Superhuman Product Market Fit: The PMF Engine Explained

Why Most Founders Think About PMF Wrong

Product-market fit gets thrown around like a magic spell. "We have traction." "Customers love us." "We just need to find our PMF." But none of that is measurable, and none of it tells you what to do next.

The Superhuman product market fit framework flips this. Instead of vibes and vanity metrics, it gives you a number - a score you can actually move. Rahul Vohra, CEO of Superhuman (the email client), turned PMF from a feeling into an engine. He published the methodology publicly and it became one of the most-referenced startup posts in Silicon Valley.

I've built and exited five SaaS companies. The pattern I see most often: founders either launch without testing fit at all, or they collect feedback with no system for acting on it. The Superhuman method solves both problems.

Here's the full breakdown - what the framework is, how to run it, what the High-Expectation Customer concept actually means, how to use PMF as a permanent tracking metric, and the specific moves that actually shift your score.

The Origin Story: Two Years of Coding, No Launch

Before we get into the mechanics, the backstory matters. Rahul Vohra didn't stumble onto this framework by accident. Before Superhuman, he built Rapportive - the first Gmail plugin to scale to millions of users - and sold it to LinkedIn. That experience exposed something: Gmail was getting slower and more bloated every year, while knowledge workers were spending enormous time managing email. That was the opening.

But building an email client is hard. The Superhuman team coded for years without launching publicly. By the time they had a real product, Rahul faced an uncomfortable question: did they actually have product-market fit, or were they just building something they personally liked?

He turned to the existing literature. Paul Graham said PMF is when you've made something people want. Sam Altman described it as users spontaneously telling other people to use your product. Marc Andreessen famously wrote that you can always feel PMF when it's happening - customers buying as fast as you can make the product, money piling up, reporters calling. But Rahul found these definitions intangible. They described the outcome of PMF, not how to get there. None of them told you what to do on a Tuesday when you're staring at your dashboard wondering why retention isn't where it should be.

So he built a system.

The Core Metric: The Sean Ellis Question

The whole engine is built on a single survey question first popularized by Sean Ellis - who ran early growth at Dropbox and Eventbrite: "How would you feel if you could no longer use this product?" with three answer choices: very disappointed, somewhat disappointed, or not disappointed.

The benchmark that matters: if 40% or more of your users say "very disappointed," you have product-market fit. Below that, you'll struggle to grow no matter how much you spend on ads or outbound. The number isn't arbitrary - it was validated across hundreds of venture-backed startups. Companies that consistently cleared 40% demonstrated strong, repeatable growth. Companies below it kept hitting walls.

Why disappointment instead of satisfaction? Because asking people if they like something invites politeness. Asking how much they'd miss it gets to whether it's actually necessary in their lives. That's the difference between a nice-to-have and a must-have. Satisfaction surveys get you feel-good data. The disappointment question gets you dependency data.

Superhuman started with a score of 22%. Not great. But the methodology showed them exactly what to do about it - and within three quarters, they drove that number to 58%.

The High-Expectation Customer: The Concept Most People Skip

This is the piece that gets left out of most PMF write-ups, and it's arguably the most important concept in the whole framework. Before you even finish segmenting your survey data, you need to understand what you're actually looking for: your High-Expectation Customer, or HXC.

The HXC framework comes from Julie Supan, a marketer who worked with companies including Airbnb and Dropbox. The core idea is that the HXC isn't a broad persona meant to describe everyone in your target market. It's the most discerning person within your target demographic - the one with the highest standards, the most options, and the least patience for mediocrity. If your product delights the HXC, it will almost certainly work for everyone less demanding in that market.

Think about how this plays out with other products. Airbnb's HXC doesn't simply want to visit new places - they want to belong somewhere. Dropbox's HXC doesn't just want file backup - they want to organize their life and keep their life's work safe. The HXC framing pushes you past functional benefits into emotional ones. What does your most demanding customer actually need to feel?

For Superhuman, working from the survey data of users who said "very disappointed," the team built out their HXC profile. They named her Nicole. Nicole is a hard-working professional who deals with many people - likely a founder, executive, manager, or business developer who processes a high volume of email daily. She prides herself on responsiveness. She's not just trying to clear her inbox; she's trying to feel fast, in control, and competent. Speed and keyboard mastery aren't features to her - they're identity signals.

