Examples are illustrative unless otherwise attributed.
Define activation as customer progress
A signup is an observable event, but activation should represent a useful result. Define it for a specific role and task. “Created a project” may be too weak if the customer came to send an estimate. “Sent a valid estimate to a real customer” may be closer to value, subject to permission and the product’s actual purpose.
Specify eligibility and the time window. A user who lacks necessary source data may require a different onboarding route from one ready to start. Do not use a single completion percentage to conceal those differences.
Replace emptiness with an honest model of success
A new product can demonstrate usefulness before it has user-generated activity. Provide labelled sample data, a completed example, a guided first task, and an easy transition into the user’s own work. Explain which actions are real and which belong to a demo. Give users a way to remove examples without losing their own records.
For a marketplace or directory, concentrate on one sufficiently useful segment. Verify any seeded public information and respect source rights. Fabricated inventory, fake reviews, or fictional users presented as real create a false impression of product liquidity.
Better evidence makes the next decision more useful.
Build a cohort with a shared need
Recruit a small set of people whose situations are similar enough to compare. Record why they are eligible, how they were recruited, prior familiarity, and what they were promised. Be explicit about pilot price, support, product limitations, and the feedback arrangement.
Warm contacts are a legitimate starting point. Their relationship with you is also a variable. A subsequent cohort from a different route helps distinguish willingness to help the founder from willingness to adopt the product.
Use manual service to expose the missing system
A concierge approach can help a customer reach value while you observe the work required. Track the manual steps and disclose them. If every useful outcome depends on an hour of founder intervention, the product has demonstrated a service possibility, not yet a scalable workflow.
Review friction by frequency and consequence. A confusing label matters less than an import that loses required data. Prioritise reliability and the core outcome before surface polish or peripheral automation.
Track adoption and durability separately
Use a cohort record with invitations, starts, first-value completions, assistance level, time to value, return behaviour, and eventual payment or cancellation. Define return at the task’s natural interval. Weekly planning, annual compliance, and daily messaging do not share one sensible retention cadence.
Keep qualitative evidence alongside counts. A person who returns because the tool is useful differs from one who returns only because you reminded them repeatedly. Note the intervention rather than attributing every return to the product alone.
Set the next investment decision
Before expanding, decide which uncertainty remains: customer relevance, onboarding, repeat value, willingness to pay, or delivery economics. Design the next cohort to address it. You may reduce assistance, test a clearer price, or recruit through a different relevant channel.
Do not turn five successful assisted sessions into a claim of product-market fit. They are evidence about five sessions. Preserve the conditions that made them work and test whether those conditions can be reproduced at an acceptable cost.
Continue the traffic and growth learning path
Sources & further reading
Google Analytics: Traffic acquisition report ↗Background references are distinguished from our original examples and proposed exercises.
