2. Who holds it

§Every startup has someone who holds product, whether or not anyone has said so. If the role has not been named, it is being held by default, usually by the founder who talks the most in meetings. The first act of product ownership is to make the implicit explicit: one name, known to everyone, answerable for what gets built.

§ 2.1One of four people#

§At the earliest stage the holder is one of four people, and at the very earliest stage all four are the same person.

§The CEO holds it when the company's founding insight is a product insight. This is the common case. The person who saw the problem and decided to build the company is the person best placed to keep deciding what the product must do, and to hear every signal that tests the founding bet.

§The CTO holds it when the founding insight is technical: a capability that did not exist before, looking for the problem it solves best. Here the technical founder has to become the customer's advocate as well as the technology's, which is harder than it sounds and the subject of a later chapter.

§The COO holds it in operations-heavy companies, where the product is a service delivered by people and systems as much as a piece of software, and the person who runs the operation is the person who knows where it breaks. Telehealth, logistics and marketplaces often land here: the operating founder sees every failure first, and product is the discipline of turning those failures into the next thing built.

§A CPO, or someone with an equivalent title, holds it when the founders have decided that none of them should. This is rarer than the org charts suggest, and it works only if the CPO has the standing to overrule the founders on product questions. A product lead who reports to a CEO who quietly keeps the decisions is not holding product. They are holding a backlog.

§Sometimes the work is genuinely split: one founder runs discovery and customer conversations, another runs delivery with the engineers. That can work. What cannot work is splitting the accountability. When the two disagree, one of them decides, and everyone in the company knows in advance which one.

§ 2.2The T-shaped owner#

§Whoever holds product at a startup does not get to hold only product. They will be in sales calls, in the support inbox, in the design tool, in the standup and in the investor update in the same week. The shape that makes this survivable is the T.

§The horizontal bar is breadth: working knowledge of marketing, sales, support, design, engineering and finance, enough to hold a real conversation with a specialist in each and to know when they are being told something that does not add up. The vertical bar is depth: one discipline the owner is actually expert in, which is where their judgment is most trusted and where the company will lean on them hardest.

§In practice the vertical bar is set by history. A founder who spent five years in design holds product with a designer's depth. One who spent five years selling holds it with a seller's. Neither is better. What matters is knowing which it is, and compensating deliberately for the other bars of the T.

§Depth without breadth produces a specialist who decides product through the lens of one discipline: the designer-founder who over-invests in polish, the engineer-founder who ships architecture, the seller-founder who promises features on calls. Breadth without depth produces a generalist whose judgment nobody defers to when it counts.

§ 2.3Switching hats#

§The daily reality of the role is switching between responsibilities faster than a large company would consider sane. A morning might hold a customer conversation, a decision about what the first version leaves out, a support ticket that reveals a bug, a sales objection to answer, and a review of an engineer's estimate. Each of those asks a different question and rewards a different instinct.

§The discipline that makes switching work is simple to state and hard to keep: change the hat, never the accountability. In the sales call, the owner sells, and also listens for what the customer says they would pay for. In the support inbox, they answer, and also notice that three people this week hit the same wall. In the engineering estimate, they respect the number, and also ask which half of the feature could ship first.

§ 2.4The solo founder#

§More than a third of new startups now begin with one founder, and the tools that appeared in 2024 and 2025 are the reason it works. The category has its own institutions now: Julian Weisser's Solo Founders program in San Francisco backs a small cohort of single founders each year on the explicit premise that one person with the right tools can carry a company, and its existence is as good a signal as any that the solo path has stopped being an exception. For a solo founder the question of who holds product has one answer, and the rest of this chapter applies with the volume turned up. There is no horizontal bar of the T to borrow from a co-founder; the breadth has to be real, or bought from advisors and customers. There is no one to hand a hat to; the switching happens inside one head, and the accountability cannot be dropped because there is nowhere for it to fall.

§ 2.5What this looks like at the start#

§At two founders and one engineer, holding product means: one founder is named, they spend most of their week in contact with customers, they write down what they hear, and they decide what the next thing to build is, out loud, with a reason attached. That is the whole practice. Everything else in this book elaborates it.

§The next chapter is about the reason attached, and about the three things the holder is answerable for.