The reason this matters mechanically: once you've defined your HXC, you have a filter. Every feature request, every roadmap debate, every positioning question gets run through the same question - does this serve Nicole? If yes, prioritize it. If no, deprioritize it regardless of how many users asked for it. The HXC is your north star for product decisions, not your average user.

This is also why building for your HXC is not the same as building for your ICP. Your ICP (ideal customer profile) is a sales construct - it describes the company or person most likely to buy. Your HXC is a product construct - it describes the most demanding person who will love what you've built and spread the word. Confuse the two and you'll optimize for conversion at the cost of product quality and retention. You want to build for the person who will love the product most, not just the one most likely to swipe a credit card.

The Four-Step PMF Engine

The survey question is just the trigger. The real work is what comes after. Here's how the engine runs:

Step 1: Survey

Send the survey to users who have actually used the product at least twice in the last two weeks. You want retained users, not drive-by signups. Rahul's team sent four questions total:

• How would you feel if you could no longer use this product? (very / somewhat / not disappointed)
• What type of person do you think would most benefit from this product?
• What is the main benefit you get from this product?
• How can we improve this product for you?

You start getting directionally useful data around 40 respondents. You don't need hundreds. You need signal from people who genuinely use the thing.

Notice how tight the survey is. Four questions. Not a 20-minute form. Not a quarterly review. The brevity is intentional - it maximizes completion rates while capturing everything you actually need. The second question is particularly clever: asking users what type of person would benefit most almost always causes happy users to describe themselves. You're crowdsourcing your ICP definition from the people already getting the most value.

Step 2: Segment

This is where most people running a basic Sean Ellis test miss the whole point. Don't look at your overall score in isolation - segment it. Product-market fit isn't uniform across your entire user base. It exists in pockets. Some segments love your product. Others are indifferent.

When Superhuman segmented their data and focused only on their highest-expectation customer profile - the user type who got the most value - their score jumped from 22% to 32% without changing anything about the product. That's not magic. That's targeting. They filtered out users whose profile didn't match the HXC they'd defined, and the underlying score for the right segment was significantly higher than the blended average was showing.

The implication here is huge: if your aggregate score looks weak, you might already have strong PMF with a subset of your users. You don't have a product problem - you have a targeting problem. Tightening who you're acquiring and onboarding can move your score before you ship a single line of new code.

For founders early in the process, this is where the work of seeding your product with the right user type pays off. If you've onboarded a wide variety of early adopters - some who fit your hypothesis, some who don't - your aggregate score will be muddy. Segment ruthlessly. Find the cluster where your score is highest. That's your real market.

Step 3: Analyze

Now you have two populations to study: users who said "very disappointed" (your fans) and users who said "somewhat disappointed" (your fence-sitters). Both matter, but for different reasons.

For your fans: look at what they say the main benefit is. This is your product's core promise - the thing you should double down on and never dilute. For Superhuman early on, users kept saying variations of "speed." Fast inbox, keyboard shortcuts, half the time in email. That's the signal. Everything else is noise.

For your fence-sitters: look at what's holding them back. These are people who almost love your product. Something is stopping them. Maybe a missing feature. Maybe an onboarding friction point. Maybe a pricing barrier. These are fixable - and fixing them has high leverage because the underlying desire is already there. The fence-sitter has already partially bought in. You just need to remove the specific thing blocking full commitment.

The analysis step also surfaces something uncomfortable: you'll almost always find that the "not disappointed" group wants completely different things from your product than your fans do. Don't try to convert them. They're not your market. Chasing them wastes resources and dilutes the product for people who actually care. This is a painful conclusion for most founders, but it's one of the most important disciplines the framework enforces.

Step 4: Build the Roadmap - the 50/50 Rule

Split your roadmap in two. Half goes toward doubling down on what your fans love - making that core benefit even better, even faster, even more undeniable. Half goes toward removing the specific blockers that keep fence-sitters from becoming fans.

This 50/50 split is more than a gut-feel heuristic. It's a structural safeguard against two common failure modes. If you spend all your time doubling down on the magic, you end up with a product that's cool but too limited for mainstream adoption - nobody can actually use it for their full workflow. If you spend all your time addressing objections and plugging gaps, you become a utility that a better-funded competitor will out-feature. You need both. The fans keep you differentiated. The fence-sitter fixes expand your addressable market.

