See pricing
An 86-minute guide to finding PMF | Rob Snyder (Author, The Power of Pull)
Lenny's Podcast: Product | Career | GrowthPMF

An 86-minute guide to finding PMF | Rob Snyder (Author, The Power of Pull)

Rob Snyder turns product-market fit into a concrete test: find a specific person with an urgent project, weak alternatives, and a reason to buy now, then see if they actually try to move forward. The episode is especially useful because it shows how to validate demand before building too much, run early sales calls, and recover when the first hypothesis is wrong.

October 25, 2026 · 86 min listen · 16 min read · Rob Snyder
0:00
–:––

Context

Rob Snyder is the author of The Power of Pull and writes The Physics of Startups. He says the framework came from painful firsthand experience as a founder who spent years pushing a product the market did not want, then later from helping more than 50 startups grow from zero to $1 million in ARR. In this episode he turns product-market fit into a practical process for finding demand, running early sales, closing first deals, and learning from churn. It matters to PMs because the same logic can be used inside startups or inside larger companies launching a new product, feature, or business line.

The Big Idea

Product-market fit is not something a team talks a market into. It is something a team discovers by finding a specific person with pull: an urgent project, inadequate alternatives, and a reason to act now.

Rob's argument is that early product work should be built to expose that pull as fast as possible. Compliments, feature requests, and design feedback are weak signals. The strongest signal is simple: the buyer tries to move forward before you even have much product.

Key Insights

Pull lives in a person

  • Rob argues that teams often define markets at the wrong level. You are not selling to SpaceX or to mid-market private equity as an abstract category. You are selling to one person, such as a head of legal or an operating partner, who is blocked on a specific job.
  • This matters because company-level segmentation hides the actual trigger to buy. A person-level view makes you name the real urgency, the real workflow, and the real champion.
  • For PMs inside larger companies, this is also a reminder that an ICP is incomplete until it names the exact role and the exact project.

All four conditions matter

  • The PULL framework only works when all four parts are true at once: a project, an unavoidable reason to do it now, a list of current options, and limitations that make those options insufficient.
  • Jump found this with financial advisors. Advisors needed to reduce time spent taking client notes and entering them in the CRM. The work was unavoidable because missing those records could create serious compliance risk. Generic AI note takers were not good enough because they did not integrate with advisors' CRMs and their compliance status was unclear.
  • The counterexample was insurance agents. They shared the annoyance of note taking, but the work was not equally unavoidable, and doing less or nothing was still an acceptable option. That missing urgency meant no real pull.

Pull is discovered, not created

Rob found this idea liberating because it replaces persuasion with search. If pull already exists in the world, your job is not to convince people to care. Your job is to find where they are already straining against a blocked workflow. That is why he strongly prefers firsthand exposure over abstract brainstorming. One founder he mentioned spent days shadowing small businesses, then independent physical therapists, and even explored buying a physical therapy franchise, because living the workflow exposed the real stuck points that a whiteboard would never have surfaced.

Test before you build

  • Rob warns against the classic move of going heads down because the idea does not feel ready for customers yet.
  • His preferred early test is a two-slide conversation:
  • Slide 1 restates the person's project, why it is urgent now, what they have tried, and why those options fall short.
  • Slide 2 shows a simple concept for how you might help.
  • The goal is not polished validation theater. The goal is to hear some version of do I have that right on slide 1, then how do we get started on slide 2.
  • AI can shorten prototyping time, but it should not delay contact with reality.

Learn from buying behavior

  • Rob treats early customer work as a search for action, not a search for praise.
  • Positive but distant reactions such as this is interesting, this is the future, or send me more information are weak signals. They show curiosity, not pull.
  • Strong signals look different. The buyer leans in, asks how it works, asks how to move forward, or offers to pay before the product exists.
  • His test is blunt but useful: if the conversation makes it feel weird for the person not to buy, keep going. If the conversation makes it feel like you are pushing, step back and revisit the hypothesis.

