Stewart Butterfield did not set out to build a workplace messaging platform. He set out to build a video game called Glitch. When Glitch failed in 2012, his team had spent years building internal tools to coordinate remote work, and those tools turned out to be more interesting than the game. Slack launched in 2013 and acquired its first hundred customers mostly by calling in favors, getting friends at other companies to try it, and then watching those companies beg for more seats. The go-to-market strategy they had written down was largely irrelevant.
This is not a cute origin story. It is a pattern so consistent across surviving startups that at this point it deserves to be treated as a law.
The Setup: Everyone Has a Launch Strategy
Picture a founding team six months from launch. They have a slide deck. Slide nine is the go-to-market strategy. It involves a few channels: maybe content marketing, maybe a product hunt launch, maybe a partnership with a distributor in their target vertical. They have done the math. If conversion rates hold and traffic scales, they hit a hundred customers in four months.
This is not delusional. This is how you’re supposed to think about it. VCs expect the slide. Accelerators teach the framework. The problem is not the thinking. The problem is that the market has not read the deck.
The customers you planned to reach are not paying attention. The channels you banked on are noisier than your models assumed. And somewhere, in a corner of the problem you were not focused on, a small group of people is doing something that looks almost like what you built, wants exactly what you built, and has no idea you exist.
What Actually Happened: The Notion Case
Notion launched multiple times before it worked. The first version, in 2013, got traction with some early adopters and then fell apart, partly because the technical architecture could not support the product vision. The team nearly ran out of money. Ivan Zhao and Simon Last moved to Kyoto, cut their burn to almost nothing, rebuilt the product from scratch, and relaunched in 2018.
The plan for finding users involved product directories and tech press coverage. What actually happened was that a small number of power users got genuinely obsessed and started making YouTube tutorials and Notion template galleries on their own, with no coordination from the company. The team noticed this was happening and began feeding it: featuring community templates, engaging directly with the people making content, building features those users specifically needed.
The first hundred real, retained customers did not come from press. They came from three or four people who cared unreasonably much and pulled in their networks. Notion’s job was to recognize that signal, not manufacture a different one.
This is the thing that distinguishes the teams that survive. Not that they stumble onto a customer acquisition channel by luck, though luck is always in the mix. What distinguishes them is that they are paying close enough attention to notice when something is working that was not in the plan, and honest enough to abandon the plan.
Why It Matters: The Map Is Not the Territory
Founders spend weeks on go-to-market strategies because the alternative, admitting that you genuinely do not know how you’ll acquire customers until you try, is uncomfortable to sit with and nearly impossible to put in a pitch deck. So they build a plan that sounds reasonable and then treat it as more load-bearing than it is.
The survivorship issue here is brutal. The startups that stuck rigidly to their original channel strategy when it wasn’t working are mostly gone. You don’t hear about them. You hear about Slack pivoting from a game studio, about Notion grinding through a rebuild in Japan, about companies like GitHub, which grew almost entirely through developer word-of-mouth in communities its founders were already part of, because the founders were just scratching their own itch and talking about it.
The common thread is not a brilliant acquisition strategy. It is founders who were embedded enough in the problem space to be in the right conversations when serendipity happened, and self-aware enough to notice.
There is a related failure mode worth naming: founders who find an early channel that works and over-index on it before they understand why it works. Early customers from a specific community or referral chain feel like validation of a strategy when they are often validation of a relationship. Your first hundred customers should be teaching you something, not just filling a spreadsheet.
What We Can Learn
The practical lesson from Notion, from Slack, from GitHub, from almost any company that survived its first year with real customers, is not that planning is useless. Planning is useful for forcing you to articulate assumptions. The moment those assumptions meet reality, though, your job is to update.
A few things that actually seem to separate the survivors:
They started with a network, not a funnel. The first hundred customers almost always come through direct relationships or the one or two degrees away from them. This means founders need to be willing to do the awkward, unscalable work of personally recruiting users. Butterfield called people. Zhao talked to every early Notion user he could find. There is no automation for this phase.
They paid attention to who was already banging on the door. In almost every case, there is a customer type that showed up differently than expected, asked for things you hadn’t built, and stuck around anyway. That customer is telling you something. Ignoring them because they do not match your ICP slide is a mistake you will pay for.
They treated the first channel that worked as a hypothesis, not a conclusion. Finding one acquisition channel that converts is the beginning of a question: why is this working, who are these people, and is this a foundation or a trap? The loudest customer in your inbox is not always the right one to follow, and the same logic applies to channels.
They did not wait for permission to change direction. The teams that survived were willing to say, out loud, that the plan wasn’t working while they still had time to adjust. This sounds obvious. In practice, it requires a specific kind of honesty that is genuinely hard when you have told investors and your co-founders and yourself a story about how this was supposed to go.
The mythology around startup growth tends to collapse the messy, contingent process of finding real customers into a clean narrative after the fact. Slack didn’t set out to pivot from games. They set out to make a good game, failed, and paid close attention to what they had left. That attentiveness is the skill. The channel, the ICP, the deck, those are all placeholders until the market tells you otherwise.