The roadmap output is also more specific than most teams expect. The analysis step gives you language directly from your users - specific features they mentioned, specific pain points they named. You're not guessing at what to build. You're working from quoted evidence. That's the difference between a data-driven roadmap and one built on internal assumptions.

Step 5: Track

Run the survey again with new users every quarter. Watch the score move. Superhuman went from 22% to 58% by running this process repeatedly over several quarters. That's not luck. That's iteration with a clear feedback loop.

Rahul treats the PMF score as a permanent operational metric - not something you measure once at launch or once when you're fundraising. The score wobbles naturally as you expand into new segments, add adjacent personas, or change pricing. As Superhuman widened its market beyond founders to other professional segments, their PMF score naturally dipped. That's expected and healthy - as long as you're tracking it and responding. The score gives you early warning before that dip shows up in your churn numbers.

What the Superhuman Story Actually Looks Like in Numbers

Let's ground this in what the methodology produced for Superhuman specifically, because the results are worth studying.

The company spent years building without a public launch. When they finally started bringing users in, they ran the Sean Ellis survey and got a 22% "very disappointed" score. Below the 40% threshold. Not yet at PMF.

By segmenting to their HXC profile and focusing their roadmap on the 50/50 split between reinforcing speed (the core love) and addressing specific friction points (missing integrations, onboarding gaps), they drove the score to 58% in under a year. They did this while capping growth intentionally - Rahul was reportedly on calls with every new user personally in the early days, both to onboard them well and to gather the qualitative signal needed to feed the engine.

The downstream results: Superhuman built a waitlist that grew to hundreds of thousands of people. They built a product so valued by their HXC that users were evangelizing it before it was widely available. They charged a premium price that filtered for exactly the high-expectation customer they'd defined. And they built a business that eventually raised over $100 million in funding at a significant valuation.

The PMF engine didn't just help them build a better product. It gave them a story - a disciplined, quantified, replicable story about why their product worked for a specific person in a specific context. That story is credible to investors, compelling to the press, and useful internally for every product and marketing decision.

The HXC vs. ICP: Why This Distinction Matters for Your Sales Motion

I want to dig into the HXC vs. ICP distinction a bit more because it has direct implications for how you prospect and who you bring into your product in the first place.

Your ICP is defined largely by firmographic and behavioral criteria: company size, industry, job title, buying power, sales cycle length. It's optimized for your ability to close a deal. Your HXC is defined by emotional criteria: what kind of person has the standards, the context, and the problem orientation that makes them most likely to love what you've built? It's optimized for retention and word of mouth.

The mistake founders make is building their prospect targeting entirely around the ICP and then being surprised when retention is weak. They're bringing in people who fit the buying criteria but don't fit the product love criteria. The deal closes, but the user never becomes a fan. They churn at renewal. They don't refer anyone. Your PMF score stays low because you're onboarding people who aren't your HXC.

The fix is to align your outbound targeting with your HXC profile, not just your ICP. That means your cold email lists, your LinkedIn sequences, your paid targeting - all of it should be pointed at the specific segment of people who are most likely to become your Nicole, not just the ones most likely to take a demo call.

When I'm building prospect lists for early customer acquisition in a specific niche, I use a B2B lead database to pull targeted lists filtered by title, industry, company size, and seniority. The point isn't to blast a thousand people - it's to put the product in front of the specific user type whose profile matches your HXC hypothesis, so the PMF survey data you collect reflects your real target market, not a random assortment of whoever signed up first.

If you're building for founders and executives specifically - the "Nicole" profile - you want to seed your beta or your early access list with founders and executives. If your early cohort is mostly junior employees and students who signed up out of curiosity, your PMF data will point you in the wrong direction. Garbage in, garbage out.

How This Applies Beyond SaaS

If you're running an agency, a coaching program, or any service business - this framework still works. The question translates directly: "How would you feel if you could no longer work with us?" The segmentation logic is the same. Find the clients who'd be devastated to lose you, figure out what specifically makes them feel that way, and build your positioning and service delivery around that signal.

