Most people begin a startup in the wrong place. They find an idea, decide it is clever, make a prototype, and hope the market agrees. Theo's argument is blunt: the idea does not make the company. Customers do.
"You shouldn't be asking another founder if it's a startup. You should be asking potential customers if it's a thing they would use and pay for."
That sounds obvious until you look at how founders actually spend their time. It is much easier to ask a smart friend whether an idea sounds promising than to ask a prospective customer how they work today, what they pay for, and whether they would switch.
Ideas are cheap. Motivation and pain are not.
At the opening, Theo pushes back on the myth that successful companies spring from a brilliant concept. A good idea helps, but it cannot carry a founder through the hard, repetitive part of company building. The work only holds up when the founder cares enough about the problem to keep going and other people care enough to change what they do.
That distinction matters because investor interest is not the same as demand. A pitch can sound good. A prototype can look good. Neither proves that someone has a painful enough problem to pay, change behavior, or keep using the product.
Ask for the workflow, not validation.
Theo suggests replacing the question "is my startup idea good?" with a much more useful conversation.
- Find the people who would pay for the thing.
- Ask how they do the job today, where it breaks, and what it costs them.
- Offer a rough solution, even if it does not scale.
- Watch whether they use it, return, and ask for more.
Do the thing that does not scale
The video’s most useful example comes from Ping, Theo's video-call company. The team wanted to know whether creators valued separate, high-quality audio and video tracks after a call. Building the finished version would require infrastructure, transcoding, storage, and time. It could take months.
So they did not build it first. Theo put together a rough system that captured the needed packets, ran an FFmpeg script on his own machine, and sent a Google Drive link to the user. It was absolutely not how a real company should operate forever. That was the point.
The manual workaround exposed two things that a polished feature would not have answered on its own: whether people wanted the result, and whether the quality was good enough. Once users showed that they needed it, the team had a real reason to invest in building the durable version.
Customer discovery is not pitching
Theo gives concrete examples. If you are making a Kubernetes tool, do not ask a founder you admire whether it sounds good. Ask Kubernetes developers what goes wrong in their day and whether your proposed tool would solve it. If you are making a grocery-store inventory product, go to a grocery store. Learn how its staff manages tags, what it pays today, and whether it would try another option.
That means resisting the urge to sell too soon. Rather than asking a creator, "Do you want to try Ping?", Theo asks how the creator collaborates with editors, what they used for a past collaboration, and what felt painful about the process. The answers shape the product better than a yes or no to a pitch ever could.
Product-market fit is a customer relationship
For Theo, a startup becomes real when customers need the product enough to make the path toward product-market fit visible. It may take one hundred conversations to find the first person who is willing to try. Those conversations are not wasted when a first product idea is wrong. They create relationships with the people who can point the team toward the right problem.
He makes the standard even higher: when you cannot find users, be your own user. Theo uses Ping for his own interviews every week. That constant use finds bugs, reveals bad workflows, and forces the company to rebuild core pieces around real behavior rather than internal guesses.
A side project becomes a startup when customers pull it forward.
Build enough to learn. Speak to the people with the problem. Do the manual work while it is still cheap. Then invest in the polished system after the demand is real.