to nest another Visit container within the Visitor container.

2. As above, it is within this new visit container that we will define the date range, so “click here to define Visits” and create a rule where “Date: Month” equals the month you want to define as this month.

3. So that I can make sure that on the visit this month the customer expressed interest in purchasing I am going to add one more criteria to the segment, that a Product View occurred on the visit.  To add this criteria drag over of the Product View event container and nest it within the Visit container for this month.

Visitors that did not purchase this month

1. Finally we need to ensure that we don’t include visitors that purchased last month and this month.  Click over to the Exclude segmentation canvas.

2. Drag over the Visitor container.

3. Drag over the Orders container and nest it within the Visitor container.

4. Since the order had to occur last month you will “click to define Orders”.

5. Add the rule where “Date: Month” equals your selected month.

6. Don’t forget to save the segment before you close the window.

If you are really familiar with segment building then you may have noticed that I could have simplified the segment.  I have chosen not to in this instance to make it easier to follow the segment creation process.

So what now?  If you plan on using the information you gather in Discover from this segment for targeted email remarketing then you may approach the analysis like the following.

1. Make sure the time period of your project includes both time periods in your segment.  In my example I would need to have the time period include both July and August, not just August.

2. Launch the report Site Metrics > Visits

3. Expand the Chart Options so that you can change the granularity to Month

4. Drag over the segment you just created.

5. Drag over the metric “Product View” under the new segment

6. Breakdown the second month you are analyzing (in this case it is August 2010) by the Products report (or some classification of that report)

7. You may want to look for a product that you would want to feature in an email and then breakdown that product or group of products by your report containing unique user ID.  Now you already know the person is interested in the product and that they have purchased recently, so they are more likely to purchase again.

To be able to pass these IDs on to your internal team that manages email campaigns or to your email vendor you will want to first make sure that you have all the values showing.  At the bottom of the report you will see “Show” and then a dropdown.  From the dropdown select the number of values you need.  Now that the report is set up perfectly you can either copy the report to the clipboard (see the 5 Helpful Discover Tips post for details) or you can email the report using the envelope icon above the workspace.  Don’t forget to save these reports as a project so that you can come back to the report and just change the dates in the segment to get the data for the next time period.

I hope these examples were helpful and sparked ideas of your own on how you can use this functionality to perform actionable analysis for your business.  As always, if you have any questions or comments, feel free to leave them below.

Is campaign X driving loyal visitors?

There are so many different areas of analysis to get a clearer picture on the performance of campaigns.  If you are a business that values loyal visitors, then whether or not a campaign drives repeat visitors is an important measure of success.  For the purpose of this blog, we are going to use a simple definition for loyal visitors as visitors who return to the site

The following steps are how I would approach creating a segment to answer the question above.

1. Launch the Segment Builder.

2. Drag over the Visitor container.  Since our criteria will span multiple visits we want to use the Visitor container.

3. Drag over the Visit container so that it nests within the Visitor container.  The nesting allows us to identify that within a single visit several criteria occurred and that we want to see all that visitors data if they had at least one visit that fit the criteria.

4. “Click here” to define the Visit.

5. Select your company’s campaign or tracking code report and set it equal to the campaign you want to analyze.

6. Then “Add Segment Rule”.  In the second rule find “Date: Month” for the report and then select July 2010 for the month.  Make sure that you are matching ALL of the rules.  Once you have done this click “OK”.

7. Save your segment to your Favorites or another Custom folder.  Since the Date segmentation criteria is not compatible with SiteCatalyst, you will receive an error if you try to save a time based segment to the SiteCatalyst folder.

To answer the question “Is campaign X driving loyal visitors?” you would want to do the following:

1. Go to Site Metrics > Unique Visitors

2. Unhide “Chart Options” at the top of the report and find the Granularity option and select Month.  By changing the granularity we will now see Monthly Unique Visitors for each month and then the total at the bottom of the report are true unique visitors for the entire time period viewed.

3.   Find your segment in the correct folder in the Segments menu and drag it over to the Unique Visitors report.

Notice that the total Unique Visitors matches the total Monthly Unique Visitors in July and the total Monthly Unique Visitors in August is significantly less, just 6% of the total.  Therefore, 6% of visitors that entered on the campaign in July returned to the site in August.

Now that you have created this fabulous segment you can run any of the reports in Discover to try to understand what drove visitors back to the site and then utilize an optimization tool like Adobe Test & Target to encourage more visitors to return to the site.

