The More Screens Customers See, the Sooner You Need a Product Manager
A simple chat box bred few arguments, while apps spread across phones, browsers, and other companies' software create conflicts only a dedicated referee can resolve.
First Round · When Should You Stop Clinging to Product and Hire Your First PM? Advice From a Founder and CPOIn June 2025, OpenAI employed thousands of people but only about 30 product managers, the people who decide what to build and in what order. The reason, argues Saumil Mehta, a former startup founder who later led product work at the payments company Square, is that the first version of ChatGPT was essentially a box on a screen.
Mehta learned the rule from Alyssa Henry, who managed him at Square and who was comparing two very different products. Square's system for processing payments, she told him, "is an inch wide but a mile deep," while its marketing and loyalty tools for small shops "are a mile wide but an inch deep." Mehta says founders should act accordingly.
By surface area, he means the number and complexity of screens a customer can touch. A web-only product with a few screens and a lot of hidden complexity can run with few product managers, because most day-to-day choices can be made by engineers. A simple product that lives on iPhone, Android, and the web, and that pulls data in and out of other companies' software, may need a dedicated manager much earlier, even if the underlying technology is simpler.
That is because every visible screen is shared territory. Prospects evaluating the product, current customers using it, salespeople demoing it, marketers rewriting its words, designers arranging its layout, and the engineers who display and supply its data all feel entitled to an opinion. More screens also mean more chances for odd combinations, bugs, and visible failures.
And opinions about what customers see run strong. Mehta recalls an engineer who insisted on dark mode, the white-text-on-black-background look, and a designer who told him, "Because everyone can see the design, everyone thinks they can be a designer." The job, then, is gathering those competing views and turning them into one plan, which is what product managers are built to do.
The test played out for one founder of a young artificial intelligence company building software agents that resolve employee computer problems. “We’re covering a lot of surface area and third party integrations, especially given our enterprise customers, likely more so than most AI startups at our stage,” he said, so he did what felt counterintuitive and hired a product manager much earlier than expected. Count the screens and integrations, in other words, and you can tell how soon you will need a referee.
Questions people ask
What does a product manager actually do day to day?
They collect requests from customers, salespeople, designers, and engineers, study how people actually use the product, and decide what the team builds next and in what order. They also write down exactly how a feature should behave, settle disagreements about the design, and track whether the change worked.
Do connections to other companies' software count as surface area?
Yes. Each link means your product must handle someone else's logins, data formats, and unexpected software updates, plus the screens to control them. Those links add stakeholders, opinions, and ways to break, which is why founders with many enterprise integrations often hire earlier.
If my technology is very complex, do I automatically need a product manager sooner?
Not necessarily. A product can be technically profound, like processing payments or running a large language model, yet have only one or two screens. In that case engineers can make many decisions themselves, and a founder can wait longer before adding a product manager.