A software idea usually arrives as a solution: a dashboard, an automation, a marketplace, an AI assistant. That is useful creative energy, but it is not yet an opportunity. An opportunity is a problem, a reachable group of people and a credible reason they might change what they do today.
Separate the idea from the opportunity
An idea describes something you could make. An opportunity describes a situation in which making it might create enough value to justify the effort. The distinction matters because many products can be built correctly and still fail to become important to anyone.
Before choosing features, write down four things in plain language:
- Who experiences the problem? Use a bounded group you can identify and reach.
- What happens today? Describe the workflow, workaround or cost without referring to your proposed product.
- Why does it matter? Look for frequency, delay, money, risk or frustration—not merely stated interest.
- What must change? State the outcome a customer would recognize, even if your preferred implementation disappeared.
If these statements remain vague, additional code will usually make the uncertainty more expensive rather than resolve it.
Turn confidence into an assumption map
Founders often debate whether an idea is “good.” A more productive question is: what must be true for this direction to work? The answer turns confidence into a set of claims that can be challenged independently.
Clinic managers lose meaningful time to a fragmented intake workflow; they control or influence a purchase; existing tools leave an important gap; a safer workflow can be delivered at an acceptable cost; and the team can reach enough suitable clinics.
Some claims concern the customer, others the market, distribution, delivery or your own constraints. Do not average them into one optimistic score too early. One weak, essential assumption can matter more than several encouraging but optional signals.
Distinguish evidence from interpretation
A useful evidence record says what was observed, where it came from and what it does—or does not—support. “Customers want this” is an interpretation. “Four of five practice managers showed the same manual reconciliation step, and three described a weekly delay” is evidence that can be reviewed.
Strong early evidence does not have to be large. It should be relevant to the claim:
- Use workflow observation to understand behaviour and workarounds.
- Use interviews to learn language, context and constraints—not to count compliments as demand.
- Use a concrete offer to test commitment.
- Use delivery to test whether the promised outcome can actually be produced.
- Use repeated acquisition attempts to test whether one encouraging result can become a channel.
Retain contradictions. A customer who declines for a specific reason may reveal the boundary of the segment more clearly than a friendly conversation.
Choose the smallest test that can change the decision
The smallest product is not always the smallest useful test. If your main uncertainty is willingness to pay, a carefully scoped offer may teach more than a polished prototype. If the uncertainty is technical feasibility, a thin implementation may be exactly right. If the uncertainty is access to customers, code may teach almost nothing.
Define the test before running it:
- the claim being tested;
- the method and participant boundary;
- a success criterion;
- the time or cost limit;
- what you will do after a positive, negative or ambiguous result.
This prevents an experiment from becoming an open-ended project and makes it harder to reinterpret every outcome as encouragement.
Finish with an explicit decision
Research has little value if it only accumulates. At a review point, record what you know, what remains uncertain and whether the next direction is to pursue, hold, reject or gather specific missing evidence.
The purpose is not to make decisions permanent. It is to preserve the reasoning that existed at the time. When new evidence arrives, you can compare it with the earlier assumptions instead of reconstructing the past from memory.
Good opportunity assessment does not eliminate risk. It makes the risk visible enough to choose the next commitment deliberately.