Measure retention
See how many people came back in each day, week or month after their first visit or first goal, and how identifying signed-in users makes it accurate.
Retention shows whether people come back. Everyone is put in a cohort by when they first visited your website (or first completed a goal, like a signup), and the table shows how many of each cohort came back in each day, week or month after.
Read the report
Open Retention in a website's sidebar. Pick:
- Who's in a cohort: people who first visited, or who first completed a goal. A goal's first time is the first ever, so someone who signed up last year isn't new this week.
- What counts as coming back: any visit, or completing a goal, like a purchase.
- Daily (the last 14 days), Weekly (8 weeks, from Monday) or Monthly (6 months), in the website's timezone.
- On counts people in each period they came back in. On or after also counts them in the periods before their last return, as people who were still around.
Period 0 is the cohort itself. The All cohorts row weighs each cohort by its size and leaves out a period that's still running until no cohort has finished it. Dashed cells are still running, so their numbers can only go up. The dashboard's filters apply to each person's first visit (or the visit they first completed the goal in), so a channel is Organic Search filter shows how people who first came from search stay.
Make it accurate: identify signed-in users
Without more, a person is one browser: a phone and a laptop are two people, and Safari forgets a browser that hasn't visited for 7 days, so later periods read low. Install the script in your app too, and call identify once someone signs in:
window.supr("identify", { user_id: "usr_123" });Every browser identified with the same user_id is then one person in the report, on any device. The Identified figure shows the share of people it knows this way.
AI agent or LLM? Read this page as Markdown