Decision Three: Will Customers Pay for the Product?
Task
Run the Stage 3 gate review.
Summary
Decide whether the product has value on its own and can be supported without recreating a custom project.
A working product is not yet a product business
Customers may praise a demonstration, use an early version, or agree to a pilot while the company still performs a custom project behind every sale. The Stage Three decision asks a narrower question: will intended customers pay for the product itself and receive value through a repeatable operating model?
The answer cannot come from enthusiasm alone. It requires evidence from sales, onboarding, product use, support, delivery effort, and customer economics.
Prepare the evidence before the review
Bring one joined view of the intended customer segment and product version. Do not combine unrelated customer types or product configurations merely to increase the sample.
| Dimension | Core question | Useful evidence |
|---|---|---|
| Paid demand | Did intended customers buy the product? | Arm’s-length sales, price, discounts, conversion, and contract terms |
| Offer consistency | Was substantially the same product sold? | Standard scope, configuration, promise, and exceptions |
| Custom work | What newly scoped labour was required? | Custom hours, one-off integrations, founder time, and change requests |
| Onboarding | Can customers start through a repeatable process? | Completion rate, elapsed time, staff hours, and exception rate |
| First value | Do customers reach a meaningful result? | Defined first-value event, time to value, and customer confirmation |
| Product use | Is value delivered through the product? | Activation, key workflow completion, and continued use |
| Support | Is support bounded and predictable? | Tickets, support hours, escalations, and engineering intervention |
| Economics | Can the model work at the intended price? | Product revenue, direct cost, onboarding cost, and contribution margin |
| Continuation | Is the value durable? | Continued use, renewal, expansion, or a clear continuation commitment |
Label missing data, immature customers, exceptional discounts, and founder-assisted results. A small honest sample is more useful than a large blended one.
Separate strong proof from supporting signals
Strong proof is a paid purchase by an intended customer, using a substantially standard offer, followed by repeatable onboarding and value delivered mainly through the product. Continued use or renewal strengthens that evidence.
Supporting signals include trial use, a letter of intent, a design-partner agreement, positive interviews, reference willingness, or a successful demonstration. These signals help explain demand, but they do not prove the complete operating model.
Weak signals include internal enthusiasm, feature requests without purchase commitment, vanity usage, exceptional customers, or a result achieved only through unrecorded senior effort.
Run the decision meeting
Assign one accountable decision owner. Include sales, product, delivery or customer success, support, finance, and the person responsible for the underlying data.
- State the proposition being tested: customer, problem, product, price, and expected operating model.
- Review facts before interpretation.
- Examine exceptions and negative evidence explicitly.
- Compare actual effort and economics with the intended model.
- Record disagreement and uncertainty.
- Decide go, keep testing, narrow the proposition, or stop.
Do not average every measure into one universal score. A regulated enterprise product may require more onboarding than a simple self-service tool. Thresholds must come from the intended price, margin, promise, implementation complexity, and support capacity.
Record a decision that can be reviewed later
The decision record should identify:
- the date and accountable owner;
- the product version, customer segment, use case, and route to market covered;
- the proposition tested and evidence period;
- facts observed, including contrary evidence;
- discounts, custom work, founder involvement, missing data, and other exceptions;
- the interpretation and rationale;
- material dissent;
- conditions, owners, measures, and review dates;
- what investment is authorized or withheld.
A conditional go is not permission to ignore unresolved work. Each condition needs an operating limit. For example, the company may continue serving the current cohort while refusing to increase sales volume until onboarding time or custom effort falls below an agreed threshold.
What must be true before moving on
Proceed to building the repeatable subscription business only when intended customers have paid for a substantially standard product, reached value through a repeatable path, required support and custom work the company can sustain, and produced economics consistent with the intended price. If that evidence does not exist, the correct result is more focused learning—not a larger acquisition budget.