Rules
Part of Mapping alternatives is real market research, done before any interview
Reading a market research case study critically, one overturned assumption at a time
Startup market research case studies for 2027 use constructed scenarios showing how geography, authority, supply, labor, and economics can overturn demand claims.
What to take away
- Read a case for the decision structure it records, not for the headline result.
- Check what was measured, what contradicted the founders, and when the research stopped.
- A constructed teaching case says so. A real case names the company, the sample and the period.
- Demand means payment, procurement, delivery or repeat use. Interest is not demand.
- Copy the stop rule, then rebuild every assumption for your own market.
The four cases below are constructed teaching examples. Treat them as templates for a reading habit, not as evidence about any real startup.
A five-step reading procedure
- Classify the case.Real, hypothetical, sponsored and independently evaluated cases carry different weight. A sponsored case answers the sponsor's question.
- Write down the decision it served.Note the boundary, sample, period and method before you read one conclusion.
- Hunt the contradiction.Ask which results the founders had to explain away, and which groups were never interviewed.
- Test demand against behavior.Payment, procurement, delivery and repeated use are what a pilot can show. Enthusiasm shows none of them.
- Copy the stop rule.Record what would have ended the research, then set your own threshold before spending more.
Critical case review
- Is the case real or hypothetical?
- What decision, boundary, sample applied?
- Which evidence contradicted the founders?
- Did action require payment or repeated use?
- Which assumptions differ from our market?
Mapping the alternatives your buyer already uses is real market research, and the mapping alternatives walkthrough shows why that belongs at step one.
What the four constructed cases overturn
Each case teaches one move. None reports a real company's results.
Local service. A founder finds a geography with the right households and few listed competitors. Interviews confirm strong interest, and the travel and staffing math later shows the service window cannot be met at the posted price.
The team narrows the radius and tests a denser adjacent area. The grounding cut is a Census BDS establishment count for the state and metro by firm age, which shows whether local supply is thin or merely unlisted.
B2B workflow. Employees praise a prototype and describe painful manual work. Procurement and finance interviews then find no budget owner and no approved data route.
The startup stops treating enthusiasm as demand and researches an existing compliance purchase instead. The check is a BLS BED cut of gross job gains by establishment age in the buyer's industry.
Marketplace. Buyer interviews and search interest suggest real demand. Provider observation shows availability at the promised hour and price is thin, so the team pauses buyer acquisition and recruits one narrow provider category.
The decision measure becomes completed matches and repeat use, not waitlist size.
Subscription. A landing page buys cheap signups. Cancellation interviews show the offer set the wrong expectation, so the team tests clearer terms with a smaller qualified audience.
Signup rate falls, retained contribution rises, and the smaller segment wins.
Public data plus customer evidence is the pairing the worked examples of that mix set out case by case.
Comparing the four cases
| Constructed case | Evidence that contradicted the claim | Decision taken |
|---|---|---|
| Local service | Travel time breaks unit economics | Narrow the geography |
| B2B workflow | No budget owner, no data route | Chase a funded trigger |
| Marketplace | Supply cannot fulfill at the promised hour | Pause buyer spend |
| Subscription | Cheap signups churn quickly | Qualify the offer |
Read each row as a decision map. The last column holds an action, not a metric.
Public datasets that frame or challenge a claim
The Census Bureau's Business Dynamics Statistics reports establishments, firm startups and shutdowns, job creation and destruction, and firm characteristics by age and size.
Use it to test whether a market is forming new firms. It cannot tell you that one product will sell. Both series are published as aggregate counts, so they describe the market rather than your buyer.
The Bureau of Labor Statistics' Business Employment Dynamics derives gross job gains and losses from establishment records and publishes cuts by size, age, industry and state. It can challenge an average-market story and cannot prove a purchase.
Pick the geography, industry, period and unit before you pull a series. A reproducible public data file is what lets a colleague re-run your cut and see the same result.
Audit the published evidence
The FTC's advertising substantiation policy requires a reasonable basis before an objective claim is disseminated. Hold any case study performance figure you plan to quote to that standard.
Check who funded the case, whether the sample and period are disclosed, and whether results appear as ranges or single points. A sponsored ranking with no method section fails that test.
Review integrity counts too. Fake reviews, selective requests, conditioned incentives and hidden relationships distort the acquisition numbers a case reports.
For interview-based research, consent and data handling belong in the audit. Confirm what you may collect, and for how long, before you reuse another team's script.
