Run customer development and early sales as one funnel, with a founder running the calls. Each stage ends when the buyer gives something up: a second meeting, an intro to the person with budget, a priced evaluation, a signature, a referral. Write down what makes an opportunity qualified: you have confirmed the problem, your contact can influence the purchase, a budget path exists, and no missing capability blocks it. Count close rate only on those deals. Steve Blank designed customer development so discovery feeds customer validation, the step where early sales prove a repeatable sales process. He has written that 3–5 early customers may be enough to show it in enterprise software, and 300–500 for a web product.
This page covers the motion from the first discovery call to the first referral, once you have something to sell or a priced offer to test. Before production code, start with how to validate demand before you build. Once a buyer wants proof in their own environment, read how to convert a pilot to a paid contract. Once they say yes, read how to negotiate your first customer contract.
Run discovery and sales as one funnel
| Stage | Exit: what the buyer gives you | HubSpot default stage | Drop or park when |
|---|---|---|---|
| Exploratory interview | A follow-up call on one named problem | Appointment scheduled | They cannot recall a recent instance of the problem |
| Focused interview | High importance, low satisfaction with today’s fix, and a next meeting | Appointment scheduled | Today’s workaround is cheap and good enough |
| Qualified opportunity | Confirmed problem, budget path, decision process, no blocking gap | Qualified to buy | Nobody you can reach can approve the spend |
| Evaluation | A demo for the people who decide, then a priced pilot or proposal | Presentation scheduled; Decision maker bought-in | They ask for one more feature instead of a signature |
| Paid | A signature and an invoice they pay | Contract sent; Closed won | Terms drift with no date to sign |
| Referral | An intro to a peer with the same problem | Closed won, tagged by source | They will not refer you. Ask what stops them |
Blank’s Customer Development model has four steps. In customer discovery, the founders test hypotheses about customer problems in front of customers. Customer validation is where you develop a sales model that can be replicated and scaled. Customer creation and company building come after. If you cannot find enough paying customers during validation, his model sends you back to discovery. He calls discovery done when customer responses converge and more interviews stop producing significantly new information.
The FounderNexus session view: research and sales are two ends of the same conversation. Treat each stage as a qualification gate. A buyer who will not give feedback on a prototype was not going to buy, and finding that out at stage three costs you little. Judgment from a FounderNexus session, not a survey.
Write down what qualified means
Koomen, who co-founded Optimizely, tells founders to spend the first call on questions and save the pitch for later. You need people who have the problem, the budget, and the decision-making authority to buy. His list starts with what made them take the call and ends with who else will weigh in. Bosmeny’s model call opens the same way: why did you agree to take my call today, and how do you solve this now?
| Check | Ask on the call | Park or drop when |
|---|---|---|
| Problem | Tell me about the last time it happened. How bad is it, and how do you measure the cost? | No recent example, or nobody can put a cost on it |
| Today’s fix | How satisfied are you with how you handle it now? | They are satisfied, or the workaround is free and good enough |
| Path to money | Does fixing it make money, save money, or cut risk? | The payoff sits many steps away from the product |
| Influence | Who makes the buying decision? Who else needs to weigh in? | Your contact cannot name the decider or get you to them |
| Budget | What is your budget for this? How does your company buy software? | No budget and no owner who can create one |
| Blockers | What would stop you from using this full-time? | A missing capability you cannot deliver soon |
The influence and budget rows come from Koomen’s question list. The importance, satisfaction, and path-to-money rows are FounderNexus session judgment: hunt for high importance and low satisfaction, and prefer a payoff a few logical steps from the product. A long chain from product to payoff is a kill signal. The same session advice on who to call first: a manager or director who feels the problem and can influence the decision. Design the first call to disqualify, and test one customer hypothesis at a time in outbound.
End each call with the next step booked
Fitzpatrick’s The Mom Test names two signals that a meeting worked. Commitment means the buyer gives up time, reputation, or money. Advancement means “they are moving to the next step of your real-world funnel and getting closer to a sale.” His rule of thumb: “If you don’t know what happens next after a product or sales meeting, the meeting was pointless.” He lists “Looks great. Let me know when it launches.” as a bad finish and “When can you come back to talk to the rest of the team?” as a good one.
Bosmeny’s advice from the Stanford startup class: get people to a yes or a no as quickly as you can. A thousand maybes will sink you, and a no frees you to move on to someone who might say yes. His follow-up slide for Clever’s second deal shows a long trail of unanswered emails, pricing calls, and redlines with a buyer who wanted the product.
