Closing Your First B2B Contract: From Vague Design Partner to Recurring Revenue

At YC, you see the same mistake over and over again. A founder rolls into Demo Day with a "design partnership" at a company with a fancy logo — six months in, still not a single dollar of revenue. They're proud of the logo on their website. They think they've made progress. They haven't.

This article condenses YC's framework for how early B2B founders actually move from first conversation to recurring revenue contracts — and why 90% get stuck in stage one. Four stages, a handful of tactical tips, and a clear "pro move" at the end.

The four stages

Most early B2B companies move through a predictable progression. The goal is to move through it as fast as possible:

1. Design partnerships (usually unpaid, usually too long)
2. Free pilots / proof of concept
3. Paid pilots
4. Recurring revenue contracts with an opt-out (pro move)

90% of founders get stuck at stage 1. Only 5–10% try to leapfrog straight to stage 4 before the product is mature — that rarely works. Most founders are too slow, not too fast.

Stage 1: Design partnerships — and why most of them are a waste

You have Figma mockups, or just an idea. You're selling into an industry with deep domain knowledge (law, accounting) that you don't have yet. So you propose a "design partnership": you sit with the customer, observe how they work, and co-design the product alongside them.

Sounds great. In practice: three months, six months, badly defined scope, low engagement because the customer isn't paying for your time and they've got their own business to run. You get a fancy logo on your website that feels like progress. You're not a dollar closer to real revenue.

What actually works in a design partnership is sitting next to the customer's keyboard for a few days. Ask: "What part of your job do you hate the most? If you could wave a magic wand, what would disappear?" Offer to do the work manually yourself. The best founders go so far as to get qualified as an auditor or a real-estate agent and actually work the job for a few months.

The goal: identify a narrow, burning problem you can build a wedge product for in 48 hours. Build it, take it back, see if they're happy to pay. If yes — don't build more. Sell the same wedge to ten similar customers.

The mistake is trying to build a broad platform to reach feature parity with the existing software. You can't, you're too small. You waste months. And the design partner won't tell you "this sucks" — they'll be helpful and suggest "just one more feature," because they don't want to hurt your feelings.

Stage 2: Free pilots and proof of concept

You have a product — maybe just a wedge — but no other happy customers as social proof. Naturally the customer wants to try it before paying. That becomes a pilot or PoC.

Same trap as a design partnership if you're not careful: three months, no clear endpoint, low engagement. One founder told me a French customer asked for a pre-proof of concept. Nobody knows what that is.

What you need: a value equation. If you're selling a customer-service AI that solves 20% of inbound queries, say it directly: "We expect to reduce your team from 100 to 80 people, saving roughly a million in salaries. We charge $200K for the software." Now you have something concrete to prove.

Techniques that work:

• Backtest on historical data — give us a thousand past tickets, we'll show how many we resolve
• Side-by-side trials — our AI answers in parallel with your agents, the answers don't go out, we compare at the end
• 1% of volume or a smaller geography — low risk, nobody gets fired if it goes sideways

Your job is to make sure your champion can walk into their CFO's office with proof and say: "This is worth a million, costs $200K. Obviously we're doing it."

The biggest trap at this stage: you're too afraid to talk money. You think it'll scare them off. It almost never does. If we deliver these metrics — what's that worth to you? Customers not ready or willing to pay — disqualify them early.

Stage 3: Paid pilots

Lesson from stage 2: shorten the pilot and get a paid commitment up front. Once a customer puts money in, they take the pilot seriously.

Also ask for the full-product price at the same time — what's the annual fee? Better to know early if they're not going to pay at all. If a long procurement process looms, ask your champion what they personally can approve on a corporate card. $10–20K without involving finance is often much better than a bigger number that takes three months to land.

Other commitments worth asking for:

• Wait until there's a suitable project to test on (if a new client engagement starts next month — wait for that)
• Customer data ready to go
• A dedicated person or team testing the product
• Check-ins every couple of days so you can fix bugs overnight

Keep the timeframe short — 7 to 14 days if the product is in shape. You're not selling a bug-free experience. You're selling yourself and the team — the promise that if something breaks they have your cell number and you respond 24/7.

Time-to-first-value is the northstar. Going from weeks to hours is the single biggest lever for pilot-to-paid conversion. In practice: do janky things. Never do a full API integration during a pilot. Excel exports work. Have the customer email you data, you email back the result. Whatever gets them to value fastest.

And book the post-pilot meeting before the pilot starts. That's where you walk through metrics and hard ROI. The decision to keep going should be crystal clear by then.

Stage 4: Recurring revenue with opt-out — the pro move

The downside of paid pilots: you still have to negotiate the main contract afterwards. It's like a second sales process when you thought you were done.

Pro move: monthly or annual recurring contract with a 30–60 day money-back / opt-out window at the start. If the customer does nothing, it converts to full recurring after the window. One sales process, one recurring revenue contract.

It's magic. In a sales meeting you can say with confidence: "This is how customers buy our product — annual contract, 30-day opt-out. X, Y, and Z all signed on those terms." Wildly persuasive.

You can't leap straight here. Maybe the sales process isn't tight enough yet, maybe you lack social proof, maybe the product isn't ready. For the first one or two customers a free pilot might still be the way in. But don't stay on all fours too long.

Investor disclaimer: be careful what you report as ARR/MRR to investors when customers are still inside the opt-out window. The tactic is fine, this is just a transparency point — be clear with investors about where each customer sits.

Customer success: the work starts after signature

A company I talked to last week had signed $4M in contracts — but implemented less than $2M. They were missing a customer-success function. Once you start closing one of these every week, you need as much or more effort on onboarding and making sure the customer is actually getting value.

Tips from YC companies that actually work

• Start SOC 2 today. Plus HIPAA, ISO 27001 if relevant. They take months to start. Earlier is better.
• Treat your champion like a co-founder. They sell for you when you're not in the room, fight budget battles for you. Set a closing date together — they'll miss it every time, but it still creates urgency.
• Map the buying process explicitly. "Describe the last time you bought software like this — who internally signed off, what was the process?" You have an economic buyer, a technical approver, a security gatekeeper, a legal team, and the daily users. Win over each one.
• Never close a meeting without the next touchpoint booked.
• Get on a plane. "Hey, I'll be in Houston next week — lunch?" If they say yes, book the flight. It works wonders.
• Don't fight contract terms. The customer's legal team loves to redline. Ask: is this clause company-ending or just an annoyance? Accept anything that's not unlimited liability or IP transfer of your product.
• Use scarcity honestly. "We're talking with seven or eight potential customers but only have capacity for two enterprise customers this quarter. Are you in, or do we talk again in six months?"

In summary

90% of early B2B founders get stuck at stage 1. The way out is to constantly push toward more structure, more paid commitment, shorter cycles. Design partnership → free pilot with a clear value equation → paid pilot with a tight timeframe → recurring contract with opt-out. And the moment the first customer signs: build a customer-success function before the second one closes.

You can't run before you can walk. But don't stay on all fours too long either.