More examples to follow in the next post…

1.  Quick segment creation

Often I find myself looking at a report in Discover and I think that a particular value in a report is interesting, interesting enough for me to want to dig deeper, which requires me creating a segment.  Instead of going to the segment builder from the left menu, you can create the segment from within the report itself.  Just right click on the value and you get a menu like the screen shot blow.  Once you select which container you want to use, Discover creates the segment and automatically places it in the report that you are in.  You can then right click on the segment to save it if you want to keep it for future use.

2.  Quick report change

I compare segments, a lot.  When comparing segments it takes a little bit of time to get the report set up exactly like I want it with the right segments.  Instead of having to go through the same process for all the other reports you want to view you can change a report with the right click of your mouse.  Just right click on the report name and go to Change Report and you will see the reports menu.  If you want to keep your original report you can copy the report first and then perform this modification on the copy.

3.  Breaking down more than one row at a time

You may want to compare multiple values in a report broken by the same report.  You can click on the green plus button by each one of them and select the breakdown report.  To save you a little bit of time you have the option to use the CTRL key to select several values and then click on the green button next to one of the values and select your breakdown report.

4.  Copy to clipboard

How many of you have been wishing and waiting for trended fall-out data?  Well, if you have Discover it is already here.  I think this is one of those features that people don’t notice or forget, which is why I wanted to take the time to cover it today.

Within the Discover interface you get a view of what the trend looks like with the spark line next to the conversion number.  If you would like the raw data it is as simple as opening up Excel, flipping back to Discover, clicking the Copy to Clipboard button and pasting in Excel.

You will notice a lot of data when you paste the fall-out report into Excel.  The first section includes the definitions of what each checkpoint is defined as.  The second section shows the totals for the top next pages from the checkpoint when visitors fall-out (do not eventually make it to the next checkpoint) or fall-through (continue on to the next checkpoint).  The third section is where you will find the trended data for the fall-out, including the percent of traffic that continues on to each step/checkpoint.

5.  Comparing the same segment over time

Sometimes my analysis starts with questions like, “Why was there a big jump in visitors from Canada to my website?”  This analysis requires me to compare the same segment for two different time periods, which is super easy to do in Discover.  Just create your segment and once it is saved you can drag it over to the report twice, so that you have two instances of the same segment in the report.  Then at the bottom of the report you will notice you now have two lines on the time range, one for each segment.  Use the handles at each of the time line to select the two time periods you want to compare.  Because the segments are different sizes I like to look at the percent of total instances each line item represents.  As in the screen shot below you may notice that a significantly greater percent of traffic came from one referral source, like in this case “yahoo.com” (5.51% vs. 3.67%).

You might be surprised to know that some of these little golden nuggets and more can be found in the Discover User Guide (go to Help > Documentation within SiteCatalyst).  I can hear the groans now, you probably dislike reading manuals as much as I do, but I promise you, this one is well worth your time and will allow you to use Discover more effectively, so I encourage you to read through it and create yourself a little cheat sheet of the golden nuggets you find.

If you are new to Discover, you may wonder, “Where are my Search Keywords – Paid and Search Keywords – Natural reports?”  To get these reports you need to first load the Paid Searches report in Discover, which is found under the Traffic Sources menu.  From here we can break down paid or natural by keyword to see our paid or natural keywords.

What is the impact of all my paid search initiatives on my business?

Please note: If you have set up your paid search to be recorded in the campaigns variable and you have utilized SAINT classifications to roll up all paid search campaigns then you can answer this question in SiteCatalyst.

1. Drag the Orders and Visits metrics to the Paid Searches report.  Note: you can select any metric that you would consider a key success on your site.
2. Create a calculated metric for conversion rate.  In this example I am going to define it as “Orders / Visit”.  If you are unsure of how to create a calculated metric in Discover, you can reference section 6.5 of the Discover User Manual.

This report quickly illustrates to me the value of my paid search initiatives in driving a significant amount of traffic to my site (21% of total visits) and converting those visitors.  In fact, I see now that my paid search initiatives are converting visitors at higher rate than the site average (13.19% vs. 10.35%).  After learning of the stellar performance of paid search, compared to other channels, you may ask…

Which of the keywords that I purchased are converting the highest?

Not every keyword is created equal and we can get a better idea of which paid search keywords should receive some TLC, or if certain keywords should just be dropped.

1. Click on the green plus sign next to Paid to break down paid search by search keywords.
2. Sort the Search Keywords by Order by clicking on the arrow in the sub-column header.
3. 
4. Of the top five keywords (by orders), “jj esquire” is the lowest performer, converting at 12%, compared to more generic keywords like “jeans”, which converts at 16%.  So now the next question I have is…

Why is keyword “jj equire” converting at a lower rate?

There is a lot of data available to try paint the picture of why one keyword may be underperforming.  I will try to take you though a logical flow, but when doing your own analysis you may skip a step, if it is not relevant to your business.

