The Playbook

Do Things That Don't Scale

The early-traction work is manual, unglamorous, and nobody's going to clap for it. Do it anyway.

← All articles · distilled from real founder research

The most common lie founders tell themselves is that there's a clever, automated shortcut to the first hundred customers. There isn't. The real early playbook is a grind of hand-sent emails, manual onboarding, and doing the job yourself — and the founders who won mostly just out-grinded the moment.

Let me give you the receipts.

When Stripe was getting off the ground, Patrick and John Collison did something that would horrify a growth team. When someone showed the faintest interest, they didn't send a signup link and hope. They onboarded them on the spot — took the person's laptop and set it up right there. It got a name: the "Collison install." Behind the curtain, they ran manual back-end plumbing that just looked automated. They faked the machine until the machine existed.

Des Traynor at Intercom sent roughly a hundred hand-written cold emails a day, read the replies, and made tomorrow's batch better. Every day. He also ran webinars at 6am to catch people in Australia and Asia. That is not scalable. That is a person deciding to do a hundred hard things per day until something moved.

DoorDash? The founders spun up a landing page with PDF menus of Palo Alto restaurants and their own phone number on it — and then did the deliveries themselves. Not because it scaled. Because driving the food over meant talking to real customers and hearing what was actually broken. Same energy as Alibaba's Jack Ma cold-calling and physically visiting manufacturers to sign them up and teach them the platform by hand.

Here's the part people miss. Doing things that don't scale isn't a scrappy phase you suffer through because you're too broke to automate. It's how you learn what to automate. You cannot systematize a process you've never run. Zapier's Wade Foster manually configured integrations for people. Salesforce invited prospects to the team's actual apartment to test the product. Every one of these founders was buying information, not just customers.

Look at Wade's move more closely, because it's a masterclass. He went onto product support forums where people were begging competitors for an integration, and he replied — offering to build it, now, with a landing-page link. Those pages got maybe ten to fifteen visitors. But about half signed up. Fifty percent conversion, because he showed up exactly where the pain was screaming and answered it personally. No ad in the world converts like that.

The physical-product founders live this even harder. D'Shawn Russell started Southern Elegance Candle Company at a local farmers market, doing roughly $200 a weekend, face to face, learning which scents people actually reached for. Better Booch spent its first months perfecting kombucha recipes at farmers markets — the markets were both the income and the R&D. Pipcorn debuted at Smorgasburg, a Brooklyn open-air market, hand-stamping bags of popcorn kernels one at a time. You do not hand-stamp bags because it scales. You do it because you're broke, you're learning, and every single sale is a conversation.

There's a mindset shift buried in all this. When you have zero customers, your job is not marketing. Your job is learning, disguised as sales. The manual work is the curriculum. Aaron Krause was in five Philadelphia-area supermarkets with Scrub Daddy before it exploded — he was learning the retail motion by hand, one store at a time, long before Lori Greiner turned it into 42,000 units in under seven minutes on QVC.

So what do you actually do Monday morning?

Pick the most manual, embarrassing, doesn't-scale version of getting a customer, and do it. Send the emails yourself, one at a time, personalized. Get on the phone. Marek Piechut got his first ten customers "on the phone and in a car," literally driving to prospects. Onboard your first users personally, watch them use the thing over a call, fix what confuses them while they're still talking. Do the delivery. Hand-stamp the bag.

It'll feel like it's not "real" growth because it doesn't look like the case studies with the hockey-stick charts. But those charts came after. Every automated growth engine you admire was a manual, ugly, one-at-a-time slog first. Earn the right to scale by doing the thing that doesn't.

---