How long should founders personally do manual work that doesn't scale?

Paul Graham · published Jul 2013

The most common unscalable thing early startups must do is recruit users manually rather than wait for them.

You can't wait for users to come to you; you have to go out and get them.

Scope: nearly all startups

source Do Things that Don't Scale

Paul Graham · published Jul 2013

Even startups solving an urgent problem should aggressively acquire early users in person rather than wait for signups.

Stripe solved an urgent problem and could have waited, but instead set up users on the spot; they became YC-famous for it.

source Do Things that Don't Scale

Paul Graham · published Jul 2013

Marketplaces are so hard to start that founders should expect to take heroic, in-person measures at first.

Airbnb went door to door in New York recruiting users and improving listings; about 30 days of in-person engagement made the difference between success and failure.

Scope: marketplaces specifically

source Do Things that Don't Scale

Paul Graham · published Jul 2013

The usual way to find users to recruit manually is an untargeted launch followed by seeking out more of whichever early users prove most enthusiastic.

Ben Silbermann noticed early Pinterest users cared about design and recruited at a design bloggers' conference, which worked well.

Scope: unless you built it to solve your own problems, in which case find your peers

source Do Things that Don't Scale

Paul Graham · published Jul 2013

Founders should take extraordinary, unscalable measures to delight early users, not merely to acquire them.

Wufoo sent every new user a hand-written thank you note; first users should feel signing up was one of their best choices.

source Do Things that Don't Scale

Paul Graham · published Jul 2013

Worrying that individual customer attention won't scale is misplaced, and delighting customers in fact scales better than founders expect.

Larval startups have nothing to lose; having too many happy users would be a great problem; you can usually make things scale more than predicted and the practice permeates the culture.

source Do Things that Don't Scale

Paul Graham · published Jul 2013

Trying too hard to make initial users happy never leads startups down a blind alley.

Based on Graham's observation across startups he has seen.

source Do Things that Don't Scale

Paul Graham · published Jul 2013

Existing customer-service conventions, set by big companies, are not the upper bound on user experience; being small lets you deliver service no large company can.

Founders' standards come from big companies they've bought from; Tim Cook cannot send a hand-written note after a laptop purchase, but a small founder can.

source Do Things that Don't Scale

Paul Graham · published Jul 2013

Over-engaging with early users is not merely permissible but a necessary part of the feedback loop that makes the product good.

Making a better mousetrap is not an atomic operation; even when you build something you need yourself, the first version is never quite right.

Scope: for most successful startups

source Do Things that Don't Scale

Paul Graham · published Jul 2013

Feedback from direct engagement with a startup's earliest users is the best feedback it will ever get.

Once big enough to need focus groups, you will wish you could still watch a handful of users use your product in their homes and offices.

source Do Things that Don't Scale

Paul Graham · published Jul 2013

Using your own software on lukewarm users' behalf is the right early move, because it teaches you how the product feels to users and creates a near-instantaneous feedback loop.

At Viaweb, building merchants' stores themselves taught them how it would feel to merchants to use the software; missing features were discovered and implemented on the spot.

Scope: early stage recruiting of lukewarm users

source Do Things that Don't Scale

Paul Graham · published Jul 2013

With few users you can do by hand what you plan to automate later, which lets you launch faster and tells you exactly what to build when you do automate.

You'll have muscle memory from doing the work yourself, so you know exactly what to build.

Scope: only when you have a small number of users

source Do Things that Don't Scale

Paul Graham · published Jul 2013

Solving users' problems manually until you can gradually automate the bottlenecks is preferable to the more common failure of having automation that solves nobody's problem.

Manual solving is frightening but less so than having something automatic that doesn't yet solve anyone's problems.

source Do Things that Don't Scale

Paul Graham · published Jul 2013

Any initial strategy that omits extraordinary founder effort — a big launch, a big partner — is by that fact suspect.

It's not enough to do something extraordinary initially; you have to make an extraordinary effort.

source Do Things that Don't Scale

Paul Graham · published Jul 2013 · hedged

Startup ideas should be conceived as vectors — what you will build paired with the unscalable things you'll do initially — rather than as scalars.

The need for unscalably laborious work to get started is nearly universal, and the pairing reminds founders they must work hard in two dimensions.

Scope: in most cases the second component is just recruiting users manually and delighting them

source Do Things that Don't Scale

Paul Graham · published Jul 2013

Early founders fall into imitating the flaws of big companies, such as indifference to individual users, because it seems more professional.

They want so much to seem big that they copy big-company behavior.

Scope: at the beginning

source Do Things that Don't Scale

David Lieb · published Sep 21, 2022

The biggest lesson from founding a startup is an ownership mentality: since there is no one else, you must personally do everything required.

A startup begins with nothing but a big vision, and getting from nothing to that vision requires the founders to do all of it themselves at the beginning.