For agency owners: your HXC isn't every B2B company in your vertical. It's the specific profile of client who gets transformational results from your work and talks about you without being asked. That person has a job title, a company stage, a problem context. When you find three or four of them saying the same thing about why you're irreplaceable, that's the signal. Build your positioning around that, not around the broadest possible description of who you could theoretically serve.

When I was growing my agency business past seven figures, the turning point was getting ruthlessly specific about who we were best for. Not every marketing agency, not every B2B company - a specific job title, a specific company size, a specific problem we solved faster than anyone else. The Superhuman method gives you a systematic way to find that pocket of fit instead of guessing.

The same logic applies to productized services, info products, and newsletters. If you're running a paid community or course, the Sean Ellis question is just as valid. Survey your most active members. Find out who would be devastated to lose access. Use those people to define your HXC. Then build your marketing and content strategy around attracting more people who fit that exact profile.

Common Mistakes When Running This Framework

Surveying too early

If you send this survey to people who signed up last week and haven't really used the product, the data is garbage. You need real usage. Two sessions minimum, ideally over two or more weeks. Otherwise you're measuring first impressions, not dependency. The question is asking about loss - and you can't miss something you've barely tried.

Looking at aggregate scores without segmenting

This is the single biggest mistake. An aggregate score of 28% might hide a segment that scores 55%. Dig into the data before you conclude you don't have fit. You might already have it - just with the wrong audience being counted in your denominator. Rahul's team discovered this when they segmented out non-HXC users and watched their score jump ten points in minutes - without changing the product at all.

Trying to fix the "not disappointed" group

This is a trap. Founders get anxious about the users who don't love the product and try to add features to win them over. Nine times out of ten, those users are simply not your market. Chasing them bloats the product and confuses your positioning. The framework is explicit about this: ignore the "not disappointed" group entirely. Focus on your fans and your fence-sitters. Let the indifferent ones go. It takes discipline to do this, especially when the "not disappointed" users represent revenue or when your team has emotional attachment to solving their problems.

Running it once and moving on

PMF isn't a checkbox. It's a moving target. As you add users, expand to new channels, or change pricing, your audience composition shifts - and so does your score. Run the survey quarterly. Treat it like a financial metric, not a one-time validation exercise. Superhuman tracks it continuously, rolling scores up monthly and quarterly. As they expanded beyond their original segment, the score naturally fluctuated - and that fluctuation is useful information, not a problem to suppress.

Using the wrong survey pool

Sending the survey to your entire email list instead of retained active users is a version of the same problem as surveying too early. You want signal from people who have actually formed a habit around your product. Cold subscribers, trial users who never converted, and churned users who got reactivated will all distort your data. Be strict about the qualification criteria: active use, at least twice, in the last two weeks.

Treating the score as the endpoint

The score is a compass, not a destination. Getting to 40% doesn't mean you stop running the engine. It means you've crossed a threshold where growth tactics will start working. But the goal is to keep the score as high as possible as you scale - and that requires continuous iteration. The score is an input to your roadmap, not a graduation ceremony.

How to Use PMF Data to Drive Your Roadmap Conversations

One underrated benefit of this framework is what it does for internal alignment. Most roadmap debates are battles of opinion - engineering wants to refactor, product wants to ship new features, sales wants whatever the last enterprise prospect asked for. The PMF survey data cuts through all of that because it's direct voice-of-customer evidence.

When you can point to 40 or 80 user responses and say "here's what our fans say the core benefit is, in their own words" and "here are the three specific things our fence-sitters say they're missing," the conversation changes. It's no longer about whose opinion is loudest. It's about what the data says.

The word cloud analysis that Rahul's team ran on open-text responses is particularly useful here. When you see the same words showing up repeatedly across dozens of responses - "speed," "shortcuts," "clean interface" - that's not one person's opinion. That's a signal. And when the fence-sitters keep mentioning the same missing integration or the same friction point, that's not a feature request. That's a blocker you need to fix before those users can become fans.

Build this into your quarterly planning cycle. Every quarter: run the survey, segment, analyze the word patterns, update your roadmap split accordingly. Make the PMF score a standing agenda item in your leadership meetings. Make it as routine as reviewing your MRR or your churn rate.

Finding the Right Users to Survey in the First Place

One practical bottleneck founders hit: you need actual retained users to survey. If you're pre-launch or early stage, getting the right people in the door matters before you can run any of this.

