Rob Snyder's Pull Framework for Product-Market Fit

A financial advisor once told the founders of an AI note-taking startup called Jump that using their product felt like how Alexander Graham Bell must have felt the day he invented the telephone (11:14). It is a strange thing to say about software that just writes down meeting notes. But that strangeness is the point. Rob Snyder, author of The Power of Pull and writer of a newsletter called The Physics of Startups, has a name for the feeling behind that comment: pull. Pull is what happens when a customer would be weird not to buy.

Snyder says he has helped about 50 startups go from zero revenue to more than a million dollars a year in recurring revenue (02:11). His central argument, made early in the conversation, cuts against how most people hunt for product-market fit: the moment when a product's value to customers becomes undeniable and growth takes care of itself.

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. (16:20) — Rob Snyder

Pull, in other words, is never a property of a market segment or an org chart. It belongs to one person: the head of legal at a specific company, the specific financial advisor filing a specific report. Snyder's framework for finding that person spells out PULL: a Project on someone's to-do list, something Unavoidable about doing it now, a List of options they've already tried, and the Limitations that make those options fall short (07:33). All four have to be true at once. Miss the U and someone will politely tell you how interesting your idea is and then do nothing.

The Jump example shows why the unavoidable part matters so much. Financial advisors are legally required to log client meeting notes in their compliance systems, and skipping that step can mean real trouble with regulators (10:28). Generic AI note-takers did not integrate with their systems and were not clearly compliant, so advisors were stuck (10:49). That stuckness is what made them grab Jump's product. Snyder contrasts this with insurance agents, who also hate taking notes but face no such regulatory requirement. Their existing habit, however annoying, was good enough, so the same pitch fell flat (11:52, 12:17). Same job, same irritation, no pull, because one letter of the acronym was missing.

Snyder knows this the hard way. As a first-time founder he built a product with hair-on-fire pain points, a differentiated solution, and daily-active design partners, then spent two years getting nowhere before a pivot finally worked (05:07, 05:31). What he eventually realized was that pain and value are not what make someone buy. People do not drop their existing priorities for a slightly better option; they buy to get an unavoidable project unstuck.

Earned Secrets, Not Whiteboards

If pull already exists out in the world, the founder's job is not to invent a clever idea but to go find it. Snyder is blunt about where good ideas do not come from: a whiteboard, a brainstorm, or ten thirty-minute Zoom interviews (18:53). He tells the story of a friend with no product idea who simply picked industries to shadow in person for days at a time. He hated working with solar panel installers. He loved independent physical therapists, rode along with them, and eventually tried to buy a franchise just to study its profit and loss statement. Somewhere in that process he spotted a real, unavoidable cost the therapists were stuck trying to cut, and that became his pull hypothesis (20:36, 21:23). Snyder calls this kind of insight an earned secret, something you cannot get by asking people what they want, only by living next to the problem.

He is also skeptical of a piece of standard startup advice: find design partners. In his experience, design partners tend to play what he calls Steve Jobs, handing out feedback with no intention of ever paying, because people with a genuinely unavoidable project do not want to help build a product, they just want the project finished (32:12).

Selling To Learn

Once a founder has a pull hypothesis, the next step is cold outreach, and Snyder's advice here is almost anti-marketing. Skip the template that opens with a pain point and a pitch. Instead, send what he calls a Founder Magic message: short, personal, and interesting enough to earn a conversation, nothing more. One startup he advised got replies with a line as simple as mentioning that Snoop Dogg had invested in them (35:39). The message's only job is to get someone on a call, not to sell them anything.

For volume, Snyder suggests adding 100 to 150 people on LinkedIn per week and messaging them once they accept (39:33). His rough benchmarks: 30 to 40 percent of connection requests should be accepted, and 30 to 50 percent of those people should reply (40:09). Fall short, and the problem is usually the profile or the message, not the market.

The sales call itself follows a specific order. Snyder recommends opening with a direct question: why did you agree to take this call (42:35)? Often the prospect will answer by accidentally filling out the entire PULL framework unprompted. Calls run thirty minutes, because shorter blocks tend to get skipped on people's calendars (43:37). His heuristic for cutting losses: do five calls, then take stock. If everyone leans back and nothing surfaces, the hypothesis is probably wrong (45:13). When a call goes lukewarm, Snyder throws what he calls a Hail Mary: asking if there is anything else on the person's to-do list they would pay someone to take off their plate right now (29:11). He has watched more than one startup pivot entirely off that single question.

We're selling to learn, not really to, to, like, push something. We're gonna learn, and we're seeing if they try to buy. (46:16) — Rob Snyder

Pricing, too, gets grounded in the framework rather than guessed at. Snyder suggests anchoring early deals to the cost of whatever the customer is already paying, whether that is a consultant, a competitor, or the time of an existing employee (56:00). It gives a number that feels intuitive to the buyer instead of arbitrary.

The Factory, Not the Widget

Landing a first customer is not the finish line. Snyder says every new customer is guaranteed to churn until they cross a specific point where leaving would feel strange. He points to a stat from sales consultant Mark Roberge: Slack found that customers who sent around 10,000 messages early on almost never churned (68:21). Snyder's own startup, which sold retention software to restaurants, found its own version of that threshold in the first two weeks after signup (68:21 context, 67:57). The job of onboarding, in his framework, is reverse-engineering that threshold and building a fast path to it for every customer.

His bigger mental model ties all of this together. Snyder argues that a startup, at its core, is a machine for repeating one case study an unlimited number of times: a customer with pull, a sales process that surfaces it, and a delivery process that locks in retention (72:59). Once that case study exists, the founder's job is not to invent a hundred different stories but to build a factory that reproduces the same one, over and over, and to look at each stage, pipeline, sales, delivery, as a potential bottleneck to fix one at a time.

When none of this is working, Snyder's last tool is what he calls goblin mode: doing something a little extreme, like flying to a conference uninvited or going door to door, just to get in front of real customers and learn something (79:38). It is not a sustainable operating mode, he says, just a break-glass option for when a founder is stuck and time is burning.

Near the end of the conversation, Snyder offers the best summary of what the whole framework is for. It will not hand anyone a guaranteed idea. What it does is narrow the search.

This is just giving you a metal detector. You might not find metal where you're looking right now, but at least I'm not giving you a spoon. (81:05) — Rob Snyder

That, more than any acronym, is the quiet promise of the whole method: not a shortcut to product-market fit, but a faster, less painful way to find out, honestly, whether it exists at all.