What is a retention curve?

What is retention analysis?

Retention analysis measures how many users keep returning to a product over time, and the retention curve is its main chart. You take a group of users who started in the same period, decide which action counts as "active", and plot the share of that group still doing it at each interval after their start.

Three choices decide whether the curve means anything:

• The starting event. Sign-up, first login or first completed setup. Pick the one that marks a real start.
• The return event. An action tied to value, such as sending an invoice or publishing a report. A login alone tells you little.
• The interval. Match it to how often the job needs doing. A daily tool gets daily or weekly buckets; a payroll tool gets monthly ones.

Retention analysis usually puts several curves side by side, one per sign-up month or acquisition channel. That comparison is how you tell whether a change actually helped.

How do you read a retention curve?

Read it left to right in three parts: the initial drop, the slope after it, and where it levels off.

Say 1,000 people sign up for an invoicing tool in March, and "active" means sending at least one invoice that week. In week one, 450 send one. By week four it's 300, by week eight 260 and by week twelve 250.

• The initial drop (from all 1,000 down to 450) is the people who didn't get value in their first week. It is mostly an onboarding problem.
• The slope from week one to week eight shows people trying the product and deciding. A steep slope here means the product isn't becoming part of their routine.
• The plateau at about 250 users, 25% of the cohort, is your retained core. Whether the line goes flat at all matters more than the level.

A curve that keeps sliding toward zero means no group of users has found lasting value yet. More sign-ups won't fix that; the product has to change.

What does a healthy SaaS retention curve look like?

A healthy curve drops early, then flattens and stays flat. Some products even curve up slightly later, as lapsed users return or teams widen their use.

There is no universal "good" level. Where a curve settles depends on the segment you sell to, the contract (monthly self-serve plans behave differently from annual contracts), how often the job comes up and how strictly you define "active". So compare your own cohorts against each other: each new monthly curve should sit above the last one, or flatten sooner.

How do you flatten a retention curve?

Work on the part of the curve that's moving.

• The initial drop: shorten the path to the first result. Cut setup steps, prefill what you can and ask for less before the user sees value.
• The slope: find what retained users did in their first weeks that others didn't, and get more new users to do it early.
• The plateau: watch for slow decline among long-time users, often a sign that their workflow changed and the product didn't keep up.

The biggest gains usually sit in the first sessions, where a new user either gets something done or leaves. An in-product agent helps there: the user asks, and the agent does the step with them instead of sending them through screens they've never seen. Foldspace includes product analytics (retention, funnels and session replay) that can compare retention across segments, and its Product Agent does the step in the product, so a team can check whether the curve for users who worked with the agent flattens sooner.