Startups fail when they try to compete head to head with big companies. It happens at the point Geoffrey Moore calls the chasm, and the way out is to stop copying the market leader and start specializing.
The copying starts for a reasonable-sounding reason. A startup at $1M or $2M in revenue has spent a year or two selling to early adopters, the people who buy because a product is new. Then those deals dry up. The next buyers want logos, a roadmap, and a security page, so the founder studies the market leader's website for clues about what "credible" looks like. The homepage gets wider. One use case becomes four. "Built for regulatory affairs teams in pharma" turns into "The document platform for modern enterprises."
Here's the thinking: "Big buyers want a big vendor, so we should look like one."
Here's what big buyers see: a 30-person company claiming to do what a 3,000-person company does, and doing less of it. Every sales call now opens with the same question, "So how are you better than [the leader]?", and there's no good answer, because the one thing the startup had going for it is gone from the page. It could build exactly what one customer needs and change it next month. The leader can't. And the homepage no longer says so.
Anthony Pierri and Leah Tharin have names for the two paths. The startup that keeps its advantage is a flexible specialist. The one that gives it up is an unstable generalist. To see why one crosses the chasm and the other stalls, you need to know who is buying on each side of it.
After the early adopters, a different kind of buyer shows up
Moore's book Crossing the Chasm came out in 1991, and the idea still holds. Early adopters buy because something is new and they want a head start. The early majority, who Moore calls pragmatists, buy because it's proven and people like them already use it. The gap between those two groups is the chasm. Moore calls it "by far the most formidable and unforgiving transition in the Technology Adoption Life Cycle."
The gap is there because the two groups want opposite things. Pragmatists, in Moore's words, "want evolution, not revolution. They want technology to enhance, not overthrow, the established ways of doing business." They also won't buy without hearing from people like them first, which Moore sums up as a catch-22: "The only suitable reference for an early majority customer, it turns out, is another member of the early majority, but no upstanding member of the early majority will buy without first having consulted with several suitable references."
Founders read that and decide the answer is to look bigger and safer. So they add personas, widen the homepage, and start saying "platform." That's the trap, and the next section is about why.
Startups win as flexible specialists and stall as unstable generalists
I first heard these two terms on Leah Tharin's podcast with Anthony Pierri of Fletch. Leah put it this way: "Startups have an advantage as flexible specialists, offering solutions that larger enterprises cannot provide." A flexible specialist is small enough to build exactly what one kind of customer needs, and to change it when that customer's needs change. An unstable generalist tries to serve everyone and ends up as wide as a big company without any of the size that makes a big company safe to buy from.
Here's the thinking that turns a startup into a generalist: "If we say we're for everyone, we won't scare off any leads."
Here's what happens instead. Pragmatists compare you to the market leader. The market leader has 200 more features and 10 years of references. You lose. Moore again: "the pragmatists are loath to buy until they can compare. Competition, therefore, becomes a fundamental condition for purchase." Since they're going to compare you either way, you want the comparison to happen inside a niche you own, where most of those 200 features don't matter.
Dunford calls this positioning style Big Fish, Small Pond, and it's her alternative to going head to head. The idea is to "carve off a piece of the market where the rules are a little bit different, just enough to give your product an edge over the category leader." The payoff, in her words: "it's easier to dominate a small piece of the market, compared to full on, head-to-head market domination."
Big companies can't afford to follow you into a niche
You might wonder why the market leader doesn't just build for your niche too. They won't, because big companies are built for volume. A niche of 40 accounts doesn't move the revenue line at a company doing $2 billion a year, so no product manager there will get the headcount to build for it. Dunford's one condition for this strategy is that "your competitive advantage must be difficult for the market leader to copy," and the cheapest way to make something hard to copy is to make it too small for them to bother with.
That's the whole advantage. Use it.
Documentum grew 35x by picking 40 accounts
This is the example Moore uses in the book, and the order of events is what matters.
In 1993 Documentum was stuck at about $2 million a year in revenue. Jeff Miller came in as CEO and picked one target: the regulatory affairs departments inside Fortune 500 pharmaceutical companies. That's the team that assembles a new drug's application for approval. About 40 companies. Roughly 1,000 people. A tiny market, on paper.
The pain inside that market was huge. A new drug application ran to hundreds of thousands of pages, and every day it sat unfinished was a day of patent life gone. Moore puts the value of a successful drug at around $400 million a year, so each day of delay cost more than most software budgets. His line on it: "In the case of the pharmaceutical industry's regulatory affairs function, the pain was excruciating."
Documentum built document management for exactly that job. Filings went out faster. The people who noticed were the executives who owned the drug's revenue, which is how Documentum got past the IT departments that would normally have picked a bigger, safer vendor.
Then came what Moore calls the bowling alley, where each niche you win knocks over the next one. Revenue went from $2 million to $8 million the year after Miller arrived, then $25 million, then $45 million, then past $75 million and a NASDAQ listing. Each new market followed the last: contractors who build and maintain plants, then regulated chemicals, then oil refining, and eventually trading desks on Wall Street. Moore's rule for picking the first one: "The size of the first pin is not the issue, but the economic value of the problem it fixes is."
Related post: The Essential Guide to Why, When, and How to Niche Down
Pragmatists buy the whole product, the references, and the price
Documentum worked because it passed three checks that every pragmatist runs before signing. A flexible specialist can pass all three. A generalist startup usually can't.
They check whether you solve the whole problem
Moore's term for this is the whole product. Early adopters, he writes, "are willing to take responsibility for piecing together the whole product." The mainstream is not. A specialist knows which integrations, templates, and workflows a regulatory affairs team needs on day one, because they've built for that team ten times before. A generalist ships the core product and a partner directory and leaves the rest to the customer.
They check whether you're a safe choice
Pragmatists are careful with young vendors, and the way past that is references from their own industry. Three pharma logos beat thirty logos from thirty industries. This part also gets easier with every deal. One case study gets you the next reference, a few references get you shortlisted by default, and eventually two regulatory affairs directors find out at a conference that they both use you.
They check whether the price matches the problem
A specialist costs less to run, because support and onboarding are the same job every time. That leaves room to price on what the problem is worth. Documentum could charge pharma companies real money because the alternative was another lost day of patent life.
The cost of skipping these checks shows up in the failure data. CB Insights' 2026 post-mortem of 385 failed VC-backed startups found that 43% died of poor product-market fit, and two-thirds of those were early-stage companies that never found a market at all. Trying to fit every market is a reliable way to fit none.
Your last ten deals tell you where to start
Pull your last ten closed deals and write down the industry, the department, and the problem for each one. If three or more share all three, that's your first pin, and your homepage should say so out loud. If none of them match, you don't have a positioning problem yet. You have a segment problem, and the homepage can wait.
The founders I see get through the chasm all did some version of this. They picked the small pond, won it, and only then went looking for the next one. The ones who stall tried to win the whole category on day one. 🍕
Subscribe to receive the latest blog posts to your inbox every week.
