Stop building things nobody wants. Here's a framework that actually works.
Most products fail before they find users. According to CB Insights' post-mortem analysis of startup failures, 42% of failed startups cite "no market need" as the primary cause — more than running out of money, more than team problems. The product worked. Nobody wanted it.
The reason is almost always the same. Founders validate by asking their network. Their network is polite, enthusiastic, and demographically narrow. The network confirms the idea is interesting. The founder builds for six months. Then launches to silence.
The fix is not to validate harder — it's to validate differently. Specifically: stop asking "is this a good idea?" and start asking "which specific groups of people have this problem badly enough to pay for a solution?"
The standard advice is to talk to users. It's good advice, but it fails in practice for two reasons.
First: founders interview their networks. If you're a software engineer in San Francisco, your network skews toward software engineers, startup people, and tech-adjacent professionals. They share your zip code, your income, your worldview. They're also your friends, which means they're motivated to be supportive rather than honest.
The result is a sample of one demographic slice of humanity expressing polite enthusiasm. That's not validation. That's confirmation bias with extra steps.
Second: even honest interviews capture what people say, not what they do. "Would you use this?" is a different question from "Do you currently spend $15/month on something that imperfectly solves this?" Polite interest is not a purchase signal. The people who have the problem acutely already have workarounds — and they'll tell you those workarounds in concrete, specific terms. People without the problem will speak in generalities.
A nurse's aide in rural Tennessee, a shift worker at a distribution center in Ohio, a freelancer managing three project streams with irregular income — these people don't appear in the average founder's interview sample. But they might be exactly who needs the product. Building without finding them means building for the wrong segment.
Before you ask "will people pay for this?" you need to ask "which specific people have this problem acutely enough that they'd pay for a better solution?"
That's problem-solution fit (PSF). It's not a binary. It's a score from 0 to 100 that reflects three factors evaluated together:
All three dimensions need to be high for a strong PSF score. A segment that recognizes a problem but already solves it adequately scores low — not because the problem doesn't exist, but because the solution gap is small. There's no reason to switch.
High-PSF segments (70+) have the problem acutely and no adequate existing solution. They're already spending money, time, or emotional energy dealing with it. They have workarounds that frustrate them. Those are the people who will pay, refer others, and give you honest product feedback.
Not: "Does this problem exist?" — it almost always does, for someone. But: "Which specific groups have it severely enough, and which ones have already solved it well enough that targeting them is wasted effort?"
"Is my idea good?" is the wrong question.
The right question is: "Which specific groups of people experience this problem acutely, and which ones don't — and why?"
Modest Idea evaluates product concepts across 250 synthetic personas drawn from Census PUMS (Public Use Microdata Sample) data. These personas span age, income, education, occupation, geography, family structure, and personality. They're statistically weighted to reflect the actual US adult population — not overrepresented with tech workers or underrepresented with service-sector workers.
The output is PSF scores for 6–8 distinct demographic segments with the reasoning behind each score. Not just who scores high, but why the problem is acute for some groups and irrelevant for others.
The critical insight: PSF is always segmented. The same product can be genuinely needed by one demographic group and irrelevant to another. Building without knowing which group is which means building for everyone — and ending up serving no one well. Read more about the methodology on the methodology page.
A habit accountability app sounds universally useful. Everyone struggles with habits sometimes. The founders expected their target audience to be young urban professionals — the same people who follow productivity content and read self-improvement books.
The segment scores were counterintuitive.
Office workers have an accountability structure they don't notice. The commute, the 9am start, the colleague who asks about the gym. Their schedule anchors habits involuntarily. They don't feel they need a habit app because their work schedule provides accountability for free — it just doesn't feel like accountability because it's invisible.
Shift workers have none of that. A rotating 3-on/4-off schedule shares no anchors with anyone in their social circle. Days start at different times every week. There's no colleague to ask about the gym because their shift ends at 11pm. Their problem is structural, chronic, and acute. The office worker's problem is mild and already handled.
The product roadmap implications are significant. A habit app for shift workers needs asynchronous accountability check-ins, partner matching that prioritizes schedule overlap, and notifications timed to rotating shift patterns — not a social feed optimized for morning routines. See the full habit accountability app analysis for the complete segment breakdown.
A marketplace connecting clients with independent contractors — where freelancers post profiles and clients post projects — showed an equally counterintuitive geographic split.
