Don’t Enter the Market. Redefine It.
Why the strongest companies escape incumbent comparisons and set new measures of value.
Nobody “finds” a great market. You build one, or you take a seat in somebody else’s.
It’s easy to end up in someone else’s market. “Searching” for product-market fit casts you as an explorer hacking toward a golden idol already buried in the jungle: the fit exists, it’s out there, and reaching it is a question of enough machete work. Tweak the messaging, adjust the pricing, and a hungry crowd surfaces.
I’ve defended that search, and I still do. Steve Blank built a methodology on it. The Four Steps to the Epiphany (2005) describes a startup as a temporary organization looking for a repeatable business model, and running one as a small version of a big company kills it. The customer development loop the book introduced is now the standard for how founders talk to buyers before they build, and it deserves to be. I’ve made the case for accepting the search and for grounding that phase with execution metrics.
The Map You Search Inside
The Four Steps to the Epiphany says startups are either in existing markets, resegmented markets, or new ones.
Blank encourages founders to figure out which market type they’re in, as though it were something the market had done to them. It’s a decision made early, often accidentally, and then defended for years by everyone who joins afterward.
Confined to a market that already exists, most founders reach for segmentation to carve out a slice of it. Sort a population by firmographics and geography and you get a TAM slide the board will nod at. That slide describes the world as it is. Whether anyone in that world will listen to a new claim is a question segmentation has no way to ask.
When you pursue an existing segment, you inherit the buyer’s mental models along with it. You enter an existing budget line. You accept a price ceiling somebody else set. A database buyer measures you against Oracle. A CRM buyer measures you against Salesforce. You’re selling on ground somebody else surveyed, and your win-loss reports come back saying “feature gap” and “price” more often than you’d like.
Why the Funnel Can’t Fix It
The reflex at that point is to work the funnel. Rewrite the headline, rebuild the demo, run the test again. A buyer who hasn’t been given a way to read a new thing will supply one himself, and he’ll reach for whatever costs him the least thought: you’re a cheaper Oracle, or you’re a line item he isn’t going to defend to anyone. Every test you run gets scored inside that sentence.
Funnel work can’t reach the problem because of where it sits in the buyer’s decision. A funnel measures conversion inside a frame somebody already set, and the setting happened upstream, out of view. By the time anyone reaches your pricing page, the comparison set has been chosen, the budget line has been picked, and your champion already has a sentence she uses to describe you to her CFO. A better headline might move her through that decision faster while the decision itself stays closed.
So the gains are real and capped. Each round of testing tunes the message toward the buyers who already understood you in the incumbent’s terms, since those are the people converting and producing the data. Anyone who might have responded to a different story never appears in the results because nobody told them one. What you get is a well-tuned machine for recruiting the customers most likely to grind you down on price.
Building the frame instead starts with a specific claim about how the world has changed and uses that claim to redraw the context your product gets judged in. Positioning leaves the product alone; it changes the surroundings until your product is the obvious conclusion. Three moves do that work.
Name the Enemy
The enemy is the status quo: the friction your customer already resents and has stopped mentioning, because mentioning it never helped. It’s a way of working that has to end. Name a competitor instead and you’ve validated the competitor, then spent the rest of the sales cycle arguing inside its frame.
At Heroku, the market that already existed was obvious and terrible. Developers wanted somewhere to host Ruby apps, which meant a commodity knife fight against every VPS provider, settled on specs per dollar. We went somewhere else. Our enemy was infrastructure management, the whole body of sysadmin chores that had nothing to do with shipping software. Nobody was calling it an enemy at the time. Developers assumed servers were part of the job, the way you assume weather.
Say Why Now
Name the change in the world that makes the old way untenable. This is the hinge, and it’s where designed markets mostly fail, because the founder reaches for a shift that turns out to be a wish.
A real one passes a test you can run before writing a word of copy: somebody who doesn’t work for you can verify it, and it happened before your pitch. If your why-now is “buyers are waking up to the need for this,” you don’t have one; you have a hope with a date attached. If it’s a change in what something costs, or a capability that shipped and didn’t exist eighteen months ago, a skeptic can check it.
Heroku’s was sitting in public. Rails had made web applications fast to write, and EC2 had made servers rentable by the hour since 2006. The distance between how quickly you could write an application and how long it took to get one running had grown absurd, and anyone paying attention could see it. We never had to convince a buyer the shift was real. We had to name what it meant.
