How a cohort exploration works
A cohort is a group of users defined by when they did something for the first time — usually their first visit, though it can be any event you choose. The exploration puts each cohort on its own row and each following period in its own column, so you read left to right to see how much of that group came back a day, a week or a month later.
Three settings decide what the grid actually means. The inclusion criterion says what puts a user into a cohort. The return criterion says what counts as coming back — any visit, or something more demanding like a purchase. The calculation decides whether a returning user is counted in that period only, cumulatively, or only if they have returned in every period since.
You can also change the metric in the cells. Instead of the number of active users you can show event counts or revenue, which turns the same grid from a retention view into a value-per-cohort view.
Why cohort exploration matters
Most reports mix everybody together, so a month of heavy acquisition looks like a good month even if none of those people ever came back. Cohorts separate the two: acquisition is the height of a row, retention is what happens along it. A business with weak retention and strong acquisition is renting an audience rather than building one.
It also grades your marketing by source rather than by volume. Add a breakdown and you can see whether the visitors from one campaign, country or channel behave differently in the weeks after they arrive. That is often a stronger argument for shifting budget than any first-visit metric.
Common mistakes with cohort exploration
The first is running it on a site where returning is not a meaningful action. A trekking agency or a law firm sells something people buy once; low retention there is the nature of the business, not a failure, and the grid will only depress the person reading it.
The second is comparing an incomplete cohort with a finished one. The most recent row has not had time to fill its later columns, so it will always look worse. The third is leaving the return criterion as any visit when what you care about is a purchase or an enquiry, which flatters the numbers considerably.
How to act on it
Decide first whether repeat behaviour is part of your business model. If it is — ecommerce, subscriptions, training, anything with a second sale — set the return criterion to the action that carries value, not to a page view.
Then read down the first column to judge acquisition and across the rows to judge retention, and treat the newest cohort as unfinished. Where a cohort from one channel holds up better than another, that is a signal worth acting on in your measurement plan and your budget. Where retention is the weak point, the fix usually sits in email, onboarding and product rather than in more traffic, and a broader customer lifetime value view will tell you how much that repeat business is worth.