Y Combinator Co-founder Paul Graham: If the Product is Good Enough, Focus on Customer Acquisition
Paul Graham, co-founder of Y Combinator, stated in a post that during office hours in mid-August, he told a startup that their product was good enough and they should stop coding to focus entirely on customer acquisition. He mentioned that the startup took this advice to heart.
During the same period, he publicly reiterated his office hours logic multiple times: giving founders permission to do simpler things than originally planned, often phrased as "You don’t need X, just Y"; when faced with multiple paths to growth, optimizing growth rates and letting data decide, the winning path might be a fourth option not yet considered. Each batch interview lasts about five minutes, directly addressing what the company is currently doing. He does not disclose the name, industry, or subsequent growth rate of the company.
"Good enough" is a judgment relative to customer acquisition costs, not a completed feature list. Stopping coding without sales and channel capabilities merely leaves engineering gaps filled with silence. Taking the advice to heart does not equate to having validated paying customers.
The YC system turns advice into a product: office hours yield executable subtractions. The unnamed case cannot be replicated into a universal formula.
In market mechanisms, buyers are seed-stage companies seeking growth curves, while sellers are accelerator mentors providing permission and attention. The driving force is the individuals’ posts, not financing terms. Funds continue to be consumed according to the original financing plan, but work hours shift from warehouse to funnel. Beneficiaries are teams that can immediately convert advice into weekly growth, while those under pressure still measure progress by lines of code submitted.
Source: Public information
ABAB AI Insight
Paul Graham frames office hours as an authorization ritual: allowing less to be more. The phrase from mid-August, "Stop coding, go get customers," is a reiteration of his frequently used "You don’t need X, just Y" in a real-name follow-up. By not naming the company, the story can circulate as doctrine. Growth rates serve as judges, acknowledging that the path may be a fourth unknown option, preventing founders from treating the whiteboard as strategy. The five-minute interviews compress sympathy, leaving only one step. As a YC alumnus, he pressures all teams still refining their products with public posts: functional debt can pause, but customer acquisition debt cannot.
The capital path is the reputation of the accelerator. The post does not introduce a new fund, only a redistribution of attention. The motivation is to pull the graduating class out of product narcissism; the strategy is to make "taking it to heart" a demonstration for the next batch of office hours. Seed-stage cash is reallocated between engineering salaries and customer acquisition experiments; misallocation will not appear in this tweet.
Analogous concepts include Steve Blank's customer development, lean startup's building testing, and Jobs cutting features. Graham's version is harsher: code itself is declared the current biggest waste. The phase shifts from demonstrable products to repeatable customer acquisition.
This is a transfer of pricing power. What counts as work is being redefined. The mechanism is: a mentor's statement reduces the value of submission records below that of customer acquisition records. The internal currency of engineering teams depreciates, while sales and growth experiments appreciate. Companies that take the advice to heart will prove it with next week's data; those that do not will continue refining until they run out of money.
ABAB News · Cognitive Laws
Once the product is good enough, additional code often serves to avoid customers.
The most expensive delivery from a mentor is allowing you to do less.
Once growth rates become the judge, the roadmap must yield.