Cohorts

A cohort is a saved group of people who match rules you choose. That is the whole idea, so do not let the word put you off. You decide the rules, for example “people who bought something in the last 30 days” or “people who signed up but never came back”, and leadmaps gathers everyone who fits into one named group. Once you save it, that group is yours to reuse. You can drop it onto your other reports as a filter, so instead of looking at everyone, you look at just the people in that group.

Think of it like a saved search at a library, or a smart playlist in a music app. You set the rules once (“songs I added this year that I have played more than ten times”), give it a name, and from then on the list keeps itself filled with whatever matches. A cohort is that smart, self filling list, but for the people who use your product.

The reason cohorts are powerful is that “everyone” is rarely the question you actually have. You usually want to know about a specific kind of person: your power users, the people a campaign brought in, the customers who are slipping away. A cohort lets you define that kind of person once, name it, and then ask every other report about just them, instead of rebuilding the same filter by hand every single time.

What you see

The page lists your saved cohorts and lets you build new ones.

• A page title that reads “Cohorts”.
• A list of your saved cohorts. Each one shows:
  • Its name. The plain name you gave it, for example “Bought in last 30 days” or “Signed up, never returned”.
  • Its size. Roughly how many people currently match its rules, so you can see at a glance whether it is a big or a small group.
  • Its rules. A short summary of what someone has to do or be to belong, so you remember what the cohort means without opening it.
• A builder for making a new cohort, where you add the rules that define who belongs, then save and name it.
• A way to take any cohort and apply it as a filter on another report.

A “cohort” is the saved group. A “rule” is one condition someone must meet to belong, for example “did the purchase event” or “first seen in the last 30 days”. When a cohort has several rules, a person usually has to match all of them to be included, which lets you describe a precise kind of person.

How to read it

Read a cohort as a sentence: these are the people who did these things, and there are this many of them.

Here is a worked example. Say you want to study your most engaged customers, so you build a cohort with two rules: “completed a purchase” and “had at least 5 sessions in the last 30 days”. You name it “Power buyers” and save it. The list now shows:

• Name: Power buyers.
• Size: 1,420 people.
• Rules: completed a purchase, and 5 or more sessions in 30 days.

What does this tell you on its own? Just that 1,420 people are both buyers and frequent visitors, which is a healthy core to have. But the real value comes when you put this cohort to work. Open the sources page and apply the “Power buyers” cohort as a filter, and now the page shows where your best customers came from, not where everyone came from. You might learn that your power buyers are far more likely to have arrived from a particular channel, which is the kind of insight that changes where you spend. Open the funnels page with the same cohort applied, and you see how your best customers move through your product, which can reveal the path you want to nudge everyone else toward.

The point of the worked example is this: a cohort by itself is just a count. Its power is as a lens you carry to every other report. Define the group once here, then look at the rest of leadmaps through that lens.

One thing to keep in mind: a cohort is a living group, not a frozen snapshot. As people start or stop matching its rules, they move in and out of the cohort over time. “Bought in the last 30 days” describes a different set of people next month than it does today, which is usually exactly what you want.

Customize it

You define the rules, save and name the cohort, then apply it wherever you like. Here is each control and when to use it.

• Set the rules. Add the conditions that decide who belongs, for example “did a certain event”, “has a certain property like plan = pro”, or “was first seen within a date window”. Add more rules to narrow the group to a precise kind of person, since a person typically has to match all of them. Use a single rule when you want a broad group (“anyone who purchased”), and stack rules when you want a sharp one (“purchased, and on the pro plan, and active this month”). The rules are the definition, so spend your thought here.
• Save and name it. Once the rules describe the group you want, save the cohort and give it a short, clear name. Name it for what it means in plain words (“Trial users week 1”, “At risk of leaving”) so that you, and anyone on your team, instantly understand it later without reading the rules. Saving is what turns a one off filter into a reusable group you can apply anywhere.
• Apply it as a filter anywhere. Take a saved cohort and use it to narrow another report, so that page shows only the people in the cohort. Apply a cohort when you want to ask a normal question (where do people come from, how do they move through a funnel, do they return) about a specific group rather than everyone. This is the main payoff of cohorts: build once, reuse everywhere.

Use cases

• Study your power users. Build a cohort with rules that capture your most engaged people, for example “purchased” plus “many recent sessions”, and name it. Action: apply it across your sources, funnels, and returning visitors pages to learn where your best customers come from and how they behave, then aim your marketing and your product at attracting and keeping more people like them.
• Build a campaign audience. Define a cohort that matches the exact people you want to reach, for example “signed up but never purchased” or “bought once more than 60 days ago”. Action: use that group as the audience for a targeted message or offer, so you speak only to the people for whom it is relevant, and then watch whether the cohort shrinks as those people convert, which tells you the campaign is working.
• Compare two groups side by side. Create two cohorts that represent different kinds of people, for example “came from search” and “came from paid ads”. Action: apply each one in turn to the same report, like a funnel, and compare how the two groups behave. If one group converts far better, you learn which kind of visitor is worth more, and where to put your effort.

Tips

• Name cohorts in plain words, not in rules. A name like “At risk of leaving” tells you and your team what the group means at a glance, while a name like “no_event_30d” makes everyone reopen the rules to remember. The clearer the name, the more the cohort gets used.
• The power of a cohort is reuse, so save the ones you will ask about again. If you find yourself building the same filter by hand on different pages, that is a sign it should be a saved cohort. Build it once, then apply it everywhere.
• Stack rules to sharpen, but do not over narrow. Each rule you add makes the group smaller. A precise cohort is useful, but if you pile on so many rules that only a handful of people match, the reports you apply it to will look empty and tell you little.
• Remember a cohort changes over time. People flow in and out as they start or stop matching the rules, so a time based cohort like “active this month” describes a moving set of people. That is usually the point, but keep it in mind when a cohort’s size shifts.
• Start from a question, not from the rules. The best cohorts answer something you actually want to know, like “who are my best customers” or “who is about to leave”. Decide the question first, then write the rules that capture those people.