There is a phrase founders use that quietly worries me every time I hear it: “We just need to find our product-market fit.”
The worry is not with the goal. Product-market fit is genuinely the most important milestone a startup can reach — the point at which you have found a group of customers who need what you are building badly enough, and in sufficient numbers, that growth becomes something the market pulls rather than something you push. Everything before product-market fit is preparation. Everything after it is execution. Getting from one to the other is the central challenge of the startup phase.
The problem is the word “find,” which implies that product-market fit is something you stumble across, or discover, as if it were waiting for you in the right market if you looked hard enough. In my experience working with more than 2,000 businesses over 25 years, the founders who reach product-market fit reliably are not the ones who found it. They are the ones who built a systematic process for understanding the gap between where they were and where they needed to be, and then iterated towards it with discipline.
This guide covers that process in full — what product-market fit actually means in measurable terms, the specific tests and signals you should be tracking, the qualitative evidence that quantitative metrics alone will miss, and the iteration framework that moves you towards it when you are not there yet.
What Product-Market Fit Actually Means
The concept was introduced into the startup vocabulary by Marc Andreessen, who described it simply as being in a good market with a product that can satisfy that market [1]. That definition is correct but not operationally useful — it tells you what the destination looks like without helping you know whether you have arrived.
The more useful operational definition comes from the experience of what it feels like on both sides of the threshold. Before product-market fit, growth requires constant effort — you push customers into the funnel through marketing and sales activity, some of them convert, many of them churn, and the business grows slowly if at all because the rate of acquisition barely exceeds the rate of loss. After product-market fit, the dynamic shifts. Customers refer others without being prompted. Churn falls because people who have found something that genuinely solves a real problem do not voluntarily give it up. Marketing starts to work at lower cost because you have a clear story about a real need, told in the language of the people who have that need.
The transition between these two states is not always dramatic. Sometimes it happens gradually, and founders only recognise in retrospect that something changed around a specific month. But the underlying mechanism is always the same: the product has become genuinely indispensable to a clearly defined group of people, and that indispensability is what drives sustainable growth.
What product-market fit is not, importantly, growth. Founders sometimes mistake early rapid growth for product-market fit, particularly when that growth is driven by a one-off catalyst — a press feature, a viral social post, a successful launch event. Growth that depends on a catalyst is not the same as growth that is structurally embedded in the product’s relationship with its market. The test is what happens when the catalyst is removed: does growth continue, and does the customer base stay? If yes, you may have product-market fit. If the numbers revert to their previous trajectory, you have had a spike, not a shift.
The Three Dimensions You Need to Measure
Product-market fit is not a binary state you either have or do not have — it exists on a spectrum, and it has three distinct dimensions that are all necessary and none of which is individually sufficient. Missing any one of them means the fit is incomplete, regardless of what the others show.
The first dimension is demand depth. Does the problem you are solving matter enough to the people who have it that they will seek out and pay for a solution? A product can have a broad reach—many people are aware of it—while demand is shallow, meaning most of those people would not particularly miss it if it disappeared. Shallow demand leads to poor retention, high churn, and price sensitivity because the customer does not feel they need the product enough to tolerate inconvenience or cost. Deep demand produces the opposite: customers who complain when things break, who tell others about the product without being asked, and who resist switching even when competitors offer cheaper alternatives.
The second dimension is segment specificity. Product-market fit is never universal — it exists for a specific type of customer in a specific context. A product that has a strong fit with one customer segment almost certainly has a weak fit with another, and treating those segments as equivalent produces a muddled message and a diluted product. Identifying not just that product-market fit exists, but precisely which segment it applies for, is critical because that segment definition is the foundation of every subsequent marketing, sales, and product decision.
The third dimension is economic viability. A product that customers love but will not pay enough for, or that can only be delivered profitably at a scale the market cannot sustain, has, at best, partial product-market fit. The economic dimension is where SGI’s Business Success Formula is most directly relevant: Appeal, Profitability, and Sustainability are the three interconnected characteristics every successful business must have. A product can score well on Appeal — customers genuinely want it — but score poorly on Profitability because the price customers will pay does not cover the cost of delivery. That is not full product-market fit. It is a good idea, but with a broken business model, it requires a different kind of work to fix.