Describe the Promised Land
Sell the new way of working. Features are evidence for it, and they only get interesting once the buyer wants the world you’re describing. Latent irritation becomes a market requirement when somebody names it out loud and gives buyers the language for it.
Heroku’s promised land was that you deploy code and stop thinking about servers. We made time-to-deploy the measure, and the incumbents had no answer for a yardstick they’d never thought to compete on.
A yardstick only counts once it leaves your marketing. Ours left through documentation. In 2011 Adam Wiggins, one of Heroku’s founders, published the Twelve-Factor App, a plain description of how an application should be built so it can be deployed and scaled without ceremony. It read like engineering guidance instead of positioning, which is precisely why it traveled. Teams that never bought anything from us adopted it. Competitors built products against it. The assumptions underneath it, disposable processes and config kept in the environment and no server anyone nurses by hand, became the default way a generation of developers thought about running software. By the time the category had a name, Platform as a Service, the argument had been settled somewhere upstream of any sales call.
Salesforce ran the same play on a bigger stage. The market that existed was people who wanted cheaper on-premise CRM. Marc Benioff’s company launched its on-demand service in February 2000 and sold it against packaged software itself: the eighteen-month implementations and the upgrade cycles and the IT tax. The web had made hosted software workable, and that was the why-now. The promised land was that you log in and start working. The red “no software” circle was a thesis compressed into a logo.
I’ve written about Mesosphere at length elsewhere. The part worth adding here is the pricing. Selling a Data Center Operating System instead of a cluster scheduler moved us against a different budget line, with a different approver and a much higher ceiling. The reframe held for about eighteen months, and then Kubernetes was free and we weren’t. Winning the argument about what a thing is doesn’t settle who gets paid for it.
The Play Isn’t Confined to Software
Dietrich Mateschitz found Krating Daeng, a Thai tonic sold to truck drivers and laborers, and saw a product with no Western market. By Red Bull’s own account of its founding, the market research he commissioned came back brutal: testers disliked the taste, the logo, and the name. Those are reasonable objections if the thing on the table is a soft drink.
Mateschitz changed what was on the table. He reformulated and carbonated it, launched Red Bull in Austria in 1987, priced the slim can well above cola, and seeded it with students and extreme-sport athletes. The price did argumentative work, since nothing costing triple a Coke is ordinary refreshment. The users did the rest, demonstrating in public what the thing was for. “Energy drink” became a Western category because Red Bull handed buyers a way to understand a product that had been sitting in Thailand for years.
What the Bet Costs
A designed market is a falsifiable bet on reality, and that’s what makes founders flinch the most. Slot a product into an existing category and you can post mediocre numbers for years without ever learning whether the premise was wrong. Every miss gets charged to execution, so you hire two more reps and run the quarter again. Design a market and the verdict arrives early and in public. You’ve named a shift in the world, built the pitch on it, priced against it, and told everyone what to expect. If the shift isn’t real, buyers won’t argue with you; they’ll just fail to show up, and that silence is legible to your board and to every competitor watching.
There’s a line here worth being honest about, because from the outside the move looks identical to spin. Reframing works when the product can actually deliver on the new measure. Heroku could genuinely take deployment down to seconds. Red Bull genuinely did something to you that a cola didn’t. Move the yardstick onto ground you can’t hold and buyers find out in their first quarter of use, and you’ve taught the market a vocabulary that now indicts you. The frame has to be true before it’s useful.
Keeping the bet honest takes the same discipline the search phase demands: write down what has to be true, decide in advance what evidence would change your mind, and hold yourself to the list. What changes when the market is designed instead of inherited is the exposure. Your hypothesis stops being private. It’s on your home page, and your competitors have read it.
The other cost is that you can’t hand this off. The organizational temptation is to push go-to-market downstream to a VP of Sales and get back to the product. But a VP of Sales optimizes inside whatever frame he inherits, because that’s the job you hired him to do. Deciding what the company means sits outside that job, and it stays with the founders. Two questions get you started, and you have to answer them about your own technology: who loses if we win, and what happens to our customer if they don’t buy from us. No feature list has ever answered either one.
Slotting into an existing category usually does get you revenue faster, and it locks you into the incumbent’s price. That trade looks like traction on a monthly board update and reads as a race to the bottom two years later.
None of this makes the outcome safe. Mesosphere named its category, held the frame for a year and a half, and lost the market anyway. What the work buys is narrower than the headline examples suggest: a market that’s yours to lose rather than somebody else’s to defend. Whether that trade is worth making is the founder’s call, and it has to be made while the pitch is still being written.