Selling is the hard part

A small team can now ship in weeks what used to take a year. The code, the design and the infrastructure arrive faster than ever. The customer does not. Someone still has to find the person with the problem, earn their attention, and convince them to change how they work and pay for it.

We build companies for a living, and this is the lesson we keep relearning. The product is rarely what decides whether a company survives. The selling is.

Why good products still fail

The people who study startup failure keep arriving at the same place: the product is usually not what breaks.

Peter Thiel put it bluntly in Zero to One: “Poor sales rather than bad product is the most common cause of failure.”

Tom Eisenmann of Harvard Business School adds an important nuance. The most common early mistake he sees is a “false start”: founders begin building and selling before confirming the problem is worth solving. His answer is not to sell harder but to learn first, and he notes that this customer research can often be done in about four weeks.

Paul Graham made a related point in 2013: early on, founders have to recruit users by hand, and many avoid it, hoping a good product will find its own users.

Put together, the lesson is not simply “sell more”. It is to learn about customers early and keep learning, before the product is finished and after it ships. Building is the part founders are drawn to. Selling, and the customer understanding behind it, is the part that decides the outcome.

Why AI makes it harder, not easier

AI has made building dramatically cheaper. In one controlled experiment, by researchers from Microsoft, GitHub and MIT, developers using an AI coding assistant finished a single programming task 55.8% faster than those without one. One task is not all of software development, but our own experience points the same way: a small team now ships what used to need a much larger one.

AI is changing selling too. It helps with prospect research, personalised outreach and follow-up. What it does not do is earn a buyer’s trust or make their decision for them. A customer still has limited attention, a routine they would rather not change, and a budget someone has to approve.

Our view is that this widens the gap. When anyone can build a credible product in weeks, more products compete for the same customers, and the product alone stops being a reason to choose you. What becomes scarce is what AI does not shorten: understanding a buyer well, being trusted by them, and reaching them repeatedly through a channel that works.

What becomes scarce is what AI does not shorten.

Thiel’s line applies more today than when he wrote it: “If you can get just one distribution channel to work, you have a great business.”

What we learned the hard way

We learned this ourselves before we wrote it down.

We are builders first. Engineering, design and product are the parts of company building we are best at, and the parts we enjoy most. Selling is the part we have had to learn deliberately, across every company we have started.

People who love building tend to keep building: one more feature, one more improvement, one more reason to wait before asking someone to pay. It feels like progress. Often it is a way of avoiding the harder conversation with a customer.

The honest lesson for us was this: a founder’s interest follows traction. When a product sells, the work is energising. When it does not, even a good product starts to feel like a burden. Fixing the selling, and learning about customers as you do, is what keeps a company alive long enough to become good.

What we do differently now

At Cooperative we treat selling as part of the foundation, not something a company adds once the product is finished.

• We start where we already know the buyer. We build in markets we know first-hand, where we understand who has the problem, how they buy and what they already use. A product aimed at people we can actually reach starts with a distribution advantage.
• Go-to-market is part of day one. Every company we start shares the same foundation: legal setup, engineering, design, quality assurance and go-to-market. Selling is on that list on purpose.
• One sales system across every company. We are building our own CRM so one small team can run sales across all of our businesses in one place. The principle behind it: every company should inherit what we have already learned about selling, from repeatable processes and follow-up discipline to what works with buyers, without assuming every market or customer behaves the same way.
• We measure companies by customers, not features. A release that nobody uses is not progress. A conversation that ends in a paying customer is.

None of this makes selling easy. It makes it impossible to postpone, which is the point.

The work that decides the outcome

AI has made the part founders love faster and cheaper. It has left the part they avoid exactly where it was. That changes where the advantage sits.

The companies that last will not be the ones that build fastest. They will be the ones that know their buyer best and reach them most reliably. Building is how you make something worth buying. Selling is how you find out whether you did.

Sources

1. Peter Thiel with Blake Masters, Zero to One: Notes on Startups, or How to Build the Future (2014), quoted via Goodreads and Goodreads quotes.
2. Tom Eisenmann, “Why Startups Fail”, talk at Stanford eCorner, August 2021. See also his book Why Startups Fail (2021).
3. Paul Graham, “Do Things That Don’t Scale”, July 2013.
4. Sida Peng, Eirini Kalliamvakou, Peter Cihon and Mert Demirer, “The Impact of AI on Developer Productivity: Evidence from GitHub Copilot”, arXiv, February 2023.