Home - Product Management Resources

Nearly half of the companies that invent a product category do not live to see someone else win it. In a study of almost 500 brands across 50 categories, marketing professors Peter Golder and Gerard Tellis found that market pioneers failed 47% of the time, while the companies that entered later and eventually led those categories failed only 8% (Journal of Marketing Research, 1993). Being first was not an advantage. It was closer to a coin flip with worse odds than the people watching from the sidelines. I keep coming back to that number when a founder or a head of product tells me their whole strategy is speed. "We have to be first." No, you almost never have to be first. You have to be right, and being first makes being right much harder, because you are guessing about a market that has told you nothing yet. The number that should settle the argument The Golder and Tellis finding is worth stating precisely, because the precise version is more damning than the summary. Across those 50 categories, pioneers were still the market leader in only about 11% of them. The companies that did lead, what the authors called early market leaders, held an average market share of 28%, roughly triple the pioneer's 10%. And here is the detail that reframes the whole thing: those winning early leaders entered the market on average 13 years after the pioneer created it (Journal of Marketing Research, 1993). Thirteen years. The winner was not a fast follower nipping at the pioneer's heels six months later. In many categories the winner showed up more than a decade after the category existed, watched what worked, and executed better with a clearer picture of what customers actually wanted. You already know the anecdotes even if you have never seen the data. AltaVista was the first serious web search engine, launched by Digital Equipment Corporation in 1995, and for a while it was the best tool on the internet. Google arrived in 1998 and passed it by early 2001 (EM360Tech). Friendster launched social networking as we know it in 2002 and hit three million users in months before its infrastructure buckled; Facebook launched in 2004 and defied the first-mover mantra outright (Friendster, Wikipedia). The pattern is not rare. It is the base rate. Why the myth survives anyway If the data is this lopsided, why does "first-mover advantage" still get treated as a law of nature in strategy decks? Because the earlier research that built the myth quietly deleted the losers. The studies that found a pioneer advantage tended to survey brands that still existed at the time of the survey. A pioneer that failed in year three was not around to be counted, so it never entered the dataset. What looked like "pioneers usually win" was actually "pioneers that survived long enough to be surveyed tended to be strong," which is a completely different and nearly useless statement. Golder and Tellis fixed this by reconstructing category histories from the beginning, including the companies that died. That single methodological change flipped the conclusion. This is survivorship bias doing what it always does, and product teams are not immune to it. When you study the winners in your market and reverse-engineer their moves, you are looking at the survivors. The graveyard is full of companies that made the same moves and lost, and you cannot interview a graveyard. I wrote about this trap in product discovery separately, because it quietly corrupts more than roadmaps, but it corrupts strategy first: see why survivorship bias eats your discovery. The three reasons being first actually pays First is not always wrong. It is conditionally right, and the conditions are narrow and knowable. The most durable framework for this is still Marvin Lieberman and David Montgomery's 1988 paper in the Strategic Management Journal, which identified exactly three mechanisms that let a first mover keep the lead (Strategic Management Journal, 1988). I treat them as a checklist. If none of them apply to you, being first is just paying to educate your future competitor. Technological leadership that compounds. If getting there first puts you on a learning curve or a patent position that a follower cannot cheaply copy, first can hold. The key word is compounds. A feature the next team can rebuild in a quarter is not technological leadership. A cost advantage that widens every month you operate is. Preemption of a scarce asset. If being early lets you lock up something genuinely limited, prime shelf space, the best supply contract, a spectrum license, the last good location, then the follower arrives to find the good seats taken. Most software has no scarce asset to preempt, which is one reason first-mover advantage is weaker in software than the industry that most worships it believes. Switching costs you can actually build. If your early customers accumulate real cost to leave, data they would lose, workflows rebuilt around you, integrations wired in, then a later and even better product has to be better by a margin large enough to overcome that friction. This is the one product teams have the most control over, and the one they most often neglect. I made the case that lock-in is a countdown rather than a moat unless you keep earning it (switching costs and the product moat), but real switching costs, honestly built, are the strongest reason to want to be early. Run your situation against those three. Most "we have to be first" arguments survive contact with none of them. What actually predicts the winner If not timing, then what? The MIT Sloan analysis that grew out of the same research names five factors behind enduring leaders, and not one of them is arrival time: vision of the mass market, persistence, relentless innovation, financial commitment, and asset leverage (MIT Sloan Management Review). Look at how ordinary those are. Vision, persistence, sustained investment, continuous improvement. These are executional muscles, not a starting-gun position. Procter and Gamble did not invent the disposable diaper; it out-executed the companies that did. Gillette's razor lead came from decades of refusing to stop improving the blade, not from being first to shave. Diet-Rite was the first diet soda; Diet Coke owns the shelf. Every one of those winners let someone else pay for the market's education, then showed up with capital, distribution, and a product shaped by everything the pioneer got wrong. Distribution deserves its own line here, because it is the factor most often confused with product quality. The best product frequently loses to the better-distributed one, which I have argued at length elsewhere (distribution beats product). A fast follower with a distribution engine beats a pioneer with a better idea more often than any of us would like to admit. The test I run before betting a year on being first Running network operations at a large telecom taught me the follower's discipline in a setting where being first is genuinely dangerous. When a vendor shipped a major new platform release, the pressure to be an early adopter was real: better features, a good story for the executive review, a chance to look ahead of peers. The operators who lasted learned to let another carrier deploy first and absorb the field failures we would otherwise have discovered live, at 2 a.m., on our own customers. We were rarely first. We were reliably second, and second knew exactly which bugs to avoid. The cost of being first was paid in outages, and the follower got the fix for free. That is the same trade every product bet contains, just with the failure hidden behind slower feedback. So before I let a team commit a year to being first, I make them answer four questions honestly: Does being first lock in a compounding technical lead, a scarce asset, or a real switching cost? If no to all three, first buys you nothing durable. Are we prepared to fund market education, the expensive work of teaching customers a category exists, and watch a follower harvest the customers we taught? Do we have the persistence and capital to still be improving this in year five, or are we hoping to win in year one? If a well-funded team copied our validated idea eighteen months from now, with none of our dead ends, what stops them? If those answers are shaky, the stronger move is usually not to abandon the market. It is to let someone else define it, watch what the market rewards, and enter with a sharper product and a distribution plan. That is not timidity. Golder and Tellis put a number on it, and the number is 8% failure against 47%. Speed still matters. Speed of learning matters enormously, and I would never tell a team to move slowly. But speed to market as a strategy, being first for its own sake, is the most expensive way I know to teach your competitors what to build. Being right beats being first, and the companies that understood the difference are the ones still on the shelf.

Recent Posts