Market Intelligence for Early-Stage Startups: What to Measure Before Launch
Key takeaways
- Poor product-market fit was cited in 43% of recent VC-backed startup shutdowns analysed by CB Insights, so demand evidence matters more than a polished launch.
- Size the market bottom-up, starting from reachable customers and a realistic price, not from an industry report headline.
- Treat search interest, waitlist sign-ups and smoke tests as signals of intent, and set pass or fail criteria before you run them.
- 20 to 30 well-run customer interviews can surface most needs in a fairly homogeneous segment.
- Test willingness to pay early, then confirm it with a real purchase decision.
Why does market intelligence matter before a startup launches?
Market intelligence matters because most early failures trace back to building something the market does not pull. In a March 2026 analysis of 385 VC-backed shutdowns with identifiable causes, CB Insights found poor product-market fit cited in 43% of cases, while running out of capital (70%) was usually the final symptom rather than the root cause.
The goal of pre-launch measurement is not certainty. It is to replace your riskiest assumptions with evidence: is the problem real, will people pay, and can you reach them at a sensible cost?
How should you size the market bottom-up?
Start from customers you can actually name and reach, multiply by a realistic annual price, and only then scale outwards. Carta describes the bottom-up method as more time-consuming than top-down but notes that it "often provides more accurate results", because it rests on sales data, surveys or competitor pricing rather than broad estimates.
A practical sequence looks like this:
- SOM first. Count the accounts or consumers you could win in 3 to 5 years with your current channels.
- Price per customer. Use competitor price lists, pilot quotes or survey answers, not a hoped-for figure.
- SAM. Extend the count to every customer in segments your model can serve.
- TAM. Add adjacent segments only if the product could serve them without a rebuild.
Carta mentions a rule of thumb that pre-seed and seed investors often look for a TAM above $1B. Treat that as context, not a target to reverse-engineer.
Which demand signals are worth tracking?
The most useful early demand signals are search interest, waitlist conversion and smoke test results, read together rather than alone. Each one shows intent at a different depth, from passive curiosity to a click on a "Buy" button, and none of them proves that people will pay until money changes hands.
Search demand. Google Trends shows relative popularity on a 0 to 100 scale, normalised against all searches in a place and period. It does not show absolute volume, so use it to compare terms and spot direction, and use a keyword tool for volume estimates.
Waitlist conversion. Measure the share of landing page visitors who leave an email, by traffic source.
Smoke tests. Amplitude describes a painted door test as a mock page or feature "with no functionality behind it" used to validate an idea before building it. Track clicks on the main call to action, sign-ups and drop-off between steps. Amplitude gives no universal benchmark and advises deciding upfront what counts as a pass or a fail. It also warns that sign-ups may reflect curiosity rather than purchase intent.
How do you map competitors without wasting weeks?
Map competitors by the job the customer is hiring a solution for, not by product category. List direct rivals, indirect substitutes and the "do nothing" option, then compare them on price, target segment, positioning and the complaints their users repeat in public reviews and forums.
A one-page map is usually enough. The useful output is a short list of gaps: underserved segments, uncovered price points, or recurring frustrations your product could remove. Finding no competitors at all is a reason to check whether anyone pays to solve the problem.
How do you test willingness to pay before launch?
Test willingness to pay with structured pricing questions first, then confirm with a real buying decision. The Van Westendorp method, explained by Qualtrics, asks 4 questions: at what price the product seems so cheap that quality is doubtful, a bargain, getting expensive, and too expensive.
Plotting the answers gives an optimal price point where the "too cheap" and "too expensive" curves cross. Qualtrics calls this a starting point to validate in a real purchase situation, so follow it with pre-orders, deposits or paid pilots.
How many customer interviews do you need?
For a fairly homogeneous segment, 20 to 30 interviews is a defensible target. In their study on the voice of the customer, Griffin and Hauser of MIT concluded that "interviews with 20-30 customers should identify 90% or more of the customer needs." Keep going if new needs are still appearing.
Ask about past behaviour, not opinions. "When did this last happen?" produces evidence; "Would you use this?" mostly produces politeness.
Pre-launch metrics at a glance
| Metric | What it tells you | Sourced threshold or guidance |
|---|---|---|
| Bottom-up SOM and TAM | Whether a reachable, paying market exists | Rule of thumb for pre-seed and seed: TAM above $1B (Carta) |
| Search interest trend | Direction and relative popularity of the problem | No threshold; values are relative on a 0 to 100 scale (Google) |
| Waitlist conversion by source | Which channels attract real intent | No universal benchmark; set it before the test |
| Smoke test click-through | Intent to buy or download | Define pass or fail upfront (Amplitude) |
| Van Westendorp price range | Acceptable price band and starting price | Validate in a real purchase situation (Qualtrics) |
| Customer interviews | Needs, pains and current workarounds | 20 to 30 interviews per segment (Griffin and Hauser) |
| Sean Ellis survey (after a beta) | Early product-market fit | 40% or more "very disappointed" (Sean Ellis, via Learning Loop) |
The Sean Ellis survey belongs to the first beta, and 40% is a guideline rather than a guarantee.
Where does a venture builder fit in this process?
A venture builder can run these measurements as a repeatable system instead of a one-off exercise. Studios such as GRS Ventures, which works with founders from idea validation through go-to-market, combine market sizing, demand tests and interviews into a single evidence base that informs both product decisions and fundraising conversations.
FAQ
What is market intelligence for startups? It is structured evidence about customers, competitors, demand and pricing, used to decide what to build and for whom.
What should I measure first before launching a startup? Start with the riskiest assumption. For most teams that is whether the problem is painful enough, so customer interviews and a simple smoke test usually come before detailed market sizing.
Is a waitlist enough to prove demand? No. A waitlist shows interest, and Amplitude warns that sign-ups can reflect curiosity. Pre-orders, deposits or paid pilots are stronger evidence.
Why is bottom-up market sizing better than top-down? It is built from reachable customers and real prices, so it is easier to defend. Carta notes it often gives more accurate results, although it takes longer.
When should I run the Sean Ellis product-market fit survey? After users have experienced the core product, ideally during a beta, with people who have used it at least twice and recently.
Sources
- CB Insights, "Why startups fail: Top 9 reasons" (March 2026): https://www.cbinsights.com/research/report/startup-failure-reasons-top/
- Carta, "Market size: TAM, SAM & SOM": https://carta.com/blog/market-size/
- Google, "FAQ about Google Trends data": https://support.google.com/trends/answer/4365533?hl=en
- Amplitude, "Understanding painted door tests": https://amplitude.com/explore/experiment/painted-door-testing
- Qualtrics, "Van Westendorp pricing explained": https://www.qualtrics.com/articles/strategy-research/van-westendorp-pricing-studies-explained/
- Griffin, A. and Hauser, J. R., "The Voice of the Customer", MIT Sloan working paper: https://dspace.mit.edu/bitstream/handle/1721.1/2425/SWP-3449-27000178.pdf
- Learning Loop, "Product-market fit survey (Sean Ellis test)": https://learningloop.io/plays/product-market-fit-survey
Related reading: Venture Builder Equity: What Studios Take and What You Get.