The quality of your PMF data is directly limited by the quality of your early user cohort. If you acquired your first hundred users through a ProductHunt launch that attracted mostly other founders and indie hackers, your data reflects that specific population - not the broader market you're actually targeting. You need to be intentional about who you're seeding the product with from day one.

When I'm prospecting for early customers in a specific niche, I'll use a B2B lead database to pull targeted lists by title, industry, company size, and location - then reach out with cold email to get beta users who actually match the profile I'm testing. You want to seed your product with the right user type from the start, or your PMF data will be muddy from day one.

If you're building for a specific persona - say, operations managers at mid-market SaaS companies - go get 50 operations managers at mid-market SaaS companies into your product through outbound. Don't wait for them to find you organically. Run a targeted cold email sequence, get them onboarded, watch how they use the product, then survey them. That's the fastest way to get clean PMF data on the right population.

If your target market skews local or you're building for small businesses in a specific category, ScraperCity's Maps scraper can pull targeted local business data fast so you're not wasting time on cold outreach to people who'll never be a fit. For specific verticals like real estate or home services, there are scrapers built for exactly those markets as well - the point is matching your outreach list to your HXC hypothesis before you survey, not after.

The idea here isn't to spray and pray. It's to put your product in front of the exact profile you're hypothesizing will love it - then run the survey and let the data confirm or redirect you. Think of outbound as your user acquisition strategy during the PMF-finding phase, not just the growth phase.

What a 40%+ Score Actually Unlocks

Once you cross the 40% threshold, the game changes. Growth tactics that felt like pushing a boulder uphill start working. Word of mouth accelerates. Retention stabilizes. CAC drops because you're not fighting churn with acquisition spend.

Below 40%, no amount of paid ads, cold email volume, or growth hacks will fix the underlying problem. The product isn't sticky enough for the audience you're targeting. The Superhuman method is valuable precisely because it tells you this clearly and early - before you've burned your runway chasing a false start. A 38% score doesn't mean you're close. It means you have real work left to do on the product or the targeting before you should scale.

I've seen too many founders hit a wall at Series A or B because they scaled on top of weak PMF. The metrics looked good enough to raise - reasonable MRR, okay retention, acceptable NPS - but the moment growth spend slowed, churn ate the business. The score would have told them the truth earlier if they'd looked. The investors would have asked about it if more founders were running this framework.

Above 40%, growth investments compound. Organic referrals start to show up meaningfully. Content and SEO start to convert at higher rates because people who find you actually stick around. Paid acquisition stops being a leaky bucket. The word-of-mouth loop that Superhuman built - every email sent via Superhuman included a branded signature that drove curiosity from recipients, which drove waitlist signups, which drove more of the right users into the product - only works if the product is sticky enough to keep people sending emails through it. PMF is the prerequisite for every growth motion that follows.

PMF as a Continuous Practice, Not a One-Time Gate

One of the most important things Rahul has emphasized since publishing the original framework is that PMF is not a gate you pass through and leave behind. It's an ongoing measurement practice. Every time you expand into a new segment, launch a new pricing tier, add a new use case, or push into a new geographic market, you're potentially changing your user composition - and with it, your score.

Superhuman experienced this directly. As they expanded beyond their original core of founders and early tech executives to a broader professional audience, their PMF score naturally dipped. That's not failure. That's expected. The engine exists precisely to catch that dip early, understand which new segments are pulling the score down, and either fix the product for those segments or decide that they're outside the HXC definition and should be deprioritized.

The practical cadence that makes sense for most companies: run the survey with each new cohort of users, track the rolling score quarterly, and make it a permanent part of your operating rhythm - right next to your MRR, your NPS, and your churn rate. If the score moves materially in either direction, that's a signal worth investigating immediately rather than waiting for the quarterly review.

Treat it the way serious companies treat financial metrics. You wouldn't skip a month of revenue tracking because things seemed fine. Don't skip a quarter of PMF tracking for the same reason.

Putting PMF Data to Work in Your Positioning and Marketing

The PMF survey doesn't just generate a score. It generates language. The open-text responses from your fans - the words they use to describe the main benefit they get from your product - are your positioning gold mine.

