Most of a company's design is decided in its first eighteen months, by people who would never call themselves designers, in conversations nobody writes down.
I have been the founder making those decisions three times, and the executive arriving afterwards four times, at Honor, Lyft, Grammarly and Pitch. An unusual pair of vantage points, and mostly a source of retrospective embarrassment. Here is what I did not understand as a founder, and could not stop noticing as an executive. By the time a company hires someone to own design, the decisions that determine what the product feels like have already been taken. And they are load-bearing.
What founders actually decide
Not the visual language, which is the part everyone thinks of as design and the part most easily changed later. The decisions that stick are the structural ones, and founders make all of them.
Who the customer is, when there are two plausible answers. At Honor we assumed our customer was the adult child arranging care for an aging parent. We went to market and roughly half the people signing up each day were arranging care for themselves or a spouse. Different person, different device, and a completely different frame of mind. We had built the whole sign-up for the wrong one. That is a design decision and it was made long before it was examined.
What the product refuses to require. We committed to Honor working as well over a landline as over a smartphone. It sounds like an accessibility footnote. It was the most constraining decision in the company's history. You cannot assume a screen, so the model of what is happening has to survive being spoken aloud, and that forces a clarity a user interface will happily let you avoid. Every downstream choice inherits it.
What the first ten customers are allowed to ask for. A founder saying yes to a specific enterprise buyer in month six writes a roadmap constraint that will still be visible in the product in year five.
None of these show up in a design review. All of them are design.
The first designer does not fix this
A standard move is to hire a designer to own design, with relief, at around fifteen or twenty people. It rarely works the way founders expect and the reason is not the designer.
The taste is already baked in. It is in the onboarding flow the founder built in a weekend, in the pricing page they argued about, in the shape of the object model, in what the company thinks is embarrassing and what it thinks is fine. A new person can change the surface. They cannot change the encoded judgment without an argument with the founder, and that argument is exactly the one the hire was supposed to make unnecessary.
So either the first designer gets real authority over those structural decisions, which puts the founder in exactly the working relationship they thought they were hiring their way out of. Or they are handed the surface and slowly become a production function. That second outcome is far more common, and usually misdiagnosed as a hiring mistake.
A founder is going to do the design either way. Only one question is open. Whether they do it deliberately.
What I look for now
I sit on the other side of this now, running Noord. The founder-design question has become one of the more reliable things I can assess in a first meeting.
I am not looking for a founder with taste, exactly. Taste is pleasant and it is not predictive on its own, and plenty of tasteful founders build nothing. What I am looking for is whether the team can put something in front of a person this month.
When we founded Human, everybody on the founding team wrote code or pushed pixels. I now think that was the clearest thing about us to the angels who wrote the first checks. Not because building it yourself is romantic. Your story will change, your market will change, your product will certainly change. What an early investor is really betting on is how fast you get through those changes.
That tell is small in a meeting. Ask what they shipped in the last two weeks (two weeks, not a quarter, because a quarter lets them answer with a plan). Founders who do this work themselves answer in specifics and often show you something on a phone. Founders who have outsourced their own product judgment answer with a plan.
The failure on the other side
I should be even-handed, because founder-led design has a characteristic failure and it is severe.
Founder taste never gets audited. A company grows. The founder's instinct keeps winning arguments on the strength of having been right in year one. Nobody can tell the decisions that were correct from the ones that were merely early. Everybody in the room learns to predict the founder rather than the customer, which is a much easier problem and produces a company that is excellent at pleasing one person.
You can spot it. A product with unusual coherence and one strange blind spot, always in the same place, always somewhere the founder does not personally go. This is where a design leader earns the job, and advocacy is the wrong tool for it. What works is something that can overrule the founder's instinct without being more senior than the founder. Research the founder actually attends. And a habit of writing decisions down with the reasoning attached, so a wrong one can be found later.
What I would tell a founder
Name the decisions that are design decisions and keep them. Who the customer is when there are two plausible answers. What the product refuses to require. What it will never do. What the first run feels like to someone who has no idea what they are looking at. That list is short, it is yours, and delegating it early is how companies end up with products nobody can describe.
Delegate everything else as fast as you can, and treat the first design hire as a partner in the short list rather than an owner of it. Be explicit that this is the arrangement, because the alternative is a competent person spending eight months discovering it by collision.
Plenty of very large companies were built by founders with no design instinct at all, who bought the capability afterwards at considerable expense. So this is an advantage rather than a requirement. I would still rather back the founder who can put the thing in front of someone this week.
And write down why. Not for posterity. Write it for the version of your company that is forty people. Somebody reasonable will ask why the landline decision exists, and the honest answer needs to be better than because the founder wanted it.
I've founded three companies. I was on the founding executive team at Honor, leading product and brand from seed to Series C, and led design or product at Lyft, Grammarly and Pitch. Today I run Noord, an AI Design Lab that occasionally invests in startups.
© 2026 Renato Valdés-Olmos