What to do after your survey closes
Product Market Fit Analysis
A 42% blended PMF score can hide a 65% score in your ICP and a 20% score everywhere else.
Here is how to segment, interpret, and act on your PMF data so the analysis drives real product decisions - not just a number in a slide deck.
Measurement vs Analysis
Most teams stop at the score. The score alone does not tell you what to build next.
PMF Measurement
Getting the number
- -Sending the Sean Ellis survey to active users
- -Calculating % who answer 'very disappointed'
- -Comparing against the 40% benchmark
- -Tracking score month over month
PMF Analysis
Understanding what the number means
- +Segmenting responses by user type and cohort
- +Reading qualitative answers from 'very disappointed' users
- +Identifying your champion segment (where fit is strongest)
- +Turning segment gaps into a prioritized product roadmap
Before you analyze: validate your sample
Bad input data produces misleading conclusions. Check these four things before drawing any conclusions from your PMF data.
40+ responses
Below 40 qualified responses, your PMF score is directional only. It can track trend direction but is not statistically reliable for product decisions.
Used core feature 2x+
Survey only users who have experienced your core value - not everyone who signed up. One-time visitors inflate 'not disappointed' and distort your score downward.
Active in last 2 weeks
Survey users who have engaged recently. Dormant users who were once active will give you historical data about why they left, not current product-market fit signals.
40+ per segment
Segment analysis requires 40+ responses per segment to be reliable. A 55% score from 8 enterprise users is not the same as a 55% score from 80 enterprise users.
Run the PMF survey first, then do the analysis
Mapster segments your PMF responses automatically by user attributes so you can find your champion segment without a spreadsheet.
Run Your Free PMF SurveyNo credit card required
The PMF analysis framework
Five steps from raw survey data to actionable product decisions.
Calculate the blended score
Start with the overall number: (very disappointed responses / total responses) x 100. This is your baseline. Do not make decisions from this number alone - it is the starting point for the analysis, not the conclusion.
Segment by user type
Split responses by your most important dimensions: ICP vs non-ICP, company size, use case, and acquisition channel. Calculate a PMF score for each group. The segment scores will almost always differ significantly from each other and from the blended score.
Read the qualitative answers
For every 'very disappointed' response, read what they said is the main benefit they receive. Group these into 3-5 theme buckets. The most common theme is your core value proposition. If you do not know what to say in your marketing, this answer tells you.
Find your champion segment
Identify the user type, company size, or use case where the PMF score is highest. This is your champion segment. Read the qualitative answers only from this group. Their answers about benefits and desired improvements define your roadmap priorities.
Map 'somewhat disappointed' gaps
Read what 'somewhat disappointed' users say they are missing or what they wish the product did better. These are the specific features or improvements that would move them to 'very disappointed'. Prioritize these over feature requests from users outside your champion segment.
How to interpret your PMF score
The number tells you where you are. The analysis tells you what to do about it.
Focus on distribution
You have proved fit in at least one segment. The product question is largely answered. Shift focus to GTM: scale the acquisition channels that bring your champion segment, systematize onboarding, and expand to adjacent segments - re-measuring PMF in each new segment before scaling.
Segment and double down
You have fit in a subset of your users but not across the board. Do not average - find the segment scoring above 40% and focus entirely on them. Stop trying to improve fit for everyone at once. The champion segment is your path to 40%+ overall.
Deep qualitative work needed
The product is solving a problem but not solving it well enough to create dependency. Run 5-10 customer discovery interviews with 'somewhat disappointed' users. Ask: what would have to be true for this to be a must-have? The answer is usually a missing core feature or the wrong target user.
Revisit problem definition
At below 10%, the issue is usually not features - it is problem definition or target market. Either the problem is not painful enough, users do not experience it frequently enough, or you are talking to the wrong people. Consider a broader pivot before adding more product surface area.
Four dimensions to segment your PMF data
Blended PMF scores mislead. These four segmentations reveal where your fit actually lives.
By customer type
Example dimensions
- -SMB vs mid-market vs enterprise
- -Founders vs product managers vs engineers
- -B2B vs B2C users
Different customer types almost always have different PMF scores. The segment scoring highest is your ICP. If enterprise scores 15% and SMB scores 55%, you have clear direction even if the blended score is 35%.
