Pick a Door. Any Door.
The market you enter determines the buyer you need, the story you tell, the metrics you watch, and the budget you burn.
A go-to-market plan usually arrives secondhand. A founder names three companies she admires, somebody reconstructs how those companies launched, and the plan gets assembled out of the parts that worked: the content cadence, the pricing page, the conference booth, the analyst briefings. The tactics are described accurately. What travels with them, unannounced, is a set of assumptions about the market the original company was standing in, and those assumptions are what made the plan work.
There are four ways in. You can enter a market that already exists and try to outperform what’s there, competing on features against incumbents who set the terms of comparison. You can resegment that market on price, taking the low end who’ll accept good-enough performance in exchange for paying a lot less. You can resegment it for a niche, serving a subset the incumbents ignore, sometimes at a higher price and often with worse performance on the attributes that subset doesn’t care about. Or you can create a market where your customers couldn’t do the thing at all before. Steve Blank laid this out in The Four Steps to the Epiphany in 2005. His categories are existing, resegmented, and new, with resegmentation splitting two ways. He reckons more than half of startups walk through one of the two resegmentation doors.
Each door changes the target, the argument, the clock, and the budget at once. Enter an existing market and you’re selling to the incumbents’ customers, who want better performance and will score you against a list they already have. Resegment on price and you’re selling to the cost-sensitive end of that same base, at the risk of the incumbent deciding to match you. Resegment for a niche and you’re telling a specific group that everything else on the market was built for somebody else, at the risk that the group turns out to be too small to live on, or not distinct enough to buy any differently from the mainstream. Create a market and you’re selling against people doing nothing at all, at the risk that adoption never arrives.
I’ve argued before that nobody finds a market; you build one or take a seat in somebody else’s. The four doors are what that choice looks like in operational terms, before anyone writes a plan.
The Door Is a Choice
Blank frames the four as a diagnosis, a question you answer about a market that already is what it is. Three things you control move a market from one type to another.
Price moves it. Come in far enough under the incumbent and what would have been an existing-market fight becomes a low-cost resegmentation, which brings a different buyer, a different motion, and a ceiling on what you can ever charge.
How wide you draw the problem does the same work. Narrow it to a subset nobody serves and you’re a niche entrant. Widen it past anything a buyer can currently purchase and you’re creating a market, whether or not you meant to.
The third lever has no slot in Blank’s scheme at all. It’s the buying center you enter through, meaning which person inside the customer you actually sell to. There are always at least two. One is whoever will use the thing every day. The other controls the budget and signs the contract. Blank’s four types quietly assume you’re selling to the same buyers the category has always been sold to.
The Fifth Door
Sell to the person who uses the thing rather than the person who signs, and the market type moves underneath you. Payments in 2011 was a thoroughly settled category, with incumbent processors, a known budget line, and a procurement process every finance team already knew how to run. Stripe sold to the developer, who had never been the buyer of anything. Everything on the seller’s side of that deal was ordinary existing-market business, and the buyer was somebody the category had never once pitched.
Those two halves want opposite plans, and running both at once is the part nobody budgets for. The category hands you named competitors, a price the market already understands, and a problem no buyer disputes is worth solving. That’s ordinary existing-market work, and at least it’s legible. You’re arguing you’re better than a specific alternative in front of people who agreed long ago that the category deserves a budget line. What you build for it is competitive positioning, a differentiation claim that survives a bake-off, and a clean migration path off the incumbent.
The buyer hands you somebody who can’t sign a contract, has never sat through a vendor pitch, and will make up his mind by signing up and trying the thing well before he talks to anyone. That’s new-market work, and it costs differently. He can’t be sold to in the ordinary sense, so the documentation becomes the pitch and the free tier becomes the demo, and the product has to close itself with nobody in the room. He also can’t buy. Adoption shows up before revenue does, and at some point you have to go find the person with the budget and sell her a tool her own staff installed months ago.
So you’re funding two motions at the same time, and in most companies they report to different people. The education work sits with docs and developer relations, measured on adoption. The competitive work sits with product marketing and sales, measured in pipeline. Each side can look at the other’s budget and see something optional, which is the argument that surfaces the first time you have to cut. Cut either one and the machine stops. Lose the education and nobody arrives; lose the competitive work and the people who did arrive can’t get a purchase order approved.
Two Quarters of Expensive Confusion
Most developer-tools companies live in this position. Datadog entered a monitoring market with entrenched incumbents, Twilio sold into telephony, a category older than software itself, and MongoDB went after Oracle. The category was settled in every case and the buyer was brand new, and the plans that come out of that combination look broken to anyone trained on enterprise go-to-market. The documentation is public and ungated, so no leads come off it. There’s a free tier that on paper eats revenue. Sales gets hired late and hired to expand accounts the product already won rather than to source new ones. The pipeline report reads thin, because the signal that matters is sitting in usage data the CRM never sees.