The Sean Ellis Test: Your Quantitative Starting Point
The most widely used quantitative test for product-market fit was developed by Sean Ellis, the growth marketer who helped scale Dropbox, Eventbrite, and LogMeIn in their early stages. The test is elegantly simple: survey customers who have used your product at least twice — people with enough experience to have formed a genuine view — and ask them a single question.
“How would you feel if you could no longer use [product]?”
The response options are: Very disappointed; Somewhat disappointed; Not disappointed (it really isn’t that important); I no longer use [product].
Ellis’s research across hundreds of startups found that if 40% or more of respondents say they would be very disappointed, the product has reached the product-market fit threshold. Below 40%, the fit is insufficient to support sustainable growth, regardless of other metrics [2].
How to Run It Properly
The test sounds simple, and it is — but it is undermined by common implementation errors that produce misleading results.
The sample matters enormously. You need to survey customers who have used the product enough to have a genuine opinion — typically at least two to three uses, or in a SaaS context at least two weeks of active use. Surveying new sign-ups who have barely engaged gives you acquisition sentiment, not retention sentiment; they’re different. The 40% benchmark was established against engaged users.
The timing matters. Running the survey immediately after a positive experience — right after a successful onboarding call, for example — introduces recency bias, inflating the “very disappointed” responses. Run it as a routine part of your engagement cadence rather than as a response to a positive moment.
Sample size matters. A 40% result from eight respondents tells you almost nothing. Aim for a minimum of forty to fifty responses before drawing conclusions. The survey can be distributed incrementally, but do not interpret early results until the sample size is sufficient to be statistically meaningful.
The follow-up questions matter as much as the headline result. Alongside the core question, ask: “What type of person do you think would benefit most from [product]?” and “What is the main benefit you receive from [product]?” and “How can we improve [product] for you?” These open-ended responses are where the product-market fit iteration work lives — they tell you which customer type is driving the positive responses, what specific value is resonating, and what the product still needs to do better.
Interpreting the Results
A result of 40% or higher for “very disappointed” is positive, but the work does not stop there. The next question is who is generating that result. If your overall score is 42% but it is driven entirely by one customer segment that represents 20% of your current user base, that tells you something very specific about where your product-market fit actually exists and where your growth efforts should focus.
A result below 40% is not a verdict on the idea—it is a diagnostic. The most useful analysis is to read each response from the “very disappointed” group carefully and compare it with those from the “somewhat disappointed” and “not disappointed” groups. The differences in how they describe the product’s value, what they use it for, and what they wish it did better will frequently reveal what the product needs to change to move more respondents into the “very disappointed” category.
I worked through this exercise with a Cardiff-based B2B SaaS founder whose product management tool had been live for eight months. The Ellis survey came back at 31% — below the threshold but not disastrously so. The insight that changed everything came from the open-ended responses: every single “very disappointed” respondent described the product’s value in terms of a specific collaboration feature, while respondents in the lower categories primarily described it in terms of the task management functionality that the founder had built first and considered the core product. The product had a strong product-market fit with teams that needed better collaboration workflows. It had a weak fit with project managers who primarily needed task tracking. The founder had been marketing to the latter and underinvesting in the former. That single analysis shaped the product roadmap, messaging, and customer acquisition focus for the following 12 months.
Retention Cohorts: The Most Honest Metric You Have
The Sean Ellis test measures how customers feel about your product. Retention cohorts measure what customers actually do — and as I have said in the context of business validation, behaviour is a more honest signal than stated intent.
A retention cohort is a group of customers who all started using your product during the same defined period — a month, a quarter — and is tracked over time to see what proportion remain active at each subsequent interval. When you plot these cohorts on a chart with time on the horizontal axis and retention rate on the vertical axis, the shape of the resulting curve tells you more about the health of your product-market fit than almost any other single metric.