Sales is structured learning

  • Rob frames early sales as discovery with the option to transact.
  • He starts calls with a simple question: why did you take the call with me today? In the best conversations, the buyer answers by describing most of the pull framework themselves.
  • If that does not happen, he earns the right to ask discovery questions before pitching. That avoids turning the call into an interrogation or a persuasion contest.
  • He also recommends reassessing quickly. After roughly five calls, ask whether anyone actually pulled, what objections repeated, and whether a new hypothesis is needed. If nothing is working, use a Hail Mary question or go into goblin mode to find better evidence fast.

Retention proves real fit

  • A first sale is only partial proof. Rob's rule is that every new customer is effectively guaranteed to churn until they hit a specific moment that makes it weird to churn.
  • That moment is usually behavioral, not attitudinal. He cites the idea of a leading indicator of retention, and in his own startup the decisive moment was when a restaurant got an alert that an employee was likely to quit and could prevent it within the first two weeks.
  • Churn can also reveal the real segment. One horizontal AI workflow company saw most customers churn, but freight forwarders kept using one workflow. That pattern exposed the real market and led to a vertical pivot.
  • Once you know the retention event, onboarding should be designed backward from it.

One case study should repeat

Rob's scaling model is unusually concrete: a startup is a factory for repeating one successful case study. The goal is not to create many unrelated customer stories. It is to create one before-and-after pattern that keeps working across pipeline, sales, and delivery. That gives teams a clean way to prioritize. If the bottleneck is too little pipeline, fix pipeline. If the bottleneck is close rate, fix sales. If the bottleneck is churn, fix onboarding and value delivery.

Mental Models & Frameworks

The PULL framework

  • Project: what the person is trying to get done.
  • Unavoidable: why that project has to happen now rather than later.
  • List of options: the current tools, people, or workarounds the person could use.
  • Limitations: why those current options are not good enough.
  • Use the framework when forming an idea, narrowing an ICP, or diagnosing weak demand. If even one part is fuzzy, Rob's view is that you probably do not have real pull yet.

The two-slide test

  • Slide 1 is a plain-language restatement of the buyer's world: the project, the urgency, the current options, and why those options fail.
  • Slide 2 is a rough concept for how you could help. It does not need to be a full product or a polished demo.
  • Use it before investing heavily in build. The point is to test whether the buyer says yes, that is exactly the problem, and then immediately starts pulling for next steps.

The Hail Mary question

  • Ask this when the main hypothesis is falling flat: what is on your to-do list that you would pay to get done right now?
  • The question works because it stops defending the current idea and reopens discovery around real urgency.
  • Rob has seen entire pivots come from this move, including a startup that shifted from an applicant tracking system to a different job-description problem inside government hiring.

The retention event

  • Treat every new customer as likely to churn until they hit an early moment that predicts staying.
  • That moment is product specific. It is not always activation, daily use, or NPS. It is the first concrete experience that makes the value undeniable.
  • Use this model when designing onboarding. Start from the retention event and work backward, rather than starting from a checklist of setup tasks.

The case study factory

  • Think of the company as producing one repeatable customer transformation.
  • The factory has three stages:
  • Pipeline: getting the right people into conversations.
  • Sales: turning the right conversations into signed customers.
  • Delivery: getting those customers to the retention event.
  • Use this model to find the current bottleneck. The bottleneck, not the loudest problem, should drive near-term priorities.

Decision Principles

Principle: Name the actual buyer

  • When: defining an ICP, starting discovery, or writing a launch brief for a new product.
  • Why: company labels blur the trigger to buy. A named role with a named project makes demand testable.

Principle: Follow urgency over praise

  • When: deciding whether to keep investing in a feature, segment, or prototype.
  • Why: compliments are easy to get and cheap to misread. Real urgency shows up as attempts to move forward now.

Principle: Price from current options

  • When: setting early pricing before you have much market data.
  • Why: Rob recommends anchoring on the value of getting the job done or the cost of the current alternative, such as a consultant or an internal team.

Principle: Work backward from deadlines

  • When: closing a larger deal or planning implementation inside a customer account.
  • Why: asking when the buyer needs results gives you a real forcing function for procurement, onboarding, and proof of value.

