The seven SaaS onboarding metrics worth tracking (time to first value, activation rate, setup completion, time to activation, early active use, breadth of adoption, 90-day cohort retention), which two predict churn and which five explain it, why setup completion is not activation, and why the numbers only mean anything by cohort.
Short answer
Which onboarding metrics actually predict retention?
Two predict it: time to first value and activation rate. Five explain it: setup completion, time to activation, first-week and first-month active use, breadth of adoption, and the 90-day retention of the cohort. An account that has not reached its activation event by day fourteen is the earliest reliable churn signal you have. Report all seven by acquisition cohort, never as a blended month.
Most early churn is not a decision. It is a customer who paid, tried, and never got to the moment they were paying for. Onboarding metrics measure how many customers reach that moment and how fast, which makes them the churn prediction system for new accounts. These are the seven worth tracking, the definition of each, the relationship between them, and the reporting mistake that hides the signal.
Every metric below depends on one decision: what counts as a customer getting value. It is a product-specific event, and it is almost never “logged in.” For an analytics tool it is the first report shared with someone else. For a billing tool it is the first invoice sent. For a customer success platform it is the first account scored from the customer's own data. Pick the event that, once it has happened, makes the customer measurably more likely to stay, and check that claim against your own retention data before you build a dashboard on it.
| Metric | Definition | Tells you |
|---|---|---|
| Time to first value | Median days from first payment to the activation event | How long a new customer waits for a reason to stay |
| Activation rate | Accounts reaching the activation event within N days ÷ accounts in the cohort | How many customers ever get the reason |
| Setup completion rate | Accounts that finished every required setup step ÷ cohort | Whether the path to value is being walked at all |
| Time to activation | Median days from setup complete to activation | Whether setup leads anywhere |
| Early active use | Share of the cohort active in week 1 and in month 1 | Whether value became a habit |
| Breadth of adoption | Median number of core features used by day 30 | Whether the account is attached to one feature or the product |
| 90-day cohort retention | Cohort still paying at day 90 ÷ cohort size | Whether any of the above translated into revenue |
Time to first value and activation rate are the ones that move retention. Every SaaS business that has run the comparison finds the same shape: accounts that activate inside the window retain at a multiple of the accounts that do not. The size of the multiple varies; its existence does not.
The other five are diagnostics. When activation drops, setup completion tells you whether customers stopped walking the path or the path stopped leading to value. Time to activation separates a slow setup from a setup that finishes and goes nowhere. Early active use and breadth tell you whether an activated account is building a habit or had one good day. Ninety-day retention closes the loop and is the number that should appear on the same row as the cohort's activation rate, so the relationship is visible rather than assumed.
An account that has not activated by day fourteen is not a customer with a slow start; on most books it is the majority of the first-quarter churn, identified with ten weeks to spare. This is why onboarding metrics belong in the churn conversation and not only in the product one. Usage decline, support sentiment and renewal dates are churn signals for established accounts. For new accounts, the signal is the activation event not having fired, and the response is an onboarding intervention, not a save offer.
The practical rule: any account past day fourteen without activation gets a human, with the specific missing step named. “How is it going?” is not an intervention. “Your Stripe connection is live but no usage source is connected yet, which is why nothing is scoring” is.
The most common onboarding dashboard mistake is treating a finished checklist as a successful onboarding. Setup completion is necessary and not sufficient. A customer can connect every integration and invite every colleague and still never open the one screen that would have shown them why they bought. Track both numbers, and when they diverge, the gap is the product's problem, not the customer's: the setup led somewhere the customer did not recognise as value.
A blended monthly activation rate answers no question anyone is asking. The question is whether the onboarding change shipped in March made the March cohort activate faster and retain better than February's, and only a cohort table can answer it. One row per acquisition week or month, the seven metrics as columns, the 90-day retention last. Segment by acquisition channel and plan: the blended median almost always hides one channel producing most of the never-activated accounts, and that channel is the cheapest retention fix you have.
Time to first value, activation rate, setup completion rate, time to activation, first-week and first-month active usage, breadth of adoption (how many of the core features a new account uses), and 90-day retention of the onboarding cohort. The first two are the ones that predict churn; the rest explain why the first two moved. Track them by acquisition cohort, not as a blended monthly number.
Time to first value (TTFV) is the elapsed time from a customer's first payment (or signup, for self-serve) to the first moment they get the outcome they bought the product for. It has to be defined per product: the first report sent, the first integration synced, the first automation that fired. Measure the median, not the mean, because a few accounts that never reach value pull the mean toward infinity.
Activation rate is the share of new accounts that reach the activation event within a set window, usually seven or thirty days. Benchmarks vary too much by product to be useful as a target; the useful comparison is your own rate by cohort and by acquisition channel. What matters is the relationship: accounts that activate within the window retain at several times the rate of accounts that do not, in almost every SaaS business that has measured it.
Churn in the first ninety days is mostly customers who never reached value, and time to first value and activation rate are direct measures of that. An account that has not hit its activation event by day fourteen is the earliest reliable churn signal a SaaS company has, earlier than usage decline, earlier than support tickets, and months earlier than the renewal date. Onboarding metrics are churn prediction for new accounts.
Setup completion is the customer finishing the steps you asked for: connecting the data source, inviting the team, configuring the settings. Activation is the customer getting an outcome from the product. They diverge constantly: a customer can complete setup and never activate (the integration is connected and nobody looks at the result), or activate without completing setup (one user gets value from one feature and ignores the rest). Track both, and treat setup completion as a leading indicator of activation, not as activation.
By weekly or monthly acquisition cohort, with each metric as a column and the 90-day retention of that cohort as the last column. A blended monthly onboarding dashboard hides the thing you are looking for, which is whether a change to onboarding in March made the March cohort activate faster and retain better than February's. Segment by acquisition channel and by plan, because the medians usually hide one channel producing most of the never-activated accounts.
The metric definitions are standard; no benchmark figures are quoted because onboarding numbers vary too much by product for a cross-industry range to be honest. The day-fourteen rule reflects how Exeechain treats its own new accounts: First scores in 15 minutes. Full accuracy in 24 hours. The activation event is the first account scored from the customer's own data, and an account that has not reached it is flagged on the readiness panel rather than left to the renewal date. The retention side of the cohort table is covered in the SaaS retention curve guide, and the first-30-days playbook in reducing churn in the first 30 days.
Keep reading
Customer success
9 min read · Sep 20, 2026
Retention metrics
8 min read · Sep 20, 2026
Churn prediction
9 min read · Sep 20, 2026
First scores in 15 minutes. Full accuracy in 24 hours. From $299/mo. Connect Stripe, add one snippet, and the readiness panel tells you what is still missing.
Not ready to switch? Size your leak from your MRR and churn rate, with no billing access at all.