Market research before launching: a go/no-go checklist

Pre-launch research often fails in one of two ways. It becomes a giant market-size presentation built on assumptions, or it becomes a handful of encouraging conversations with friends. Neither answers the practical question: should we invest the next block of time and money in this idea?

A go/no-go review should be proportionate to the next bet. Before a two-week prototype, a few days of research may be enough. Before inventory, a lease, or a major build, the evidence bar should be higher. The checklist below is designed to reveal what you know, what you are only assuming, and which cheap test should come next.

Before research: define the decision and stop rules

Write the decision in concrete terms: “By Friday, decide whether to spend up to $3,000 building a pilot for independent accounting firms.” Name the customer, the offer, the amount at risk, and the decision date. “Validate the idea” is too vague.

Then pre-commit to evidence that would change your mind. A stop rule might be: “Pause if fewer than three of ten relevant buyers describe this problem without prompting,” or “Do not build if every reachable customer requires a certification we cannot obtain.” The exact threshold depends on the risk and context; the important part is setting it before enthusiasm moves the goalposts.

The six-part go/no-go checklist

1. Customer: can you identify a narrow first buyer?

• Can you describe the buyer by an observable situation, not just age, industry, or company size?
• Who experiences the problem, who approves the purchase, and who can block it?
• Can you find and contact at least a small set of these people without relying on mass advertising?

“Small businesses” is not a usable first market. “Bookkeeping firms with five to twenty staff that still collect client documents by email” is researchable. A narrow entry point does not limit the eventual market; it makes the first learning cycle possible.

2. Problem: is there a costly or urgent job?

• What happens immediately before the problem appears?
• What does the customer do today, including doing nothing?
• What does the current workaround cost in money, time, delay, risk, or frustration?
• Has the buyer already tried to fix it?

Ask for a recent story: “Tell me about the last time this happened.” Avoid pitching during the interview. Questions such as “Would you use an app that…?” invite politeness and imagination. Past behaviour, current spend, and actual workarounds are stronger signals than a hypothetical yes.

3. Trigger: why would someone act now?

Many real problems remain unpurchased because there is no buying event. Look for a deadline, regulation, growth stage, staff change, failed audit, new contract, seasonal peak, or other event that turns inconvenience into action. Record how often the trigger occurs and whether you can observe it from outside.

If the problem is persistent but never urgent, your offer may need to attach to a stronger event. For example, “organise your operations” is abstract; “be ready for your first enterprise security review” has a moment and a consequence.

4. Alternatives: what already gets the budget?

• List direct competitors, indirect substitutes, internal labour, and the status quo.
• Record public pricing and terms exactly; label anything estimated.
• Compare the outcome, setup burden, proof, and risk—not just features.
• Identify where customers praise or reject the alternatives.

Competition is evidence that buyers may spend, but it does not prove room for your offer. No visible competition can mean an open gap, a hidden service market, or insufficient demand. Use a structured competitor teardown to distinguish those possibilities.

5. Access: can you reach buyers at a workable cost?

Name the first channel, not a list of every possible channel. Can you reach buyers through a professional community, targeted outbound, partnerships, a relevant search query, a marketplace, or a physical location? Run a small test: ten carefully researched messages, a partner conversation, a search landing page, or a manual sales attempt.

Track the full path: reached, replied, qualified, met, and committed. A large audience is irrelevant if you cannot get a credible offer in front of the right people. For outbound, start by building a qualified lead list around a visible trigger instead of buying a broad database.

6. Economics and delivery: can this work without heroic assumptions?

• What price range is consistent with alternatives and the value of the problem?
• What does it cost to deliver one unit, including founder or staff time?
• How many customers can you serve with the initial process?
• Which legal, data, licensing, support, or operational constraint could stop delivery?
• What must be true for a customer to buy again, expand, or refer?

Use ranges rather than a single polished forecast. Create a conservative, plausible, and optimistic case, then expose each assumption. Early projections are a map of uncertainty, not a prediction.

Score the evidence without hiding uncertainty

For each of the six areas, mark the evidence red, amber, or green. Red means a critical assumption is contradicted or untested. Amber means you have a plausible signal but need a defined test. Green means multiple relevant sources or a real commitment support the claim.

A “go” should not mean “build the whole thing.” It should approve the next bounded test: a concierge version, paid pilot, preorder, prototype, or manual service. A “conditional go” identifies one assumption that must be resolved first. A “no-go” stops this version of the idea; it does not mean the underlying problem is worthless. Record why, so the team does not rerun the same unsupported concept six months later.

A one-page decision brief