In this article
An overview of the Cohort Builder and it's functions.
The Cohort Builder is used to create cohorts in Permutive DMP. If you'd like to know more about what cohorts are, read our Introduction to Cohorts.
The screenshot above shows a basic cohort being set up in the builder. There are a number of features available to you when creating cohorts, which will be covered at a high level in this document. The best way to understand the Cohort Builder is to try building a cohort.
Note: When building cohorts utilizing'
client.url contains' avoid the trailing
/ on the URL, even if that is part of the full URL in the browser. This is because our SDK ‘sanitizes’ the URL for Insights, i.e. example.com/123/ would be counted towards example.com/123
Cohorts are built around events. This is the first thing you need to select when building a cohort, the event you want to base it on. The simplest cohort you can build is one that states whether an event has been performed, or not, as shown below.
The next factor you can use in cohorts is time. This allows you to define the recency in which a user has to have performed (or not performed) your event to be included in the cohort.
There are a lot of different options for specifying the time range. Below is an example of selecting people who haven't performed a Pageview in the past 30 days.
You can also specify the number of times a user has to have performed the action to fall into a cohort.
Below is an example of a user who has to have performed a Pageview 3 times to fall into the cohort.
Another way you can filter people's behavior to include them in cohorts is using properties. When events are sent into Permutive DMP, they include properties (eg. client.url in the Pageview event). These properties are then available in the cohort builder.
You can link together as many different properties as you like. Below I've specified properties around the URL or title to create an example 'Sports Enthusiast' cohort.
AND / OR
What we've run through so far are the basic building blocks of a cohort.
Once you've specified one criteria, you can chain as many different statements together using AND and OR rules.
In the example below, I've used a few of the cohort builder tools mentioned so far, combined with an AND.
Once you're happy with the setup of your cohort, you can see how many people would have fallen into it over the past 30 days, had your cohort been set up. This is done by clicking 'Calculate' next to Audience Size.
Once it has finished calculating, you'll have an idea of whether the cohort you've built is suitable for its purpose. If your cohort is too small, you might want to think about broadening your criteria.
In our example above. The audience for our cohort is 3.8M, 5.8% of all users.