Rob Snyder argues pull is a person's unfinished priority that existing options cannot solve, so founders should find it rather than persuade buyers.
Snyder cites Jump, an AI notetaker for financial advisers, whose team spent eighteen months before finding pull around CRM-ready meeting notes.
Snyder says founders should test pull with firsthand research, skip design partners, and ask prospects a Hail Mary question about paid to-do items.
Finding product-market fit means finding people with stuck projects
Product-market fit starts with finding a specific person who has a project they cannot finish with the options available. That pull exists before the founder shows up, so the work is to find it through firsthand research and sales conversations, not to convince people. Snyder frames a startup as a factory that repeats one customer story through pipeline, sales and delivery.
Takeaways
Pull means a stuck project, not a pain point. Pull is a person with a project on their to-do list that they cannot finish with current options, so they would be weird not to buy. It belongs to a specific person, not a company, and exists regardless of the founder.
PULL needs four conditions at once. Project, Unavoidable, List of options considered, and Limitations of those options must all exist together. Jump took eighteen months to find this with financial advisers, whose CRM note-taking is legally required and poorly served by generic notetakers. Insurance agents had the task but no pull.
Good ideas come from firsthand research. Ideas do not emerge from a whiteboard. About 10% of the ideas behind Snyder's 50 companies came from founders' own experience and 90% from firsthand research, such as a friend who shadowed small-business owners for days before forming a hypothesis.
Selling is learning, so skip design partners and early building. Founders should treat early sales as learning, not pushing a product. Snyder has not seen design partners help, since they give feedback and never buy. He also advises against building first and against proactively offering pilots.
Outreach should be short, interesting and high volume. Snyder prefers LinkedIn, adding 100 to 150 people weekly, with 30 to 40 percent acceptance and 30 to 50 percent replies. Messages should intrigue rather than pitch. He is skeptical of AI personalization. Calls open by asking why the prospect took the meeting.
Early pricing tests willingness to transact. PULL sets price bounds through the value of the project or the cost of alternatives, such as market research consultants at about $50,000 a year. Early on, Snyder tests willingness to transact rather than willingness to pay, and favors service-shaped offers first.
Design onboarding around the retention moment. Snyder says every buyer is likely to churn until a key moment, so onboarding should drive toward it. Examples are Marc Roberge's Slack message-count indicator and a restaurant product that flagged churn risk in the first two weeks. Churn of 90% or more must be investigated.
How to test a pull hypothesis
Form a hypothesis about who is stuck, based on time spent with them firsthand.
Tap a name to go there · tap a line to see why they are linked
On this episode 2
Show 1
Reach out with short, interesting messages, mainly on LinkedIn.
Open each thirty-minute call by asking why the prospect took it, then run discovery.
Show two slides, one describing their stuck situation and one sketching how to help, and watch whether they try to buy.
Review what you learned after about five calls.
Ask the Hail Mary question about any to-do item they would pay to have done. One government software startup won its first contract this way.