What the Curve Tells You
A retention curve that continues to decline towards zero — where the proportion of active customers from each cohort approaches nothing over time — is the clearest possible signal that product-market fit does not yet exist. It means customers are trying the product and leaving. The specific shape of the decline gives you additional information: a steep initial drop followed by gradual flattening suggests the product has problems with onboarding or early-stage value delivery, while a slow progressive decline suggests a problem with long-term value delivery or competitive alternatives.
A retention curve that flattens and stabilises — where the proportion of active customers plateaus at some level above zero and stays there — is the strongest quantitative indicator of product-market fit. It means there is a core group of customers for whom the product has become genuinely sticky, and they are not leaving even as time passes and alternatives exist. The absolute level at which it flattens matters less than the fact that it flattens. A curve that stabilises at 25% is far more positive than one that is still declining at 45%.
For subscription businesses, the relevant metric is the monthly or annual subscription renewal rate. For transactional businesses, it is the repeat-purchase rate and the interpurchase interval. For platforms or marketplaces, it is the proportion of users who complete a meaningful action in each subsequent period after sign-up. The specific metric varies, but the underlying question is always the same: when the immediate novelty of your product has worn off, and normal life reasserts itself, do customers choose to keep using it?
Benchmarks by Business Type
Retention benchmarks vary significantly by product category, and comparing your numbers against the wrong benchmark produces misleading conclusions. For B2B SaaS products, annual retention rates above 80% are generally considered healthy, and rates above 90% are considered strong. For consumer subscription products, monthly churn below 5%—equivalent to roughly 46% annual retention—is the common baseline for early-stage businesses with product-market fit. For mobile apps with a transactional model, day-30 retention rates above 10% from first download are considered positive; above 20% is strong [3].
These benchmarks are starting points, not verdicts. The most useful benchmark is always your own trend line — whether your retention is improving or deteriorating over successive cohorts is more informative than where any single cohort sits relative to an industry average.
Qualitative Signals: What the Numbers Cannot Tell You
Quantitative metrics — the Ellis score, retention cohorts, churn rate, NPS — tell you whether product-market fit exists. Qualitative signals tell you why, and they are essential to the iteration work that moves you from insufficient fit to strong fit.
The Language Customers Use Unprompted
One of the clearest qualitative signals of genuine product-market fit is when customers start describing your product in ways you did not teach them. When the language they use to explain what your product does — to colleagues, to people they are referring, in reviews or social posts — is specific, vivid, and outcomes-focused, that is a strong indicator that the product has created a genuinely distinct experience in their perception.
When customers struggle to articulate what the product does or describe it in generic terms that could apply to any competitor, the fit is probably weak. The product has not made a strong enough impression to create a distinctive narrative in the customer’s mind. That is not a marketing problem — it is a product problem.
Unsolicited Referrals
The single most powerful qualitative signal of product-market fit is customers referring others without being asked and without a formal incentive. Referral programmes and commission structures can generate referrals without product-market fit — those are incentivised behaviours, not evidence of genuine conviction. The signal I look for is the email a founder receives from a new customer that starts “I heard about you from…” where the customer did not receive anything for the recommendation and the referrer did not tell the founder they were doing it.
Tracking the proportion of new customers who came through unprompted referrals, and whether that proportion is growing or shrinking over time, is one of the most reliable leading indicators of product-market fit development available.
Customer Behaviour When Things Go Wrong
How customers respond to product failures, price increases, or temporary deteriorations in service quality is a revealing indicator of the depth of their product-market fit. Customers who have a genuine, deep fit with a product complain loudly when things go wrong — not because they want to leave, but because they care enough about the product to want it fixed. They engage with your team on the problem. They wait for the resolution rather than immediately switching.
Customers who do not have genuine fit respond to problems by quietly leaving, because the product was not indispensable enough to make the friction of a complaint worthwhile. If your customer service data shows that the most engaged, highest-value customers are also your most vocal complainers, that is a signal worth recognising for what it is: evidence of the depth of their dependence on your product.