Most founders write their positioning copy from the inside out. They describe what they built, how it works, what features it has. But your HXC doesn't care about features. They care about outcomes. The survey shows you exactly how your fans describe those outcomes in their own language. Use that language in your copy, your sales emails, your landing pages, your outbound sequences.

When Superhuman's fans said "speed" and "keyboard shortcuts" and "half the time in email," that language showed up everywhere in their marketing. The tagline wasn't "an email client with a clean interface." It was about being the fastest email experience ever built. That positioning resonated because it was sourced from the people who loved the product most - not from a marketing brainstorm session.

Your cold email subject lines, your LinkedIn headlines, your paid ad copy - all of it should be rooted in the language your fans use to describe the core benefit. That's what makes messaging feel authentic rather than generic. And it's one of the practical outputs of running this framework that most write-ups don't emphasize enough.

Applying the Framework to Idea Validation Before You Build

You don't have to wait until you have a live product to start using the logic of this framework. The underlying question - would this be a must-have or a nice-to-have for my target user? - can be explored through customer discovery interviews before you've written a line of code.

The pre-product version: instead of surveying active users, you're doing exploratory interviews with the profile you think will be your HXC. You're asking them about the problem you're solving, how they currently handle it, what they've tried before, and how frustrated they'd be if nothing new ever came along to solve it. The emotional tenor of those answers will tell you a lot about whether you're solving a vitamin or a painkiller problem before you've invested months in development.

If you're still in idea-validation mode and want a structured way to stress-test whether a concept has the bones to find real PMF, check out the Business Idea Roaster - it's a free tool I put together that pressure-tests the core assumptions before you go deep.

For SaaS ideas specifically, the SaaS AI Ideas Pack is a free resource that can help you identify concepts with stronger PMF potential from the start - before you're three months into building something the market doesn't need badly enough.

Common Mistakes When Running This Framework

Surveying too early

If you send this survey to people who signed up last week and haven't really used the product, the data is garbage. You need real usage. Two sessions minimum, ideally over two or more weeks. Otherwise you're measuring first impressions, not dependency.

Looking at aggregate scores without segmenting

This is the single biggest mistake. An aggregate score of 28% might hide a segment that scores 55%. Dig into the data before you conclude you don't have fit. You might already have it - just with the wrong audience being counted.

Trying to fix the "not disappointed" group

This is a trap. Founders get anxious about the users who don't love the product and try to add features to win them over. Nine times out of ten, those users are simply not your market. Chasing them bloats the product and confuses your positioning. Focus on your fans and your fence-sitters. Let the indifferent ones go.

Running it once and moving on

PMF isn't a checkbox. It's a moving target. As you add users, expand to new channels, or change pricing, your audience composition shifts - and so does your score. Run the survey quarterly. Treat it like a financial metric, not a one-time validation exercise.

A Worked Example: Applying This to a B2B SaaS

Let me make this concrete. Say you're building a project management tool targeted at creative agencies. You launch, onboard your first 200 users, and run the Sean Ellis survey on the 80 who have used the product at least twice in the last two weeks. Your aggregate "very disappointed" score is 31%.

Below 40%. Feels bad. But before you spiral, segment the data.

When you break the "very disappointed" group down by user type, you find:

• Freelance designers: 18% very disappointed
• Agency owners and creative directors: 52% very disappointed
• Junior project coordinators: 24% very disappointed

Your product is already past the threshold for agency owners and creative directors. You don't have a product problem. You have a targeting problem. Your acquisition is bringing in a mix of freelancers and coordinators who aren't getting the same value the agency owners are.

Next: look at what the agency owners (your fans) say the main benefit is. They keep saying things like "finally a tool that works the way we actually run projects" and "the client-facing view saves us three hours a week." That's your core promise. Protect it. Build on it.

Now look at what the somewhat disappointed agency owners say is missing. Three of them mention the lack of a time-tracking integration. Two mention that the onboarding is too long. One mentions mobile access.

Your roadmap split: the next two sprints go 50% toward making the client-facing view even more powerful (double down on the love), and 50% toward shipping a native time-tracking integration and shortening the onboarding flow (fix the blockers).

Your acquisition strategy pivots: you stop marketing to freelancers and junior coordinators. You start running cold outreach specifically to agency owners and creative directors, using the language your fans gave you. You build your content strategy around the problems agency owners have that your product solves.