Urban freelancers have a problem the rural ones don't: too many options. Upwork, Fiverr, local referral networks, coworking community Slack channels. They can find work. Their market is liquid. A new platform is marginally useful — another tab to check.
Rural freelancers have the inverse problem. Their local market is thin. The three dentists in the county already have a web designer. Remote-first work culture has helped, but discoverability across geographic distance remains a genuine friction point. A centralized marketplace solves a real access problem they don't have other ways to solve. Their workaround is cold outreach — expensive in time and with low conversion.
The go-to-market strategy changes completely. Urban-targeted paid acquisition is expensive and reaches the low-PSF segment. Outreach to rural freelancer communities, state-level chambers of commerce, and rural co-op networks reaches the segment that actually needs the product. See related findings in the freelancer cashflow planner analysis.
Four common pre-build validation approaches, compared honestly:
Fast and easy. Your network is not your market. Respondents share your worldview and are motivated to be supportive. Useful as a starting point, not sufficient as the primary signal. The demographic homogeneity of your network will systematically under-sample the segments that might be your best customers.
More honest than surveying friends. Requires deliberate effort to recruit across demographic groups — which most founders skip. Done well (10–20 interviews across diverse segments, focused on behavior not hypotheticals), it's the most reliable signal. Done poorly (interviewing whoever responds to your tweet), it replicates the network bias problem.
Conversion data doesn't lie — clicking is more honest than nodding. A $500 campaign to a waitlist page will tell you more than 50 interview responses. The problem: you need the landing page first, you need to run multiple campaigns to compare segments, and $500 per segment test adds up quickly. Best used after you've narrowed your target audience.
Evaluates your concept against hundreds of demographically diverse personas simultaneously. Covers demographic breadth that would take months to replicate via interviews. Results in minutes. The caveat is real: synthetic personas are statistical models, not people. The PSF scores are directional, not predictive. Use them to identify which real segments to investigate, not to skip investigation entirely.
The most effective sequence: use synthetic persona analysis to identify high-PSF segments, then conduct 8–12 interviews focused on those segments. The analysis tells you where to look; the interviews tell you what's actually there.
Paste a 2–3 sentence description of your product — what it does and what problem it solves. The Pitch Doctor extracts a problem statement and solution interpretation. You confirm or edit them before the analysis runs.
The analysis evaluates your concept against 250 Census-grounded personas using multiple AI models at varied temperatures — designed to reduce the systematic bias that comes from using a single model at a single temperature. The personas span demographics that rarely appear in founder networks: service-sector workers, rural residents, variable-income earners, older adults, single parents.
The output is segment-level PSF scores with the reasoning behind each score. You see not just which segments score high, but why — which factors create the acute need, which factors explain the low-scoring segments, and what the mechanism is.
Use those scores to decide where to spend your interview time. Don't interview everyone. Interview the two or three highest-scoring segments. Then interview the lowest-scoring segment and ask: "Why don't you need this?" Their answer will sharpen your understanding of exactly what creates the acute need in the high-scoring groups — and that understanding will shape your positioning, copy, and feature priorities.
Confusing medium PSF scores for good news. If every segment scores 45–55 with no standout high-scoring group, the product either needs to be repositioned or the problem genuinely isn't acute for any segment. A uniform medium score is a warning, not a green light.
Synthetic personas are statistical models, not people. They predict behavior from demographic and situational proxies — occupation, income, schedule type, family structure, personality traits. Real people are messier. A shift worker with exceptional self-discipline won't be captured by the aggregate. A suburban commuter who works fully remote has a different problem profile than the standard segment description. The personas represent distributions, not individuals.
Modest Idea is a pre-build tool. It surfaces which segments are worth investigating before you commit to building — it's not post-launch proof of product-market fit. If your shift worker segment scores 84, go talk to shift workers. Not two of them — ten. See whether the reasons they give match what the analysis predicted. Where they diverge significantly, the analysis was wrong about this specific product in ways that are worth understanding before you finalize your build scope. Use this to point your research, then do the research.
Get PSF scores for your product concept across 250 Census-grounded personas and 6–8 demographic segments. Free to run.
Try Modest Idea →Upload your product idea and get Problem-Solution Fit scores across 8 demographic segments — backed by 250 Census-grounded personas.
Try Modest Idea →