How to measure usage with activation, retention and a north star metric.

Once real users arrive, opinions about the product become far less useful than evidence of how it is used. A small, well-chosen set of usage measures tells you whether users get value, whether they come back and which changes help.

Why it matters

Sign-ups and page views feel encouraging but say little about value. Activation and retention show whether the product actually does its job. They are also what serious customers, partners and investors will ask about.

Step by step

1. Choose a north star metric. A single measure that reflects customers getting value, such as “documents collected per week” or “inspections completed per week”. It should rise when customers are succeeding. See north star metric.
2. Define activation. The moment a new user first gets real value, for example “sent first reminder batch and received at least one upload within seven days”. Look at early users who stayed and find the action they had in common.
3. Define an active user. Decide what counts as active for your product and how often it should happen. Weekly is right for some products, monthly for others. Match it to how often the problem occurs.
4. Instrument events. Track a short list of meaningful events: signed up, completed set-up, completed the core job, invited a colleague, upgraded. Name them consistently. Use an analytics tool rather than building your own.
5. Run cohort analysis. Group users by the week or month they started and track what share are still active over time. Cohort analysis shows whether newer cohorts retain better than older ones, which tells you if the product is improving.
6. Review weekly. Put the north star, activation rate and latest cohort retention on one page and review them in every weekly demo.
7. Respect privacy. Tell users what you track in your privacy notice, get consent for non-essential cookies, and avoid collecting personal data you do not need.

Reading a retention curve

Plot the share of a cohort still active in each week after sign-up. If the curve keeps falling towards zero, users are not getting lasting value. If it flattens, a core group has found the product useful; your job is to raise where it flattens and reach that point faster. As a rule of thumb, a flattening curve matters more at this stage than any growth figure.

Worked example

PlanPantry tracked sign-ups closely but found it could not explain why revenue was flat. The team defined activation as “cooked at least two planned meals in the first week”, added five events and built weekly cohorts. Only 22 per cent of new users activated, but those who did were still active after eight weeks at more than three times the rate of those who did not. The team focused the next six weekly cycles on getting new users to their first cooked meal faster. Activation rose to 38 per cent and the retention curve flattened at a higher level.

Common mistakes

• Vanity metrics such as total sign-ups or downloads.
• Tracking everything and acting on nothing.
• No definition of active, so numbers change depending on who pulls them.
• Looking at averages instead of cohorts, which hides whether things are getting better.
• Instrumenting late, losing the early data you most need.
• Changing definitions quietly. If you redefine activation, note the date so trends remain comparable.

What to measure

• North star metric, weekly.
• Activation rate for each weekly cohort.
• Cohort retention at weeks one, four and eight, or months one, three and six.
• Time from sign-up to activation.
• Share of active users completing the core job.

Keep the list short. If a measure has not changed a decision in the last month, consider dropping it from the weekly page and checking it monthly instead.