Photo by Aditee Bhargava
There are two opposite mistakes founders make with this question, and both are common enough that neither should be treated as the safer default.
The first is starting too late, treating GTM as something to think about once the product is essentially finished, as if distribution is a problem to solve after the real work is done. The second is starting too early, building an elaborate GTM plan before there’s any real signal about who the customer actually is, optimising a strategy against assumptions that haven’t been tested against a single real conversation yet.
Starting too late is the more visible failure, because it produces a recognisable, painful moment: a finished product, genuine confidence in what’s been built, and silence where customers were supposed to show up. The mistake isn’t that GTM thinking happened after building; some sequencing of effort is unavoidable. The mistake is that GTM was never thought about as a parallel discipline at all, developing alongside the product, informed by the same early conversations that were shaping product decisions. By the time the product is finished, an enormous amount has been learned about the customer that never got translated into a distribution plan, because nobody was deliberately capturing it for that purpose.
Starting too early is less visible and, in some ways, more wasteful, because it produces something that looks like progress without actually being it. An elaborate GTM strategy built before a single real customer conversation has happened is built entirely on assumption, however well-reasoned. Channel sequencing, messaging, pricing, all of it optimised against a guess about who the customer is and how they decide, dressed up with enough confidence that the team starts treating the guess as settled fact before it’s been tested even once. The plan looks complete. It’s actually fragile in a way that won’t become visible until real customers behave nothing like the plan assumed they would.
The right starting point sits earlier than most founders expect, but in a much rougher form than the word “strategy” implies. The moment there’s a working hypothesis about who the customer is, even an imprecise one, is the moment to start having real conversations with people who might fit that description, specifically to test and sharpen the assumptions a GTM plan will eventually depend on. This doesn’t require a finished product. It requires enough of a concept to have a genuine conversation about the problem, the alternatives the person currently uses, and what would actually make them switch. That input should be shaping the product simultaneously, not arriving after it as a separate workstream bolted on at the end.
Here’s the part that gets skipped even by founders who get the timing broadly right: having early conversations isn’t the same as having evidence, and the gap between the two is where a lot of “we did our customer discovery” GTM plans quietly fail anyway. Five encouraging conversations feel like validation. They’re often measuring something closer to politeness than intent, because a conversation costs the other person nothing, and people are generically agreeable when a founder is visibly excited about something they’re building. The conversation that actually counts as evidence is the one that costs the other person something to have honestly: a commitment to pay, a willingness to be a reference, an introduction to someone else who’d need this, a specific date they’d actually switch by. A GTM strategy built on five warm conversations is still built on assumption. It’s just assumption with better vibes attached, and vibes don’t survive contact with a real launch any better than a guess does.
The second thing that gets missed is that the right moment to formalise a GTM strategy isn’t a single gate a company passes through once, early, and never revisits. It’s a threshold that has to be re-earned every time something material changes: a new pricing tier, a new feature that shifts who the product is genuinely for, a new market or city the company starts operating in. Founders who correctly waited for real evidence before their first GTM strategy often skip that same discipline the second and third time, because it feels like the company has already “graduated” past the customer-discovery phase. It hasn’t. It’s just accumulated evidence for a version of the customer that may no longer be the one actually buying, and a GTM motion built on stale evidence fails in exactly the same way one built on zero evidence does, just later, and with more money already spent defending it.
Picture two founders building competing products in the same category, six months apart in timeline but identical in ambition. The first spends four months building in near-isolation, launches with a polished GTM deck nobody has pressure-tested against a real buyer, and discovers in week one that the pricing, the channel, and the core message all assumed a customer who doesn’t actually exist in the form the deck described. The second spends those same four months building a rougher product in parallel with fifty real conversations, a dozen of which involved someone actually agreeing to pay before the product was ready, and launches with a thinner-looking GTM plan that happens to be built entirely on things real people already did, not things they said they might do. The second founder’s plan looks less impressive in a deck. It’s the only one of the two that was ever actually tested before the money went out the door.
A useful check that goes further than the standard version: has anyone outside the founding team who fits the intended customer profile had a real, substantive conversation about the problem this solves recently, and did that conversation include something that actually cost them to say yes to, or is the current GTM thinking still running on encouraging conversations that never asked for a real commitment? A second, quieter check worth adding to that: when did the company last re-verify that its GTM assumptions still describe the customer actually buying today, rather than the one it validated at the very beginning?
GTM thinking shouldn’t wait for the product to be finished, and it shouldn’t get formalised before there’s anything real to base it on, and it shouldn’t stay formalised forever once it’s earned that status the first time. The right moment is earlier and messier than most founders expect, and it comes around more than once.
Every business has its own version of this story. If you're working through something similar, I’d love to hear from you. Whether it's to exchange ideas, brainstorm a challenge, or just have a thoughtful conversation, feel free to reach out at [email protected].