The Iteration Framework: Moving Towards Product-Market Fit
Understanding where you currently are relative to product-market fit is the diagnostic. The iteration framework is the treatment — the structured process of moving from insufficient fit to strong fit through a disciplined cycle of hypothesis, change, and measurement.
Step 1: Identify Your Current Best-Fit Segment
Before you try to improve anything, you need to understand who in your current customer base has the best fit with the product as it currently exists. This is the starting point for all iteration work, and the Ellis survey follow-up questions I described above are the primary tool for identifying it.
Look at the “very disappointed” group and identify what they have in common that distinguishes them from the rest. Industry sector, company size, role, use case, the specific problem they were trying to solve when they found your product, the specific feature they rely on most heavily. The more precisely you can characterise this group, the more useful the characterisation becomes as a guide to product and marketing decisions.
The question to ask is not “how do we make everyone love our product?” but “how do we make more people like the people who already love our product?” That framing focuses the iteration work on extending the existing fit rather than attempting to build fit with segments the product is not naturally suited to.
Step 2: Understand What Prevents Others From Having the Same Fit
The gap between your “very disappointed” group and your “not disappointed” group almost always has a specific explanation. The product does something for one group that it does not do for the other — or it does it in a way that works for one and does not work for the other. Understanding this gap is the core of the iteration work.
The most reliable way to understand it is to talk to both groups directly. Interview five to ten people from each group and compare their answers to the same set of questions: What were you trying to accomplish when you started using this? What does the product do for you that other solutions do not? What is the most frustrating thing about it? What would need to change for you to rely on it more heavily? The contrast between the two groups will identify the specific changes — to the product, to the onboarding experience, to the messaging, to the customer support model — that would most improve fit across the broader customer base.
Step 3: Make One Change at a Time
The most common iteration mistake is changing multiple things simultaneously and then being unable to determine which change produced which result. Product-market fit iteration requires the same discipline as any other experiment: change one variable, measure the effect, learn from it, and then change the next variable.
This is harder than it sounds when the pressure to grow is high and the gap between current fit and target fit is large. The instinct is to accelerate by making several changes in parallel. That instinct produces faster activity and slower learning, which is the opposite of what the iteration process requires. Disciplined single-variable iteration is slower in the short term and faster in the long term.
Step 4: Measure the Right Leading Indicators
Full retention cohort data takes months to accumulate to statistical significance. For the week-to-week and month-to-month iteration work, you need leading indicators that move faster and still tell you something meaningful about the direction of travel.
For most businesses, the leading indicators that are most sensitive to product-market fit changes are: early engagement depth (how many of the features or actions that “very disappointed” users rely on are new users reaching within their first two weeks?); early churn signals (what is the proportion of users who disengage within the first thirty days, and is that proportion falling?); and the proportion of new customers coming through unsolicited referrals.
These three metrics, tracked weekly, will show you the directional effect of product changes faster than retention cohorts can, and they will tell you whether you are moving towards or away from the characteristics that distinguish your best-fit customers.
Step 5: Resist the Urge to Scale Before Fit Is Confirmed
This is the iteration principle that founders most frequently violate, and it is the one whose violation is most expensive. Scaling customer acquisition before product-market fit is confirmed does not accelerate reaching fit — it obscures the signal by introducing noise, depletes the capital needed to continue iterating, and creates operational complexity that slows the iteration cycle.
The resources spent acquiring customers who will churn because the fit is insufficient are resources that could have been spent deepening the product’s fit with the customers who are already retained. The most common pattern I see in startups that run out of capital before reaching product-market fit is not bad ideas or bad execution — it is premature scaling, where growth investment was deployed before the product had earned the right to grow.
The test before scaling is simple: is your retention curve flat at a meaningful level? If yes, you have the foundation to grow on. If it is still declining, growth investment will accelerate burn, not product-market fit.
What Product-Market Fit Looks Like at Different Business Stages
Product-market fit is not a single milestone — it evolves as the business grows, and what constitutes strong fit at seed stage looks different from what is required at Series A and beyond.