1. My first thought is maybe there is a difference in where visitors from the “jeans” keyword and the “jj esquire” keyword are landing on the site.  This can be uncovered by breaking down both terms by Paths > Pages > Pages Entry.  Tip: If you have more than one row highlighted when you click on the green plus button to breakdown then all the rows highlighted will breakdown.
2. 
3. In both instances, the majority of visitors (99% and 83%) are landing on the Home Page.
4. Are more of the visitors from the “jj  esquire” keyword exiting immediately from the Home Page than visitors from the  “jeans” keyword?  To answer this question, drag over the metric Single Page Visits and create a calculated metric for Bounce Rate (Single Page Visits/Entries) to add to the report.
5. 
6. Again, we see similar behavior from the Home Page for both keywords.  A possible next step would be to A/B test changes to the Home Page better tailor the user’s experience.  We know that “jeans” converts at a higher rate, so I would consider using a banner to highlight jeans to help drive visitors from generic brand terms like “jj esquire” deeper down the conversion funnel.

You may still be scratching your head, trying to understand why there is a difference in conversion between these two keywords.  Perhaps asking the following questions will lead you to the answer you are yearning for.

Since visitors are not immediately leaving the site at a significantly greater rate, are they not converting at a higher rate because they cannot find what they are looking for?

1. Create a Visit segment where Paid Search = Paid and Search Keyword = jj esquire.  Note: the Paid Search and Keywords criteria need to be defined together in a Page View container so that we guarantee that “jj esquire” was a paid search.  Then the Page View container must be nested within the Visit container so that we can capture visitor path data in the segment.
2. 
3. Open the Next Page Flow report (Paths > Pages > Next Page Flow) and drag your newly created segment to the top of the report.
4. 
5. A significant portion of visitors navigates to the Children & Toys section and then to Search.  With this greater insight into visitor behavior, you may want to test different configurations of the Home Page to highlight key areas of interest for visitors entering the site from the paid search term “jj esquire”
6. In contrast, when visitors are coming from the paid search team “jeans”, a greater percent of visitors navigate directly to the Womens section of the site.  This difference in behavior form the Homepage for the “jj esquire” and “jeans” terms further highlight the opportunity to increase conversion by providing a more targeted experience, either through landing pages or targeted offers.
7. 

There are many different paths for analysis to find opportunities to optimize your paid search and I have only covered a few of them.  You will notice that with my sample data set, the results of the analysis did not  yield that golden nugget, excitement inducing result, but  I hope that this sample analysis will inspire you in your paid search optimization efforts.  If you have a question that you would really like be to able to answer, but don’t know how to go about it, feel to post it in the comments and we will consider it for a follow-up post.

A recent Forrester research article states that, “As the economy rebounds, many companies are seeking to use a great customer experience as a way to retain customers, grow wallet share, and bring new customers on board.”  The first step in creating the best online user experience is to understand overall how visitors are using your site, which can be accomplished using analytics tools such as SiteCatalyst.  Usually as you start to understand your visitors on a whole, questions arise about who these visitors are and how they can be grouped to provide better experiences.  This is where a real-time advanced segmentation tool, such as Discover, can help.

You may be looking to optimize the user experience based on how users got to the site, what content they view, products they purchase, articles they download, how frequently they visit the site, etc.  You may want to identify characteristics or behaviors that visitors have when they convert and optimize the site to encourage other visitors to follow a similar path.

The purpose of this blog is to help you with these questions.  We really want to provide you with insights and examples of how you can drive actionable analytics through segmentation more quickly and effectively.  All posts will fall under one of three categories:
1. Fundamentals:  Introduction of things that we would consider to be fundamental, such as segmentation best practices, using metrics in the correct context, etc
2. Enhancers: Tips and tricks to enhance how you can take advantage of all the functionality in Discover
3. Analysis: Sample analysis and case studies utilizing different features of Discover to help kick start your own analysis

Because of the wealth of knowledge and experience here at Omniture we are going to tag-team the posts between three main authors.  Kevin Willeitner, the Discover Product Lead within Omniture Consulting.  Laura MacTaggart, Discover’s biggest fan.  Tim Lott, the illustrious Product Manager for Discover.   This tag-team approach also allows us to regularly post articles to quench your Discover thirst.

We are very excited to share our knowledge and experience with you and we also look forward to receiving feedback or ideas from you.  If there are specific topics you are anxious to see covered, you can submit them in your comments. And don’t forget to add your Discover feature requests to the Omniture Idea Exchange (http://ideas.omniture.com/). This is the best way to get your ideas in the proper forum where they can be voted on and hopefully implemented in a future release.