FounderNexus session judgment adds three stage-gate questions for later calls. What would we need to build for you to use this full-time? To pay full price? To refer us to peers? The referral answer doubles as a gap list. Whatever stops a happy buyer from referring you belongs on the roadmap. For deals with several people involved, write the remaining steps, owners, and dates in a plan the buyer edits with you, and keep your own close-date guess apart from dates the buyer agreed to.
Fill the top by hand
Bosmeny frames prospecting with Everett Rogers’s adoption curve. The innovators, about 2.5% of companies, are the ones who will take a call from a startup with no users and no revenue. That makes the top of the funnel a numbers game. His three channels at Clever: the founders’ network, small industry conferences where the buyers gather (email the attendee list ahead and book meetings), and short cold emails that say who you are, what you built, and ask for time.
Paul Graham’s 2013 essay Do Things That Don’t Scale calls recruiting users by hand the most common unscalable thing founders do at the start. YC’s name for Stripe’s version is the Collison installation. Anyone who agreed to try Stripe heard “Right then, give me your laptop,” and the brothers set them up on the spot.
FounderNexus session judgment on outbound: invite prospects into your content, such as an interview series or a podcast about their work, with no pitch. Keep prospects who are not buying now on a nurture cadence, and go back with something useful each time. The same session called AI-written first touches a poor start to a relationship, which matches the advice on AI SDR vs a human SDR.
Sell it yourself until it repeats
Graham, in a footnote to the same essay: “you can’t avoid doing sales by hiring someone to do it for you. You have to do sales yourself initially.” Bosmeny says founder passion and industry knowledge outweigh sales experience. Clever peeled one founder off to work on sales full-time. Koomen ends the funnel at habitual use of the product, after the signature and the rollout.
Bosmeny names two closing traps. A buyer who will sign once you build one more feature is passing. Either put the feature in a signed agreement or wait until more customers ask for it. If a buyer asks for a free trial, offer an annual agreement with an opt-out in the first 30 or 60 days instead, because a free trial gets you no revenue and no commitment.
FounderNexus session judgment on the handoff: GTM has no autopilot. Scale out of founder-led sales by hiring someone for your lowest-value bottleneck task. Hiring a VP to do the learning for you skips the stage where the company learns who buys and why. The sequence for that seat is on when to hire the first VP of Sales.
Measure the funnel by cohort
| Metric | Definition | What it tells you |
|---|---|---|
| Discovery calls held | First calls per week, by channel | Whether the top of the funnel is full |
| Focus rate | Focused interviews ÷ exploratory interviews | Whether you are calling the right segment |
| Qualified rate | Qualified opportunities ÷ focused interviews | Whether the problem has a budget and a decider |
| Close rate | Closed-won ÷ qualified opportunities in the same cohort | Whether the offer and the process repeat |
| Cycle time | Days from qualified to signed | Cash timing and how many deals one founder can carry |
| Referral rate | Paid customers who introduce a peer | Whether buyers stake their name on you |
HubSpot’s default deal pipeline has seven stages, each with a preset probability: Appointment scheduled 20%, Qualified to buy 40%, Presentation scheduled 60%, Decision maker bought-in 80%, Contract sent 90%, Closed won 100%, Closed lost 0%. HubSpot multiplies each stage total by that probability to show a weighted amount. Those figures are software defaults. Replace them with your own cohort rates before you show a weighted pipeline to your board.
There is no reliable public benchmark yet for interview-to-opportunity, opportunity-to-paid, or discovery-to-referral conversion at Seed to Series A. Set the bar from your own cohorts and the revenue plan. Blank’s 3–5 and 300–500 figures describe how many early customers can show a repeatable process. They are not conversion rates.
Run the funnel in order
Customer development funnel
- Write the stage exits. Six stages, each ending on a buyer commitment you can count.
- Write the qualified line. Problem, today’s fix, path to money, influence, budget, and blockers.
- Fill the top by hand. Network, small industry conferences, and short cold email (Bosmeny).
- Run exploratory interviews. Past behavior, no pitch. The method is on the demand validation page.
- Score the pain. Importance and satisfaction with today’s fix (FounderNexus session).
- Qualify before the demo. Budget, buying process, and who decides (Koomen).