Ninety days later, you run the survey again with your new cohort. If you've tightened your acquisition and fixed the top blockers, your score should move. That's the engine.

How to Start This Week

You don't need a fancy survey tool or a data team. Here's the minimum viable version:

• Identify 40-100 users who have used your product at least twice in the last two weeks
• Send them a short email with the four survey questions (use Typeform, Google Forms, whatever)
• Calculate your overall "very disappointed" percentage
• Segment responses by user type, job title, or use case
• Find your highest-scoring segment - that's your real ICP and the starting point for your HXC definition
• Pull the open-text responses from your fans and identify the 2-3 words or phrases that keep showing up - that's your core benefit
• Pull the open-text responses from your fence-sitters and identify the top 2-3 blockers
• Audit your roadmap: is it doubling down on what the fans love, or chasing features for people who don't care?
• Adjust your acquisition strategy to bring in more people who match your HXC profile

Run it again in 90 days. Compare scores. If you're not moving, your roadmap split is probably off - either you're not leaning hard enough into the core benefit, or you're not addressing the right blockers. Or your acquisition is still bringing in the wrong user types, which means your survey pool is still dirty.

The Superhuman product market fit framework isn't complicated. The discipline is in actually running it, acting on the data, and resisting the urge to build features for the wrong people. That last part is harder than it sounds, especially when the wrong people are paying you and asking for things loudly.

If you want to talk through how to apply this to your specific product or service - and get feedback on your positioning and ICP from someone who's been through the PMF process multiple times - I work through exactly this inside Galadon Gold with founders and operators building real businesses.

And if you're still in idea-validation mode, sign up for the Daily Ideas Newsletter - I share raw business concepts and frameworks regularly that can sharpen your thinking before you commit to a direction.

Frequently Asked Questions About the Superhuman PMF Framework

What is the Superhuman product market fit framework?

It's a four-to-five step process developed by Superhuman CEO Rahul Vohra for measuring and systematically improving product-market fit. The framework uses a single survey question ("How would you feel if you could no longer use this product?"), segments the responses to identify the highest-value user cohort, analyzes what fans love and what holds fence-sitters back, and generates a data-driven roadmap that balances doubling down on core strengths with fixing specific blockers. The output is a numerical PMF score that can be tracked over time.

What does the 40% benchmark mean?

The 40% benchmark refers to the proportion of surveyed users who answer "very disappointed" when asked how they'd feel if they could no longer use the product. This threshold was identified by Sean Ellis after benchmarking hundreds of startups. Companies above 40% consistently demonstrated strong, sustainable growth. Companies below it struggled to scale regardless of marketing investment. Superhuman used this as their north star metric throughout their early development phase.

What is a High-Expectation Customer (HXC)?

The HXC is the most discerning person within your target demographic - the user with the highest standards who will love your product for its greatest benefit and help spread the word. The concept comes from marketer Julie Supan. Superhuman's HXC was a persona they named Nicole: a high-performing professional who processes high email volume daily, prides herself on responsiveness, and craves speed and efficiency above all else. Defining your HXC gives you a filter for every product, marketing, and sales decision.

How is the HXC different from an ICP?

Your ICP (ideal customer profile) is a sales construct focused on firmographic and behavioral criteria - who is most likely to buy. Your HXC is a product construct focused on emotional and psychological criteria - who is most likely to love what you've built and become an evangelist. Confusing the two leads to optimizing for acquisition at the cost of retention and word of mouth. You want to build for the HXC and sell to the ICP - ideally, they overlap significantly.

How often should you run the PMF survey?

Quarterly is the standard recommendation, though many teams track it on a rolling monthly basis. The key is consistency - running it with each new cohort of users and comparing scores over time. PMF is not a one-time measurement. It shifts as your audience composition changes, as you expand into new markets, as you change pricing or positioning. Treat it as an operational metric, not a launch gate.

Can you use this framework if you don't have a software product?

Yes. The core question translates directly to any business where customers develop a usage habit - agencies, coaching programs, newsletters, productized services, physical products. "How would you feel if you could no longer work with us / read this / use this?" The segmentation and analysis logic is identical. You're still looking for the cluster of customers who'd be most devastated by losing you, defining that profile, and building your service and marketing around it.