How to measure retention when you only have thirty users
Measure early retention with a handful of users by counting individuals rather than computing rates: list every user, mark whether they performed the core action in each of the last four weeks, and count how many did it twice unprompted. With under fifty users, percentages are noise and names are signal. If fewer than a quarter of your users return unprompted in week two, no amount of acquisition will fix it — you would be pouring people into a product that does not hold them.
The workflow
Not "logged in" — the action that delivers the value. Sent the report, logged the workout, completed a reconciliation. Logins measure curiosity and habit-formation research consistently shows the meaningful unit is the value moment, not the visit. Write the action down as something you could see in a database with one query.
A spreadsheet: one row per user, one column per week, a mark where they performed the core action. Thirty rows fits on a screen and shows you more than any dashboard would at this size. Cohort charts need volume to be meaningful, and a 33% retention rate calculated on three people is a decorative number.
Exclude anyone you nudged, onboarded personally or reminded. What remains — people who came back because they wanted to — is your real retention. Seven of thirty returning unprompted is a genuinely promising early product; two of thirty means the product is not yet solving a recurring problem, whatever the total signup number says.
With this few users you can contact all of them. Ask what they used it instead of, and what would make them stop. Returners are the only population who can tell you what the product actually is, and at this scale you get qualitative depth that no analytics tool would give you at ten thousand users.
- Input
- 31 signups, 19 performed the core action once, 6 returned in week two unprompted, 4 still active in week four.
- What comes back
- CONTINUE, narrowly. 6 of 19 activated users returning unprompted (32%) is a workable early signal, and 4 still active at week four means the value is real for someone. The activation gap is the bigger problem: 12 of 31 never reached the core action at all, which is a first-run problem, not a demand one.
Run it yourself — free, no signup:
Who it's for, what it does, what they pay. The more specific the sentence, the sharper the read.
Two different failures look identical in a signup number. If most users never perform the core action once, you have an activation problem — the first run is too hard — which is fixable with a smaller first step. If they do it once and never return, the product does not solve a recurring problem, which is far more serious and no onboarding change will address it.
Acquisition spend on a product that does not retain is money converted into churn. Get the small number returning first, even manually, then scale. Founders reverse this order constantly because acquisition feels like progress, and it produces a large number of users who tried something once.
- Input
- Week 8: 31 users, 4 weekly active, no growth in 3 weeks, founder does manual onboarding for everyone
- What comes back
- PIVOT the first run, do not kill. Four durable weekly actives out of 31 is thin but real, and manual onboarding for everyone is masking whether the product can activate anyone alone. Cut the first run to a single step and rerun with 20 new users before deciding.
Run it yourself — free, no signup:
1.Has anyone paid you actual money for this?
2.When you stop pushing for a week, what happens?
3.Of the people who tried it, how many still use it a month later?
4.Do you have one channel that reliably brings strangers?
5.Honestly — do you still want to work on this?
6.In the last month, did you learn something that changed the plan?
Answer all six for a verdict. None of them ask what you've already spent — that's the point.
Questions founders ask about this
- How do you measure retention with very few users?
- Count individuals, not rates. A spreadsheet with one row per user and one column per week shows more at this size than any cohort chart, because percentages computed on thirty people are mostly noise.
- What's a good week-two retention rate for an early product?
- As a working guide, around a quarter or more of activated users returning unprompted is promising at this stage. The important qualifier is unprompted — returns you personally triggered do not count.
- What's the difference between activation and retention?
- Activation is reaching the core action once; retention is coming back to do it again. They look the same in a signup total and need completely different fixes, so measure them separately from the start.
- Should I do manual onboarding for early users?
- Yes for learning, but track those users separately. Manual onboarding hides whether the product can activate anyone without you, which is exactly what you need to know before spending on acquisition.
- When should I start spending on user acquisition?
- After a small group retains. Acquisition on a product that does not hold people converts money into churn, and the leak gets more expensive with every user you add.
Next, founders usually do this
- How to validate an app idea without writing code (or hiring a developer)6 steps
- How to get the first 100 users for a free tool nobody has heard of6 steps
- How to know when to kill a side project (and how to do it cleanly)6 steps
The tools used above have their own pages — Idea Validator and Kill-or-Continue Quiz — and the SOP SOP: Validate a SaaS idea in 7 days runs the same ground in more depth. Also worth reading: the Nine Lives Doctrine, and real verdicts from ideas kitty has run this workflow on.
kitty.build runs this entire workflow for you
Every step above — the research, the competitor read, the numbers, the honest verdict — is what nine specialist AI boards do automatically when you feed her an idea. She will tell you to kill it if it deserves killing. First idea is free.
Feed her an idea — freeNo card · Failed tasks are free · Your repo, your domain, your revenue