- Book the next step before you hang up. Advancement or a no (Fitzpatrick; Bosmeny).
- Demo against their problem, then ask for money. A priced pilot or the order form.
- Ask what stops a referral. Put the answer on the roadmap (FounderNexus session).
- Read the cohort. Close rate on qualified deals, cycle time, and whether interviews still teach you something new (Blank).
Mistakes that fill the pipeline with noise
Counting interviews as pipeline. An interview becomes an opportunity when it passes the qualified line you wrote.
Moving deals on enthusiasm. Move a deal when the buyer gives you the exit for that stage.
Keeping maybes alive. Push for a yes or a no and spend the time on the next prospect (Bosmeny).
Building the one more feature. Sign first with the feature in the agreement, or wait for more buyers to ask (Bosmeny).
Offering a free trial. Offer an annual agreement with an early opt-out, or charge for the pilot.
Hiring a seller to find the motion. Graham and the FounderNexus session agree: do sales yourself first.
Trusting default stage probabilities. HubSpot’s presets are placeholders until your cohorts replace them.
Pair with demand, pilots, contracts, and enterprise
Test willingness to pay before you build with how to validate demand before you build. Turn an evaluation into a purchase order with how to convert a pilot to a paid contract. Write the first order form with how to negotiate your first customer contract. If the first buyers are large companies with a buying group and a security review, read selling to the enterprise as an early-stage startup. Decide whether tooling can take any of the outreach with AI SDR vs a human SDR. Orient the rest of the motion from the GTM hub.
Sources
- Steve Blank, Customer Development Manifesto: The Path of Warriors and Winners (part 5). 17 September 2009. Four steps: customer discovery (founders test hypotheses about customer problems in front of customers), customer validation (a sales model that can be replicated and scaled), customer creation, company building; validation loops back to discovery when you cannot find enough paying customers. Blank’s reply in the comments, 1 October 2009: discovery is done when responses converge and more interviews stop producing significantly new information; validation is done with enough early sales to declare a repeatable and scalable sales process, 3–5 early customers for enterprise software, 300–500 for a web product.
- Rob Fitzpatrick, The Mom Test. Chapter five, Commitment and advancement: commitment as time, reputation risk, or cash; advancement as moving to the next step of your real-world funnel; “If you don’t know what happens next after a product or sales meeting, the meeting was pointless”; good and bad meeting finishes.
- Y Combinator, Enterprise Sales (Startup School). Pete Koomen, November 2024. The first call is for questions, not a pitch; find people with the problem, the budget, and decision-making authority; question list from why they took the call to who else weighs in; the funnel ends when the customer uses the product habitually.
- Y Combinator, How to Start a Startup, Lecture 19: Sales and Marketing. Tyler Bosmeny (Clever), Stanford CS183B, 2014. Funnel of prospecting, conversations, closing, revenue; Rogers innovators about 2.5% of companies; over 400 companies contacted in the first two months of YC; network, conferences, and cold email; ask why they took the call; get people to a yes or a no; one-more-feature trap; free trials replaced by an annual agreement with a 30 or 60 day opt-out; founder passion and industry knowledge over sales experience.
- Paul Graham, Do Things That Don’t Scale. July 2013. Recruiting users by hand as the most common unscalable thing founders do at the start; the Collison installation; footnote 2, founders do sales themselves initially and hire a salesperson later.
- HubSpot Knowledge Base, Set up and manage object pipelines. Last updated 29 September 2026. Default deal pipeline: seven stages with preset probabilities (20%, 40%, 60%, 80%, 90%, 100% won, 0% lost) used to compute weighted amounts in board view.
- FounderNexus. Founder sessions, 2026. Research and sales as two ends of one conversation; stages as qualification gates; score importance against satisfaction with today’s fix; short path to money; manager or director first; first call designed to disqualify; one customer hypothesis at a time; stage-gate questions on full-time use, full price, and referral; content-led outreach and nurture; no autopilot for GTM; hand off the lowest-value bottleneck task first. Not a survey. No closed-session numbers.
Related
- Go-to-market
- How to validate demand before you build
- How to convert a pilot to a paid contract
- How to negotiate your first customer contract
- Selling to the enterprise as an early-stage startup
- AI SDR vs a human SDR
- When to hire the first VP of Sales
Founders who have turned discovery calls into their first paid customers will pressure-test your stage exits, qualification line, and next-step habit in a FounderNexus session.