By use case
Example dimensions
- -What core workflow they use the product for
- -Primary job to be done
- -Feature set they rely on most
Your product may serve multiple use cases at different levels of fit. Find the use case where users score highest and double down on it - even if it means deprioritizing other use cases temporarily.
By acquisition channel
Example dimensions
- -Organic search vs paid vs word-of-mouth
- -Content-led vs outbound-led
- -Self-serve vs sales-assisted
Users acquired through different channels often have different levels of intent and fit. Organic and word-of-mouth users typically score higher because they arrived with context. If one channel produces significantly higher PMF scores, that is your best channel.
By user tenure / cohort
Example dimensions
- -Users from 6+ months ago vs recent signups
- -Early adopter cohort vs growth cohort
- -Pre-feature-X vs post-feature-X
Cohort segmentation reveals whether fit is improving or decaying over time. If users from 6 months ago score higher than recent signups, you may have expanded to a less-fit segment. If recent cohorts score higher, your iterations are working.
Reading the qualitative data
The score tells you how much fit you have. The open-text answers tell you what kind of fit, where it comes from, and what is missing.
"What is the main benefit you receive from this product?"
Ask: Very disappointed users only
Group answers into 3-5 themes. The most common theme is your real value proposition - not the one on your homepage, but the one users actually experience. If 60% of 'very disappointed' users say 'saves me time on X', that is your headline.
"What type of person do you think would benefit most from this?"
Ask: Very disappointed users only
Read how your best users describe the ideal user. They often describe themselves - and in much sharper language than your internal ICP definition. Use their exact words to rewrite your ICP, targeting criteria, and ad copy.
"How could we improve this product for you?"
Ask: Very disappointed users - what more do they want?
These are your highest-priority features. 'Very disappointed' users want to stay and want more. Their requested improvements are the roadmap items most likely to increase PMF score further and drive expansion revenue.
"What would you use instead if this product no longer existed?"
Ask: Somewhat disappointed users - what are they comparing you to?
If 'somewhat disappointed' users say your competitor, read what they say the competitor does better. That gap is exactly what you need to close to move them from 'somewhat' to 'very disappointed'. This is competitive analysis from your own data.
From analysis to roadmap
PMF analysis answers four product questions that no other data source can answer as directly.
What is our core value?
Most common benefit cited by 'very disappointed' users
The thing users say they get most from the product - in their words. This is what to protect at all costs and what to build on. Never remove or dilute it for the sake of expansion.
Who should we focus on?
The segment with the highest PMF score
Your champion segment. Build the roadmap for them first. Resist the pull to improve fit for everyone simultaneously - it dilutes focus and rarely works. Find fit in one segment, then expand.
What should we build next?
Improvement requests from 'very disappointed' users
Features requested by users who are already at 'very disappointed' are your highest-leverage investments. They extend and deepen fit in your best segment. Prioritize these above all other feature requests.
How do we improve the overall score?
What 'somewhat disappointed' users say is missing
The specific gap that would tip 'somewhat' users to 'very disappointed' - usually one or two features or use-case improvements. Close this gap and re-survey. Each iteration should move the score higher.
Common PMF analysis mistakes
Surveying all users, not active ones
Including users who signed up but never experienced core value inflates 'not disappointed' and artificially depresses your score. Filter for active users only.
Drawing conclusions from fewer than 40 responses
Below 40 qualified responses, small fluctuations in behavior can swing the score 10+ percentage points. Use it as directional data only.
Only reporting the blended score
The blended score hides where fit is strong and where it is weak. A 32% blended score with a 58% ICP segment score is a very different situation than a 32% score across all segments.
Ignoring the qualitative answers
The open-text responses are the most valuable part of the survey. The score tells you how much fit you have. The text tells you what kind, where it comes from, and what is missing.
Re-surveying too soon after changes
Wait 4-6 weeks after significant product changes before re-surveying. Users need time to experience new features in their workflow before their 'very disappointed' threshold changes.
Treating PMF analysis as a one-time event
PMF can decay. Market conditions shift, you expand to new segments, competitors copy your core value. Run the survey quarterly and treat analysis as an ongoing practice, not a milestone.
Frequently asked questions
Get the PMF data you need to analyze
Mapster segments your PMF responses automatically by user attributes so you can find your champion segment without a spreadsheet.
Run Your Free PMF SurveyNo credit card required