Scope: especially in the earliest days

5:57 20 Product: The No 1 Metric You Need To Look at When Building Product | Why the Best in Product Have No Domain Experience | Why You Should Not Hire From Incumbents & The Difference Between Good vs Great PMs with David Lieb, Visiting Group Partner @ Y Comb

Sanjit Biswas · published Dec 8, 2023

Willingness to brute-force unscalable manual work has repeatedly been the unlock in his projects and companies

Scarce resources force you to find a way; the high school roster typed from a yearbook and installing antennas on roofs for Roofnet both built the platform when nothing else would

Scope: drawn from early-stage projects with no money

5:39 20VC: $18BN Market Cap and $1BN in ARR in 8 Years; Samsara | How to Find Product Market Fit Reliably | How to Create a Multi-Product Company | The Pros and Cons of Serial Entrepreneurship with Sanjit Biswas, Founder & CEO @ Samsara

Harry Stebbings · published Sep 13, 2024 · hedged

In the zero-to-one phase of a company, brute force effort is really important

He slept in the office when he started, in the Elon Musk mold

Scope: specific to zero-to-one stage

55:26 20Product: What Facebook, Monzo and Deliveroo Do and Do Not Do To Build Great Products | How to Structure Product Teams For Success | Is Simple Always Better in Product and The Art vs Science of Product Design with Mike Hudack

Luke Harries · published May 23, 2025

Founders should be shameless and manually DM or email their entire personal network to boost a launch

Launches need a big initial boost for the algorithms to care, and any founder who has raised money already has thousands of reachable contacts in Gmail, Twitter and LinkedIn

Scope: rebuttal to founders who say this only works with a big team and famous investors

26:15 20VC: ElevenLabs Head of Growth on Why You Do Not Need PMs | The 7-Part Launch Playbook That Gets 700K+ Views Per Product | The Truth About CAC, Payback & Performance Marketing in AI with Luke Harries

Harry Stebbings · published May 23, 2025

Founders are far too nervous to ask for help and to do unscalable manual work like personally DMing every follower to cross-promote channels

He personally DM'd his first 50,000 Twitter followers to cross-advertise the newsletter and that is how it reached several hundred thousand subscribers

27:01 20VC: ElevenLabs Head of Growth on Why You Do Not Need PMs | The 7-Part Launch Playbook That Gets 700K+ Views Per Product | The Truth About CAC, Payback & Performance Marketing in AI with Luke Harries

Cliff Weitzman · published May 9, 2026

Founders underrate the sheer number of product iterations required, and the way to get them is direct high-friction founder contact with users

He watched Stanford users click the wrong buttons with duct tape over his mouth, routed the in-app help button to his personal iMessage, and personally called every churned user — painful but it surfaced the real problems

79:25 20VC: What I Learned from 100 of the Best CEOs in the World | What I Learned from Staying with Mr Beast for 3 Weeks | How We Will Spend More on Tokens than Salaries with Cliff Weitzman, Speechify

Paul Erlanger · published Jun 27, 2026

The way to scale a consumer product to its first thousand users is to talk directly to users and iterate on their feedback at each order of magnitude — 10, then 100, then 1,000.

Direct early feedback from passionate users compounds product quality fast; giving traders a week of early access to fomo's web app roughly doubled how good it was, with no payment or strings attached.

Scope: Paul notes crypto/on-chain users are unusually passionate, which he treats as a competitive advantage

7:49 20VC: How We Got Fred Wilson, Benchmark and Index to Invest $94M | Why Robinhood's Strategy is Wrong | Why 1-1s are BS and What Every Founder Gets Wrong About Equity | Why Taste Beats AI But How AI Kills Org Charts with Paul Erlanger, CEO @ fomo

Sanjit Biswas · published Dec 8, 2023

At some point the founder must stop doing the unscalable hands-on work because they become the bottleneck, and turning it into a scalable process is the key unlock for the next stage

Continuing to do it yourself slows the company down; scaling requires 10x-ing and then 10x-ing again

6:38 Stop once the founder becomes the bottleneck

Paul Graham · published Jul 2013 · hedged

If the market exists, a startup can start with manual user recruitment and gradually transition to less manual methods.

source Manual recruiting gradually gives way to scalable methods as the market develops

Paul Graham · published Jul 2013

In a startup's first months, 'insanely great' should apply to the experience of being a user rather than to the product itself.

Jobs meant insanely great about an already-established Apple's product; early on the product is only one component of the experience, and attentiveness can compensate for an incomplete, buggy product.

source Insanely great should describe the user experience not the product itself early on

Paul Graham · published Jul 2013

In the best case the unscalable early work is not merely a necessary evil but permanently improves the company by shaping its DNA.

Aggressive early user acquisition, manufacturing your own hardware, or delighting a handful of users become habits that persist at scale.

source Unscalable early work shapes lasting company dna not just temporary necessity

Paul Graham · published Feb 2009

In the earliest stages it pays to offer customer service at a level that wouldn't scale, because it's a way of learning about your users.

Customers are used to being maltreated by quasi-monopolies, so surprisingly good service overwhelms them, and unscalable service teaches you about users.

source Unscalable early customer service is a tool for learning about users