Cohorts

Cohorts overview

Cohorts are groups of people that are interesting to your business–they're the building blocks of your customer predictions. You'll start working with cohorts by creating the most important one–your current customers.

You'll define your customers cohort by choosing which event best describes how you define your customers–maybe it's those individuals that have made a transaction, or who've signed up for one of your services. Regardless of how your business defines its customers, your customers cohort will be the basis of your predictions in Faraday.

When using outcomes to create propensity predictions, for example, you'll point your other cohorts–your leads, prospects, and other interesting groups–toward your customers cohort and ask Faraday to find which individuals in those groups are most likely become customers. Similarly, with persona predictions, you'll choose the cohort that you want to predict personas for.

With your foundational cohort in your customers defined, you'll then create others that best fit the use cases that you're interested in powering with predictions. You'll accomplish this using both your first-party data, and third-party data from the Faraday Identity Graph to create specific events and traits that define each group. Each cohort you create can include events, traits, or a mix of both.

For cohort creation instructions using both the Dashboard UI and API, see our how-to docs for cohorts.

Event-based cohorts

The majority of your cohorts in Faraday will be defined, at least in part, by an event. This event might take the form of a transaction, a form signup, a cancellation, or some other important occurrence. Beyond your customers, you might create more nuanced versions of your customers, like "repeat purchasers" or "churned customers."

As you add an event to a cohort, you're able to specify recency, frequency, earned value, and other properties to allow you to add these more nuanced restrictions to cohorts to empower other predictions.

• Recency: allows you to set days-ago ranges on the event. An example of this would be if you're interested in creating a cohort based on customers who've placed orders between 30 and 90 days ago.
• Frequency: allows you to set a range on the number of occurrences of the event. An example of this would be if you're interested in creating a cohort based on customers that have made more than 3 orders.
• Value: allows you to set a range on the value of the event–generally orders. An example of this would be if you're interested in creating a cohort based on customers who've spent over $100.

Trait-based cohorts

Using both your data and Faraday's, traits can be added to cohorts to create further nuances. This can be useful to in several scenarios:

• Geographies: if you're looking to expand with a new physical storefront, you might score cohorts of leads in specific geographies to find the highest-impact location.
• Subgroups: if you're a business that focuses on a specific subgroup of people, is launching a new product focused on a specific subgroup, or something similar, you might create a cohort of people based on that subgroup–like "all women leads" for a womens' swimwear brand.
• Direct mailers: if you want to launch a direct mail campaign, you can create a cohort using geographic targeting in addition to mailer-specific traits, such as "USPS-indicated in service" to ensure they can receive mail.

Deleting a cohort

Before deleting a cohort, ensure that resources using it, such as a pipeline or an outcome, are no longer using it. Once there are no other resources using the cohort, you can safely delete it.

• Dashboard: click the options menu (three dots) on the far right of the cohort you'd like to delete, or upper right when viewing a cohort, then click delete.
• API: use the delete cohort API request.