Drop a conventional enterprise plan on top of that and every change in it is defensible on its own. Gate the documentation to capture leads. Retire the free tier to protect the average selling price. Hire SDRs to call the person who controls the budget and buy the analyst placement while you’re at it. What breaks is the top of the machine. The practitioner who would have found you through the docs and tried you for free never arrives, and he or she was the entire acquisition motion.
It takes a couple of quarters to see because revenue lags badly here. Practitioners who already installed you keep converting into paid accounts on their own schedule, so the numbers hold while acquisition quietly stops. By the time the revenue line bends, the cohort that would have fixed it stopped showing up two quarters earlier, and the plan that caused it has been running long enough to look like the status quo.
One Team, Two Doors
Donna Dubinsky ran Palm Computing when the Pilot shipped in 1996 and was CEO of Handspring when the Visor shipped in 1999. The two go-to-market plans were opposites, and both were right. At Palm the job was convincing people a handheld was worth wanting at all because nobody was shopping for one. By 1999 buyers compared handhelds on specifications, and Handspring differentiated on expandability and performance in a market Blank describes as already worth a billion dollars. Handspring running Palm’s evangelism would have spent a year teaching buyers what they already knew, and Palm running Handspring’s feature comparisons would have burned its cash arguing against products nobody was cross-shopping.
Blank calls 1996 a market that didn’t exist, which is tidier than the record. Handhelds had been on sale since the Psion Organiser in 1984 and had simply never sold in volume. What changed between 1996 and 1999 was buyer literacy: whether people knew what the thing was for, what to compare it against, and which budget it came out of. Neither Palm nor anyone else bought that literacy; it accumulated over four years of people watching a colleague pull one out in a meeting.
Your Dashboard Can’t Tell You Which Door You Walked Through
The uphill direction of travel is running a new market on existing-market assumptions. You fund demand capture when there’s no demand yet to capture. You buy search terms nobody types. You staff a sales team against a buying process that hasn’t formed. None of it looks like waste while it’s happening, because every line item is something a competent company does.
The obvious correction is to run the plan and watch the numbers, adjusting when they disappoint. But early success in a new market looks just like failure in an old one. Low volume. Long cycles. Prospects who engage warmly and never convert. Read as an execution problem, those numbers call for more spend and more reps. Read against the right market type, the same numbers are on schedule, and the correct response is to protect the runway. No metric tells you which situation you’re in, and the dashboard will quietly recommend the wrong one.
Each door has its own objective, and the numbers only mean anything once you know which one you’re walking through. In an existing market the objective is share: a slice of spending that already exists, taken from somebody who currently has it. The work is capture against a buying process that already runs. In a resegmented market it’s share and education at once, which is harder than either alone and also the most common way in. In a new market there’s nothing to take and nobody to take it from, so the objective has nothing to do with share. The measure is whether the number of people who could plausibly buy is moving toward something a business could live on.
The Flattering Answer Is the Expensive One
Picking a door is supposed to be a judgment about the market. It’s also a judgment about which conversation you’d rather have with your board. “New market” buys a longer runway and more patience, and it’s a claim a founder can always argue for, because nobody can disprove it inside a year. “Existing market” obliges you to explain how you’ll beat a funded incumbent with a tenth of their go-to-market budget. The answer that’s easiest to defend is also the most expensive one to be right about.
Money settles what the analysis can’t. Add up what each door costs, set it against what’s in the bank, and where two types both fit, take the one you can pay for. Each carries a different bill. A low-cost resegmentation needs a cost structure that survives the incumbent matching your price, and they can usually hold a lower price longer than you can. A niche resegmentation needs a subset both large enough to live on and cheap enough to reach because a small audience can’t absorb a big acquisition cost. A new market needs enough runway to pay for the education before anybody buys. When the only type you can afford isn’t the one you want, you have a financing problem wearing a market-type label, and it belongs in front of your board as a financing problem.
The Door Closes Behind You
A market can be new when you enter it and ordinary three years later. Nothing announces the change when it comes. So you keep spending to teach people what the category is for long after they’ve learned it, and you’re still funding evangelism while your buyers have started comparing feature lists.
So market type belongs on the quarterly agenda, next to pricing and segment, which you already revisit. Ask whether you’re still in the market you planned for. If you aren’t, ask what your objective and your spending should have changed to and how long ago.
The longer you wait, the less of a choice it is. Price, how wide you draw the problem, and which buyer you sell to are all yours to set before launch and expensive to change after. A published price fixes what the market thinks you’re worth. A buying center you’ve sold into for two years has headcount attached to it, comp plans written around it, and a content library aimed at it. You can pick the type while it’s still on paper. Once you’ve spent two years executing it, you aren’t choosing anymore, you’re living with it.