A growth funnel can look healthy at registration and still fail to create customer value. More accounts are encouraging, but they do not tell you whether people can complete onboarding, experience the product or return when they need it.
September Series 01 · A practical growth diagnosis framework. All numerical examples in this article are hypothetical, not client results.
Before increasing acquisition spend, I want to understand what happens after someone signs up. A cheaper registration is useful only when the people behind it can reach the outcome the business needs.
Why registrations can hide a broken growth funnel
Registration is an early commitment. It may reflect curiosity, a reward, a persuasive creator video or a real need. Those motives can produce similar signup counts and very different downstream behaviour.
In a remittance app, creating an account is not the same as getting verified, funding a payment or completing a transfer. For a learning platform, it is not the same as attending a lesson. For a consulting website, an account or email subscription is not a qualified enquiry.
The problem begins when the team optimises the easiest visible event and treats the rest of the journey as someone else’s responsibility. Marketing celebrates signups; product sees stalled onboarding; operations sees support requests. Each team has a piece of the explanation.
Define activation before building the report
Choose the earliest observable action that indicates someone experienced meaningful value. Define its status, time window and unit of counting. A completed transfer is more precise than a transfer-button click. A lesson attended is different from a booking started.
For this hypothetical app, activation means one successfully completed first transfer within 14 days of registration. A separate repeat metric counts a second completed transfer within 30 days of the first. These windows are example definitions, not universal industry benchmarks.
Agree on exclusions such as test accounts and duplicate users. Decide how refunds, cancellations and delayed status updates should appear. Keep a written definition so a dashboard change does not silently change the meaning of success.
A simple example: more signups, fewer activated users
| Metric | Cohort A | Cohort B |
|---|---|---|
| Acquisition spend | $2,000 | $3,000 |
| Registrations | 1,000 | 1,500 |
| Activated within 14 days | 200 | 180 |
| Registration-to-activation rate | 20% | 12% |
| Cost per activated user | $10 | $16.67 |
Both cohorts have a $2 cost per registration. Signups rise by 50%, yet activation falls by 10%. If the report ends at registrations, the second cohort looks like a successful scale-up. Once activation is included, the decision changes.
This example does not prove that the campaign attracted worse users. A release, verification delay or funding problem could also explain the result. The numbers locate a question; investigation establishes the cause.
Read the growth funnel as connected transitions
1. Registration to onboarding started
Check whether users understand the next step and can find it. Review the landing-page promise against the first screen. If a campaign implies immediate access but onboarding contains several unexplained stages, the mismatch may discourage a motivated user.
2. Onboarding started to approved
Separate people who never submit from those waiting for a decision or needing resubmission. Those groups need different interventions. My existing KYC drop-off diagnosis guide explains the importance of locating the actual stopping point.
3. Approved to first value
An approved account may still face an unfamiliar funding method, unclear fees or a lack of immediate need. Review the journey described in KYC to first transfer before assuming approval automatically creates activation.
4. First value to repeat use
A first action encouraged by a reward may not become a habit. Check whether customers have a recurring use case and whether the product remains convenient after the incentive ends. Use completed outcomes rather than repeat visits alone.
Compare cohorts with equal time to convert
Do not compare a cohort that has had two weeks to activate with yesterday’s signups. Anchor the analysis to registration date, allow the defined observation window and mark immature cohorts clearly.
Break the result down by acquisition source, market, device and relevant product version where the data supports it. Start with the largest meaningful difference. Avoid dozens of tiny segments that produce unstable percentages.
Track both the transition rate and the number of users lost. A weak step affecting ten people and a moderately weak step affecting thousands may deserve different priorities. Include median time to activation when delays are part of the problem.
Turn a dashboard pattern into one experiment
Write a testable explanation: “New users do not recognise the supported funding method.” Then gather evidence from support themes, consented user research and the actual interface. Check event accuracy before testing a redesign.
A bounded experiment might clarify the funding instruction on one screen. Choose first-transfer completion as the outcome, with support contacts and payment failures as guardrails. Use an appropriate comparison group when feasible, and wait for the activation window before judging the result.
A reminder belongs in the plan only when the user can complete the step and needs a nudge. It cannot repair an unavailable button or a misleading instruction.
A weekly growth review that supports decisions
- Acquisition: spend, registrations and source mix.
- Activation: mature-cohort completion rate, activated users and cost per activated user.
- Friction: the largest stalled group and its documented explanation.
- Next action: one owner, one hypothesis and one review date.
If your team has several dashboards but no shared diagnosis, my Marketing Audit and 1-on-1 Consulting can help connect campaign performance with the onboarding journey and prioritise a useful experiment.
Frequently asked questions
Are registrations a vanity metric?
Not automatically. They measure an early stage. They become misleading when used as evidence of activation or revenue without checking the downstream journey.
What should replace cost per registration?
Keep it for acquisition analysis, then add cost per activated user and later value metrics appropriate to the product.
Should I stop acquisition when activation drops?
Investigate the size, cause and business impact first. A technical failure may justify a pause; an immature cohort may simply need more observation time.