At early stage, with a small number of customers, product-market fit is primarily evidenced by intense qualitative signals from a tight segment: customers who refer others without prompting, who complain when things break rather than leaving, who describe the product in vivid and specific terms. The Ellis score threshold still applies, but the sample sizes are too small for cohort analysis to be statistically meaningful. At this stage, qualitative depth of fit is the primary signal.
At growth stage, with enough customers to generate statistically meaningful cohort data, the quantitative metrics become the primary measurement tool. The question shifts from “does strong fit exist in our early adopter community?” to “does strong fit extend to the broader market segment we are trying to reach?” Retention cohorts across multiple growth cohorts should be stable or improving; any deterioration as the customer base broadens is a signal that the product is being sold to a segment with weaker fit than the early adopters.
At scale, product-market fit becomes a maintenance challenge as much as an achievement. Markets evolve, competitive dynamics shift, customer expectations rise. Businesses that maintain strong product-market fit at scale do so by treating it as an ongoing measurement and iteration discipline — not a problem that was solved once and can now be safely ignored.
A Case Study in Product-Market Fit Iteration: Planetary Processing
Planetary Processing, a Cambridge-based game development infrastructure company we worked with on their funding strategy, offers a useful illustration of the product-market fit iteration process applied in a technical B2B context. The founding team had built a multiplayer game server management platform initially targeted broadly at game developers needing scalable server infrastructure.
Early Ellis survey results were below the threshold, and cohort analysis showed progressive churn even among developers who had integrated the platform technically. The qualitative analysis — interviews with their highest-engagement users — revealed something the broad targeting had obscured: the product had exceptional fit with a very specific sub-segment of the market, specifically independent studio developers building real-time multiplayer games with rapidly fluctuating concurrent user loads, where the platform’s auto-scaling capabilities delivered distinct, measurable commercial value.
For studios in adjacent categories — single-player games, turn-based multiplayer, studios with stable user bases — the value proposition was weaker and the switching cost from existing solutions was not justified. The iteration work that followed focused on deepening the product’s capabilities in the areas most valued by the core segment, repositioning the messaging around the specific pain of unpredictable concurrent load management, and tightening the customer acquisition focus to reach studio developers with that specific profile.
The result was a measurably stronger Ellis score among users acquired after the repositioning, a flatter retention curve in subsequent cohorts, and a product-market fit story strong enough to support the VC funding round the company subsequently secured.
Frequently Asked Questions
How long does it typically take to achieve product-market fit?
There is no standard timeline, and anyone who gives you one is working from anecdote rather than evidence. The honest range, based on the startups I have worked with and the broader research on early-stage company development, is six months to three years from first launch. The primary determinants of speed are how clearly the initial target segment was defined, how quickly the iteration cycle runs, and how honestly the team reads the evidence the metrics produce. Teams that run a disciplined weekly iteration cycle and are genuinely willing to change direction based on data consistently reach fit faster than teams that move slowly or that reinterpret negative signals as positive. Pre-launch validation of the kind I described in the business idea validation guide also meaningfully reduces the time to fit, because it eliminates the weakest assumptions before they are built into the product.
Can a business have product-market fit with one segment but not another?
Yes, and this is not an unusual situation — it is the norm at early stage. Almost every product that eventually achieves broad market success started with very strong fit in a small, specific segment and expanded from there. The strategic question is whether the segment with strong fit is large enough to build a commercially viable business within, or whether it is a bridgehead — a base of strong fit that can be extended to adjacent segments over time. The mistake is treating the absence of fit in some segments as evidence that the fit in other segments is not real. Find where the fit exists, build on it, and expand deliberately rather than trying to be relevant to everyone simultaneously.
Is NPS a reliable measure of product-market fit?