Trade-offs & Nuance

Services can teach faster

Manual or service-shaped delivery can validate pull and reveal the real workflow before full automation. That speed is valuable when the market is still unclear. The risk is getting trapped in a consulting model, so the team has to keep asking which pieces are repeatable and which pieces still need software.

Better UX is contextual

Linear could win against Jira because better experience mattered to the people evaluating issue tracking. Rob has tried the same logic in HR software and found that buyers did not care enough about UX to switch. A superior experience only creates pull when it unblocks something the buyer already cares about.

Goblin mode has a cost

Goblin mode can surface truth quickly. Rob describes it as the break-glass option when nothing is working: book the flight, go to the conference, line up twenty meetings, do the unhealthy thing that gets you in front of customers now. The trade-off is energy and sustainability. It is useful in a dead end, but it is not a healthy default operating model.

Common Mistakes

Mistake: Building before evidence

  • Founders often assume the idea is too rough to show, so they disappear to build.
  • Rob's warning is that this delays the cold wind of reality. You can spend months refining something that never had pull.
  • A better move is to test the hypothesis with simple language, rough concepts, or a lightweight prototype before serious build.

Mistake: Selling to broad companies

  • Saying we sell to banks or we sell to private equity firms is usually too vague to be useful.
  • The harm is not just bad messaging. It also leads to shallow discovery because no single person's urgency is defined.
  • Rewrite the target as one role with one blocked project.

Mistake: Mistaking praise for demand

  • Nice meetings are dangerous because they feel like progress.
  • Rob looks for pull-shaped actions, not polite enthusiasm. If the buyer is not trying to move the deal forward, the signal is weak.
  • The better approach is to judge conversations by attempted next steps, not by how positive the mood felt.

Mistake: Using design partners

  • Rob is skeptical of design partners because they often create a false positive. People buy into the vision, give lots of feedback, and never buy.
  • His reasoning is simple: people with a truly urgent problem usually do not want to co-design your roadmap. They want the problem solved.
  • He gives the same warning for proactively offering pilots or proof-of-concept arrangements. If a buyer requires a pilot, fine. Do not lead with free collaboration as the default.

Mistake: Copying the category winner

  • Rob calls out the trap of entering a space with an obvious winner and no clear theory of why a different buyer is still blocked.
  • Teams often justify this by saying their version is better. His response is that better is meaningless if the current solution is already good enough for the target buyer.
  • The safer version of the idea is to name the overlooked customer and the specific limitation that the winner does not solve.

Practical Application

Write one pull hypothesis

  • Fill out four lines for one exact role: the project they are trying to complete, why it is urgent now, what they use today, and why those current options fail.
  • Do not accept industry labels or generic pain statements. If any line feels broad, treat it as an open question rather than as evidence.

Run a two-slide interview

  • Put the hypothesis on slide 1 in plain language. Put a rough concept on slide 2.
  • Ask whether slide 1 is accurate before you explain anything. Then show slide 2 and watch for action, not approval.
  • End every conversation with either a concrete next step or a clear no.

Use curiosity-based outreach

  • Start with a short message that makes the person interested enough to talk. Rob's examples were a surprising founder credential, a relevant backstory, or a timely reason for the conversation.
  • Do not lead with a generic pain-point template. The job of outreach is to earn a conversation, not to close the deal in text.
  • Rob's rough LinkedIn benchmarks:
  • 100 to 150 connection requests per week
  • 30 to 40% accept rate
  • 30 to 50% reply rate from accepted connections
  • If you miss those numbers, first fix the profile or opener, not the deck.

End with a Hail Mary

  • When the original hypothesis is flat, ask what is on the person's to-do list that they would pay to get done right now.
  • Capture those answers separately from normal interview notes. They are possible pivot material, not random anecdotes.
  • If several answers cluster around the same blocked job, write a new pull hypothesis and test it immediately.

Onboard to the retention event

  • For the first few customers, do not just hand over access and hope.
  • Define the first moment that predicts long-term retention for your product. Then design onboarding so every customer reaches that moment as fast as possible, even if the process is manual at first.
  • Review every churned account against that milestone. Ask whether the team failed to get the customer there, or whether the customer never truly had pull.

