How to build a cohort analysis table in Analysis Workspace.
Hi. I’m Ben Gaines from Adobe Analytics product management. And today I want to share with you one of the most exciting new additions to Analysis Workspace in Adobe Analytics: cohort analysis. For those of you who may not be aware cohort analysis is simply the grouping of your customers or users based on when they completed a certain activity in your site or app and then following those users or customers over time to see how they continually engage with your brand or churn and don’t continue to engage. Within Analysis Workspace cohort analysis is available to add to any project, All I have to do is go to my visualizations and drag Cohort Table over onto my project and I am able to configure it. This tool does require a little bit of configuration so that we know how you want to build the cohorts. Then how you want to track their ongoing engagement or churn. So here I’ve got a six month date range and I’m going to break up my cohorts by month so you’ll see what this means in just a minute. I’m gonna say that I want to group my users by people who placed an order on my site and if I wanted to make that more than one order I could do that here or if I had a metric where I wanted to do less than or equal to I can do that as well. And I want to say that ongoing engagement with my brand is represented simply by return visits to the site. So I’m going to add visits as my return metric and when I run that report I get a cohort table and you’ll see that all of my customers are grouped by the month in which they made a purchase and if they made multiple purchases during different months they would be represented in each of those cohorts or in each of those months here in the included section and you can see that I had about 6,000 customers place orders during the first three months of my cohort analysis and it dropped to around 4,000 customers during the second three months of my cohort analysis. And then for each of those groups I can see how many of those customers came back and visited my site during subsequent months. And in fact I can see that looks like by month 4 I have nobody coming back which is really terrible and gives me an indication that maybe I should engage with these customers earlier maybe month one or month two to make sure that they are aware that they can come back and get great deals on my side or whatever the case may be. The other tremendous thing that you can do with cohort analysis, it’s all color coded making it very easy for you to visually spot trends. I can see that engagement is pretty decent really in the first month after purchase except for May. I can see that without needing to even really comprehend the numbers in the table because I can see that the shading is so light and that may lead me to analyze these people who are a month removed from their May 2015 purchase and try to understand why they fell away so hard.
In fact with that I can take that group and select that cell and create a segment based on that cell and then go analyze against that segment elsewhere in Adobe Analytics. I can do that with an entire column as well if I want to. so I could look at all of the people in month two after a purchase or after these specific purchases to see what they are doing what behaviors they are exhibiting when they come to my site. If I wanted to do that. The only thing to let you know about with those segments is that when it builds a segment it’s looking at these specific customers in this cohort analysis. It is not a dynamic segment that says all people in any date range in their second month after placing an order. It’s these people who are in their second month during the date range that I’ve defined for this cohort analysis. So that’s something to be aware of when you build that segment. Cohort analysis has tons of great uses especially with mobile apps but also with Web sites of every shape and size and variety. So hopefully you find this to be a really valuable addition to Analysis Workspace in that you’re able to work with it in your projects as you go.
For more information on this feature, please see the documentation.