Why the Hard, Manual Work That Doesn’t Scale Is Exactly What Your Startup Needs
Have you ever been told that in the startup world, you just need to build a great product and the users will come? Turns out, that’s only half the story. In his celebrated piece “Do Things that Don’t Scale,” veteran founder‑and‑mentor Paul Graham flips conventional startup wisdom on its head and argues that if you want to truly take off—which most startups do—you’ll need to do things that don’t scale. Let’s dive into what that means, why it works, and how it applies to today’s tech and business environment.
🚀 The Big Idea
At its core, Graham’s essay says: almost every successful startup begins with manual, messy, unscalable efforts. Yes, building the product matters—but before automation, growth hacks, and viral loops, you need sweat equity: recruiting users one‑by‑one, obsessing over feedback, delighting early adopters, even doing things by hand until you can’t. According to Graham, ignoring this phase is a top mistake founders make. (Paul Graham)
Key Insights in the Essay
1. You build the engine before it has momentum
Graham uses the metaphor of a car engine with a hand‑crank: exceptional performance won’t take off by itself—you have to push it. (Paul Graham) Startups rarely just “go viral” out of the gate. Founders must recruit early users manually, test the product, refine it, and push it until the engine hums.
2. Focus on the very early user‑experience
He argues that early on, your users’ experience—how they feel signing up, using you, getting responses—matters more than you might assume. For example, instead of waiting for users to sign up, go out and set them up personally. That kind of attentiveness creates a culture of “insanely great” service. (Paul Graham)
3. Narrow your focus to fuel the fire
Rather than launching broadly, start small: pick a narrow market segment where you can ignite real engagement, then expand. The example: Facebook originally for Harvard students only—worth fewer users, but each one deeply engaged. (Paul Graham) This “contained fire” strategy lets you build momentum rather than spreading thin.
4. Manual = learning
When you do things by hand (e.g., you are the “software,” you set up each user, you visit their homes) you intimately learn how the product is used, where it fails, what truly matters. That feedback loop is often faster and richer than designing everything for “automated scale” from day one. (Paul Graham)
5. Scaling comes later; don’t worship it at first
Founders often worry about scale too early—automation, big launches, large partner integrations. Graham says these are distractions at the very beginning: what matters is depth, not breadth. Launching wide with no depth usually fails. (Paul Graham)
Why This Still Matters in 2025
Even though Graham wrote this in 2013, the underlying truths hold strong today. With AI tools, low‑code stacks and instant global reach, it’s tempting to think automation and virality are everything. But for many new ventures—whether SaaS, marketplaces, AI products, or B2B tools—the manual groundwork still holds:
- AI/ML startups: You might have sophisticated models, but until you engage deeply with early customers, iterate manually, your product might miss real needs.
- Marketplaces: They’re especially fragile early. Without manual seeding (user acquisition, service quality), the “network effect” never triggers.
- Niche or vertical applications: They might not scale like consumer apps—but by deeply serving a tight niche you build a strong foundation for expansion.
What’s also interesting is how this mindset bleeds into organizational culture: founding teams who go roll‑up‑their‑sleeves early often retain the “obsession with user experience” even after they grow. That becomes a differentiator.
Reflection: How Can Founders Use This Today?
- Start recruiting users manually: Even with digital launches, allocate founder time to reach out, onboard, and service early users.
- Invite feedback actively: Go beyond surveys—watch users, talk to them, tweak your product based on real actions, not assumptions.
- Choose your initial target narrowly: Don’t try to please everyone at once. Pick a segment where you can dominate or be deeply loved. Then expand.
- Automate later: Resist the urge to build full‑scale systems on day one. Use manual hacks to validate product‑market fit and user behavior.
- Embed the attitude: If you’re going to promise “insanely great experience,” don’t reserve that for the “big launch” later—make it part of your DNA now.
Glossary
- Unscalable: Activities that cannot be maintained efficiently as you grow (e.g., manually setting up each user).
- Contained fire strategy: The idea of focusing on a small segment or niche, getting intense engagement there, before scaling outward.
- Product‑market fit: When your product satisfies a strong market demand and you see self‑sustaining growth.
- Feedback loop: The cycle of getting user input → improving product → getting more users → more input.
- Network effect: A phenomenon where each additional user adds more value to other users, typical in marketplaces and platforms.
In short: If you’re building something new and you’re tempted to skip the messy “doesn’t scale” phase in favour of big launches or automation—don’t. Embrace the grunt work, the hand‑holding, the early manual effort. Those are exactly the steps that turn a good idea into something that grows.
Source: https://www.paulgraham.com/ds.html?utm_source=linkedin
