Signups and page views feel good, but they do not tell you whether new customers get value. The accounts that stay are the ones that reach a meaningful result early, and the ones that leave usually showed warning signs in their first days.

The metrics below are widely used to spot those signs. Which ones matter most depends on your product, so the last section shows how to find your own.

1. Time to first value

Time to first value (TTFV) is how long it takes a new user to reach the moment your product first proves its worth. For an invoicing tool that might be sending a first invoice. For an analytics tool it might be seeing real data in a dashboard.

  • Define the moment in plain words, then measure the median time from signup to that moment.
  • Look at where people stall. Every extra step, form field or setup task before the first value is a place to lose them.

2. Activation rate

Activation rate is the share of new signups who complete the key action or actions within a set period, such as their first seven days. It is more useful than raw signups because it counts people who actually started using the product.

The important part is how you define "activated". Do not guess. Compare customers who stayed with customers who left, and look for the early actions that separate them.

3. Team adoption

In many B2B products, an account used by several people is harder to lose than one used by a single champion. It is worth tracking whether new accounts invite teammates, and whether those teammates become active. Test whether this holds in your data before you rely on it.

4. Depth of use, not just breadth

A user who tries ten features once is often less engaged than one who uses a core feature again and again. Track how often new users repeat your core action in the first couple of weeks, alongside how many features they touch.

5. Early retention signals

Simple return behaviour is a strong early signal: did the user come back in week one, and did they complete setup? Look at these by cohort, which means grouping customers by signup week so you can see whether changes to onboarding are helping.

How to find your own predictors

  1. Group customers into cohorts by signup date and by plan or customer type.
  2. Compare retained and churned customers. What did the retained group do in their first week that the others did not?
  3. Pick one to three metrics with the clearest difference, and set a threshold such as "completes the key action within three days".
  4. Act on the signal. Trigger an in-app prompt, an email or a call from your team when a new account misses the threshold.
  5. Re-check regularly. Products and customers change, so review your chosen metrics every quarter.

Pitfalls to avoid

  • Mistaking correlation for cause. A behaviour that goes with retention may not cause it. Run an experiment before rebuilding onboarding around it.
  • Small samples. With few customers, patterns can be noise. Be cautious about drawing firm conclusions.
  • Hiding differences in averages. Segment by plan, company size and use case before you decide something works.