How to Start a Startup

HTSaS · Chapter 02

Getting Started

Launch small, do the unscalable things nobody wants to do, and treat press as a multiplier — never a substitute for traction.

After this lesson

  • Ship a first version fast, on purpose imperfect
  • Recruit your first users manually instead of waiting for scale
  • Know when (and when not) to seek press

Picking up where Chapter 1 left off

You've decided to start, on a specific problem. Lecture 8 (Stanley Tang, Walker Williams & Justin Kan) is where the original course turns from 'should you' to 'how, this week' — how to get started, do things that don't scale, and think about press. This chapter compresses that lecture plus Paul Graham's essay on the same theme.

Launch fast, launch small

Waiting to launch until the product is 'ready' is usually a way of avoiding the scary part: real users with real reactions. A tiny, embarrassing launch to ten people who actually have the problem beats a polished launch to an audience that doesn't care.

Read the original

Lecture 8 — How to Get Started, Doing Things That Don't Scale, and Press — Stanley Tang, Walker Williams & Justin Kan

Extract the specific unscalable tactics each speaker used to get their first users — the mechanism matters more than the anecdote.

This chapter is a working summary. When the idea matters, read the source once — then come back and do the practice.

Read the original

Do Things that Don't Scale — Paul Graham

Pull the general principle — almost all successful startups did something embarrassingly manual early on — and one concrete tactic you can copy this week.

This chapter is a working summary. When the idea matters, read the source once — then come back and do the practice.

Unscalable things that actually work

  1. Recruit users one at a time, by hand, through people you know or can reach directly
  2. Do the service manually before you build the automation (concierge MVP)
  3. Deliver the product in person and watch people use it
  4. Give your personal phone number or email to your first users
  5. Onboard people yourself instead of writing a self-serve flow too early

Case

Stripe's 'Collision installs'

Early Stripe founders would sit down next to a developer at a coffee shop and install the payments integration for them on the spot, in minutes.

Takeaway: They traded scale for signal. A handful of hand-installed users taught them more about onboarding friction than any amount of marketing copy testing could have.

A minimal launch sequence

  1. Ship a version that does one job well, nothing else
  2. Recruit 5–20 first users by hand, from your own network outward
  3. Watch them use it in person or on a call; don't just ask what they think
  4. Fix the top one or two blockers; re-ship within days, not weeks
  5. Only then consider a broader or public launch

Press is a multiplier, not a strategy

Press amplifies something that's already true — real usage, a real story, real traction. It cannot manufacture traction that doesn't exist. Journalists want a story with a hook (a number, a first, a controversy), not a pitch deck. Getting covered too early, before you have anything to show, burns the story for later.

What actually earns coverage

  1. A surprising number (growth rate, scale, savings) you can back up
  2. A clear, specific problem a journalist's readers also have
  3. Being first or notably different — not 'another X for Y'
  4. A relationship with the reporter built before you need something

Is this press-worthy yet?

You're tempted to pitch TechCrunch this week. What's actually true?

Practice