Questions to Consider

  • Which exact role in the target account has a project urgent enough that it would feel strange not to buy, and have you written that role down precisely?
  • What evidence shows that the current alternatives are truly failing, rather than merely being annoying or inelegant?
  • In the last five discovery or sales conversations, how many people tried to move forward now instead of offering compliments or future interest?
  • What specific user action in the first two weeks would make churn feel strange for your product, and is onboarding designed to get every customer to that action?
  • If the current hypothesis fails after five calls, what is your team's Hail Mary question or goblin-mode move for learning faster?

Bottom Line

Product-market fit becomes easier to reason about when you stop asking whether the market likes the idea and start asking whether a specific person is trying to buy right now. Rob Snyder's framework is useful because it turns that question into a repeatable process for discovery, sales, pricing, onboarding, and even pivoting when the first idea is wrong.

Resources Mentioned

ResourceTypeWhy it was mentioned
The Power of PullBookRob's book distills the zero-to-one lessons behind the episode.
The PULL frameworkArticleThe written overview of the framework discussed throughout the conversation.
The Physics of StartupsNewsletterRob's newsletter is where he keeps refining the model and sharing zero-to-one lessons.
The ultimate guide to JTBDPodcast episodeRob compares the Project part of PULL to a Job to be Done and points to Bob Moesta's work for deeper context.
Sell the alpha, not the featureArticleLenny referenced Jen Abel's enterprise sales ideas when discussing buyers, champions, and deal process.

Case Studies Mentioned

Jump's regulated workflow

Jump found pull by focusing on financial advisors who had to take meeting notes and enter them into their CRMs. The work was unavoidable because of compliance pressure, and generic AI note takers were not good enough because they lacked the right CRM integrations and clear regulatory fit. The lesson is that urgency plus inadequate alternatives creates far stronger demand than a broad promise to save time.

Clearly AI's own pain

Rob pointed to Emily at Clearly AI as an example of solving your own problem the right way. She had been in security engineering at AWS and found something she personally would have paid to remove, but no product existed. The lesson is that solve your own problem works only when the problem is real enough that you would have bought a fix immediately.

Freight forwarder pivot

One horizontal AI workflow company saw heavy churn across most customers, but one pattern stood out: freight forwarders kept using one specific workflow. Instead of treating that as a small edge case, the company pivoted around it and became a more premium vertical product. The lesson is that churn data can reveal the real market faster than top-of-funnel excitement can.

Teero's door-to-door sprint

Teero, a dental staffing company, used an extreme version of customer discovery when ordinary outreach was not enough. Rob said the founder's parents went door to door to dental offices to get people to talk to him about becoming customers. The lesson is that when you are stuck, speed of learning matters more than elegance of method.

Experiments to Try

Experiment: Five-call pull check

  • Hypothesis: a specific role has an urgent project, weak alternatives, and enough frustration to try to buy now.
  • Change: write one pull hypothesis and turn it into a two-slide conversation.
  • Test: run five calls in one week with people who match the role closely.
  • Measure: count how many buyers ask how to move forward now, not how many say the idea sounds useful.

Experiment: Hail Mary rescue

  • Hypothesis: when the current idea is weak, buyers will reveal a better opportunity if asked what they would pay to get done right now.
  • Change: add the Hail Mary question to every flat discovery call.
  • Test: review ten answers and cluster them by repeated job, role, and urgency.
  • Measure: look for repeated jobs that generate stronger follow-up interest than the original pitch.

Experiment: Find the retention event

  • Hypothesis: one early customer action predicts whether an account will stay.
  • Change: manually onboard the next set of customers and log the first moment when value becomes undeniable.
  • Test: compare retained accounts and churned accounts against that early moment.
  • Measure: identify the behavior that best separates accounts that stay from accounts that leave.

Notable Quotes

One of the main mistakes people make is thinking that pull is a thing that a company has. It's a thing a person has. (Rob Snyder)

It would be weird if they bought, but a person with pull is weird not to buy. (Rob Snyder)

We're seeing if they try to buy. If they don't, cool. We've learned something. (Rob Snyder)

We don't make software here, we make case studies. (Rob Snyder)

AI PM course

Everyone hears the same episodes.
Few can do what they describe.

Start for free