Net Promoter Score — the “how likely are you to recommend us?” metric — is a useful supplementary signal but a poor primary measure of product-market fit. The correlation between NPS and actual retention and referral behaviour is weaker than its widespread use implies, partly because the question asks about hypothetical future behaviour rather than measuring current behaviour, and partly because high NPS can coexist with significant churn if customers like the product in principle but do not rely on it in practice. Use NPS alongside retention data and the Ellis survey — not as a substitute for them.
What should I do if my Ellis score is below 40% and my retention is poor?
Run the qualitative analysis first. Review every response from your “very disappointed” group and compare it to your “not disappointed” group. The insight about where the fit exists — which sub-segment, which use case, which problem framing — is almost always in that comparison. Then make one specific change based on that insight, measure the effect on your leading indicators over four to eight weeks, and iterate. If after three to four iteration cycles there is no measurable improvement in any of the signals, that is the moment to conduct a broader strategic review — which may include a customer pivot or a product pivot of the kind I described in the piece on pivoting versus persevering.
Does product-market fit look different for service businesses versus product businesses?
The underlying dynamic is the same — you are looking for a group of customers who find your offering genuinely indispensable — but the measurement tools adapt. For service businesses, the Ellis survey is directly applicable. Retention measurement becomes repeat engagement rate and average client tenure rather than software usage data. The referral signal is arguably even more important in service businesses, where the referral rate is both a measure of client satisfaction and a primary growth mechanism. One important difference: for service businesses, the “very disappointed” threshold can be meaningful at lower absolute customer numbers, because service relationships typically have higher depth and richer qualitative signals than product usage data.
How do I talk about product-market fit with investors?
Investors, particularly at seed and Series A stage, will expect you to have a clear, evidence-based answer to this question. The framing they find most credible is not “we believe we have product-market fit because our customers like us” — it is specific metrics, honestly presented, with an honest assessment of what they mean. If your Ellis score is 38%, do not round it up to “nearly at the threshold” — explain what you know about the segment that is driving the strong responses and what specific changes you are making to improve the overall score. Investors fund startups before product-market fit is fully established all the time; what they are assessing is whether the founder understands the metrics clearly, is iterating deliberately, and has a credible hypothesis about the path to strong fit.
References
- Andreessen, M., “Product/Market Fit”, Stanford University lecture notes, reproduced at pmarchive.com, 2007
- Ellis, S. and Brown, M., “Hacking Growth”, Crown Business, 2017 — original source for the 40% benchmark and survey methodology
- Andreessen Horowitz, “16 Startup Metrics”, 2015, https://a16z.com/2015/08/21/16-metrics/ — widely cited benchmarks for SaaS and consumer app retention
- Rachleff, A., “Andy Rachleff’s Law of Startup Success”, Wealthfront blog, 2011 — influential early framing of product-market fit as a prerequisite for growth investment
- Startup Genome, “Global Startup Ecosystem Report”, 2023, https://startupgenome.com/reports/gser2023 — data on failure rates and causes, including premature scaling as a primary failure mode
- Reichheld, F., “The One Number You Need to Grow”, Harvard Business Review, 2003, https://hbr.org/2003/12/the-one-number-you-need-to-grow — original NPS research, with the limitations of the metric noted in subsequent critiques
If you are at the stage where your startup has initial customers but unclear product-market fit, and you want an objective outside assessment of where the fit exists and what the iteration priorities should be, our startup consultants work with founders on exactly this diagnostic. We have run the Ellis survey and retention cohort analysis across dozens of early-stage businesses and know what the results typically mean in practice, not just in theory.
If you are preparing for an investment conversation and need a clear, credible product-market fit narrative supported by data, our business plan writers can help you structure and present that evidence in a way investors find compelling. And if you are at an earlier stage and want to maximise your chances of reaching product-market fit faster, our business consultants can build the iteration framework into your operating model from the start.
Related Posts
Kurt Graver is the founder and CEO of SGI Consultants, a business consultancy that has helped over 2,000 entrepreneurs establish successful startups using systematic business development methodologies. An accountant with an MBA and 25 years of commerce and consultancy experience, Kurt specialises in strategic planning, market analysis, and sustainable business growth