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

Rob, I kept hearing from founders and product people in my community that your book, The Power of Pull, was unusually useful. That got my attention, because there is no shortage of startup advice. You’ve helped dozens of companies move from zero to meaningful revenue, so I wanted to understand where this very practical way of finding product-market fit came from.

I definitely didn’t plan to write a book. I got punched in the face trying to build a company from zero to one, eventually stumbled into what worked, and then friends pulled me into their companies. I became weirdly obsessed with this stage because it’s fascinating, painful, and there wasn’t much that helped me make sense of it.

The useful thing here is that you make a vague feeling concrete. Pull is not just, “people say they like it.” It’s that someone has such a pressing need that not buying your product would feel strange.

Right. Most founders are pushing: trying to persuade people that a valuable product matters. Pull is when a person has an active project, needs to complete it now, and can’t get there using their current team, tools, or alternatives. They’re like a compressed spring, and if you can unblock them, they want to grab the thing from you.

PULL is the checklist. P is the project they’re attempting, not “buy recruiting software,” but maybe “hire more people.” U is unavoidable: why can’t this wait? The first L is the options they’ve already considered, and the last is why those options fail. You need all four; otherwise people may complain, praise you, or say it’s interesting, but they won’t transact.

The financial-advisor example makes that real. They need meeting notes in their CRM, and compliance means they can’t simply skip it. Generic AI note tools didn’t connect to their systems or make the regulatory situation clear, so Jump arrived with something that actually removed the blockage. But insurance agents had a similar annoyance without the requirement, so the urgency wasn’t there.

And pull belongs to an individual, not a logo. You’re not selling to SpaceX or to mid-market private equity. You’re selling to a particular legal leader, partner, or operator whose own work is stuck. That distinction keeps you from treating a broad market label as if it were a buyer.

So before building, you write a pull hypothesis: this exact person is trying to do this, must do it now for this reason, and has exhausted these inadequate routes. Then the real work is getting close enough to reality to test it.

Please don’t treat ten quick Zoom interviews as some heroic search when your company’s life depends on it. Go work alongside people, shadow them, and see the job firsthand. One friend tried industries from solar installation to physical therapy, spent days in the work, even considered buying a practice, and only then saw a specific cost problem worth solving.

I’d put the hypothesis in plain language on a slide: “Here’s what I think you’re trying to accomplish, why it matters now, and why the usual approaches aren’t cutting it. Is that basically true?” Then show a simple idea for help. On the first slide you want, “Yes, exactly.” On the second, you want them asking how to move ahead.

The key signal isn’t polite excitement, a request for more materials, or a future-quarter maybe. It’s a pull-shaped move: they ask how to start, whether you can do it, or offer money before there’s much product. And solving your own problem only counts if you’d honestly have paid someone to solve it now, not merely if you think it ought to exist.

When you get lukewarm reactions, throw the Hail Mary: ask what is on their list that they would pay to remove immediately. A team selling government applicant tracking heard a soft maybe, asked that question, and discovered an urgent job-description problem that became their first deal. For an introvert like me, this reframes selling: I’m not forcing a purchase; I’m finding out whether I can help with something they’re genuinely stuck on.

For cold outreach, don’t send a giant value-prop pitch. Give them a short, credible reason a conversation could be interesting. It might be an unusual founder story, a relevant background, or a timely event. Make it fit the customer—an HVAC owner probably doesn’t care about an impressive Silicon Valley credential—and don’t rely on AI personalization that just recites facts they know.

Once someone takes the call, you recommend a relaxed thirty-minute conversation, not a high-pressure sales performance. Start by asking why they accepted. If they don’t reveal the need themselves, agree on an agenda, ask permission to understand their situation, and then investigate the four pieces of pull.

Exactly. I’m trying to learn whether they have a project, a deadline, failing alternatives, and real limitations—not persuade them through force of personality. If five conversations are all leaned back and teach you nothing, stop and reassess. Sometimes one referral into a totally different sector changes everything. We’re selling to learn: if they don’t try to buy, that is useful information.

If someone does pull, return with a concrete concept and plan, even if it’s a rough mockup or a service-shaped delivery rather than a polished platform. Price can be bounded by the value of completing the project and the expense of current workarounds. Don’t volunteer a free design partnership; somebody with an urgent job usually wants it handled, not wants to play product adviser.

After the sale, assume every customer will leave until you find the moment that makes leaving weird. It’s rarely a generic usage metric. For Slack, a meaningful early behavior was sending a large number of messages; for my restaurant software, it was helping a customer prevent an employee departure. Build onboarding so customers reach that moment quickly, instead of tossing them login details and hoping.

And if churn is brutal, that’s awful, but it can expose the narrow group that truly gets value. A broad AI-workflow company saw almost everyone leave except freight forwarders using one workflow, then focused the whole business there. The certainty never becomes perfect, but each honest test gives you a better next move.

My broader model is a case-study factory. Find one kind of customer who eagerly buys, gets the promised result, and becomes strange to churn; then repeat that story rather than inventing a hundred unrelated ones. Look for the bottleneck—pipeline, conversion, or delivery—and work on that constraint. The simple version is: state the pull hypothesis, show it to people, see whether they try to buy, and think less.

It’s not an easy treasure hunt, but it’s a better tool than digging through every grain of sand with a spoon. And Rob’s still developing the framework, so his invitation is simple: use it, challenge it, and send him what you learn.