Imagine opening the commercial review with twelve paying customers. That sounds like a useful starting point. Then you look more closely. Three bought a bespoke project. Four arrived through the founder’s former colleagues. Two needed weeks of unpaid implementation. One has already stopped using the product. The remaining two look like the customers you actually want more of.
This is a hypothetical company, but it raises a very practical question. Which of those twelve relationships should determine the next quarter?
I would resist putting them into one growth curve. Each customer has taught the company something. They have not all taught it the same thing, and they may not represent the same business.
Our first two playbooks examined market evidence and the people involved in a buying decision. Once a product is in market, the next job is to understand the evidence already sitting inside the company. A signed contract answers whether someone bought. It takes more work to establish whether you can find another suitable customer, sell a comparable offer and deliver value without rebuilding the process.
Start with the customer rather than the dashboard
There is a reason to do this before increasing acquisition. RevenueCat’s 2026 subscription app report describes a market where nearly 15,000 new subscription apps now launch each month, compared with roughly 2,000 three years earlier. This is evidence about the wider app market, not a count of SportsTech companies. For a consumer product, however, it is a useful reminder that shipping another app does not give you a reliable route to demand. RevenueCat State of Subscription Apps 2026.
For a business selling software to operators, the equivalent question is more local. Can you identify the next twenty organisations with the same buying situation as your strongest customers? If the answer depends on knowing everyone personally, that needs to appear in the growth plan.
Pull your most recent customers, cancellations, lost opportunities and active evaluations into one working file. Ten or twenty records are enough to begin if that is what you have. They are not enough to establish a universal conversion rate. Keep the sample size visible.
For each customer, record what was purchased, who paid, why they acted then, how they found you and how long the decision took. Add the first useful outcome, the work required to produce it and what happened afterwards. An account with six months of use tells you something different from one that signed last Thursday.
Include costs that sales usually leaves out. A padel operator may pay the same monthly fee as another operator while requiring much more support. A training app subscriber may arrive cheaply through a race community but disappear when the event finishes. An elite team may pay well and still require a level of custom reporting that you cannot offer to the next fifty teams.
Those are different growth questions. Revenue alone will not separate them.
Explain the win before you repeat it
Ask the customer to reconstruct the purchase. I would use a short conversation built around these questions:
What was happening when you first started looking for a solution?
What were you using before and why did that stop being sufficient?
What nearly prevented the purchase?
What made the first month useful or difficult?
What would make you continue paying at the next renewal?
These are questions I would use for this review, rather than a published interview script. The principle of learning from past behaviour is central to Michele Hansen’s Deploy Empathy. Her preview also distinguishes interviews about buying, staying and cancelling. Those are separate research jobs, so I would recruit for each rather than interview only happy customers. Deploy Empathy preview.
Keep the founder’s explanation beside the customer’s explanation. If you believe you won because of the prediction model and the buyer says you were the only provider willing to clean their data before the season started, investigate the difference. Perhaps service is an important part of the offer. Perhaps the product is solving the wrong part of the problem. You will not discover that by asking for a testimonial.
A lost deal deserves the same care. Separate a competitor win from a decision to do nothing, a budget deferral and a technical obstacle. They lead to different responses. Another demo will not repair an unavailable integration; a discount may not change a decision that has no internal owner.
Put each relationship in its proper place
I would sort the evidence into four working groups. These are management categories, not a maturity score.
Customers to reproduce. The buying situation is identifiable, the offer is reasonably consistent and the customer reaches useful value at a delivery cost the business can support. Renewal is either observed or still explicitly unknown.
Customers to learn from. There is promising use or payment, but one important part of the process remains uncertain. You may need another buying cycle, a renewal decision or a delivery comparison before committing more resources.
Customers that need redesign. They pay and may be satisfied, but their economics depend on hidden founder work, repeated exceptions or services priced as software. The relationship may still be worth keeping. Its scope and price need attention before it becomes a template.
Relationships to keep outside the growth forecast. An unpaid test, one-off event or warm introduction can be useful. Give it a specific purpose. Do not count it as evidence for a recurring motion until the relevant customer action happens.
The uncomfortable part is that a prestigious account can land in the third group. A smaller customer can land in the first. The exercise is useful precisely because it separates the commercial pattern from how good the logo looks on a slide.
A small example with a different investment decision
Consider a hypothetical company selling booking and retention software to racket sport operators. It has eight single-site customers and four larger groups.
The single-site customers buy a standard package, set up within days and need limited ongoing help. The larger groups pay more, but each has negotiated different reporting, approval and migration requirements. Their contracts look stronger in the revenue report. Their contribution after delivery may be weaker.
There are at least two reasonable strategies. The company can focus on the standard package and improve acquisition among single-site operators. Or it can deliberately build an enterprise offer, price implementation separately and invest in the capabilities larger groups need.
What I would avoid is selling both as one undifferentiated product and asking marketing to produce more leads. That transfers an unresolved business decision into the acquisition budget.
Write the decision in operational terms. Which package gets promoted? Which prospects get qualified out? Who can approve an exception? Which delivery work must be measured? If nothing changes outside the presentation, the audit has not done its job.
Check whether another person could reproduce the result
Take one recent win and give the record to someone who was not involved. Could they explain who should be approached next, what buying trigger to look for, what was promised and which customer action moved the deal forward?
You are not trying to remove the founder from the customer relationship overnight. You are checking how much of the process exists outside the founder’s memory. A warm introduction can remain part of the motion. Its availability, cost and limits simply need to be understood.
For a consumer product, run the same exercise by acquisition cohort. Compare people who arrived through a coach, a training community and search. Look at whether they paid, completed a useful activity and returned when the next relevant activity occurred. Avoid comparing a race cohort after six weeks with a general fitness cohort after six months.
Seasonality matters here. A decline in sessions after the competitive season may be expected. A team failing to use the product during the weeks it was purchased for is a different signal. Define the job and its cadence before interpreting the graph.
The decision I would make this week
Spend a morning rebuilding the evidence behind your last ten customers. Then choose one of four responses: pursue more of the same buying situation, reshape the offer, fix adoption or test a different route to suitable buyers.
Assign one owner and one date to the response. Leave uncertain renewal and acquisition economics marked as uncertain. You can make a useful decision without pretending the company has a statistically stable model.
The next stage of growth should be based on the customer relationships you can support and reproduce. That is a more demanding question than whether the company has traction, and a much more useful one.
I work with SportsTech founders and leadership teams on these commercial decisions, from diagnosing current traction to building a repeatable motion. You can discuss the challenge with me at GTM Sports. Subscribe to Building SportsTech for the next playbook.