In one of the go-to-market sessions at Draper U, the speaker wrote three short lines on the board:
- Who you sell
- What you sell
- How you sell
He then said: do not reverse the order.
At the time it sounded suspiciously tidy, the sort of framework that behaves itself a little too well on stage. It only became useful later, when I started bringing it back into actual product and commercial choices. The uncomfortable part is that it is not really about neat sequencing. It is about how early you are already making trade-offs.
Who you choose to sell to changes what you are actually selling. What you sell changes how you have to sell it. And once those three get misaligned, a lot of later work is just a more elaborate way of deepening the mistake.
That is one reason I have become less persuaded by the line that a market is attractive simply because it is big.
This is not because TAM is irrelevant. Investors do care about ceiling. But a large market has never been the same thing as a good starting point. The harder early-stage question is usually narrower than that:
with the product you have now, which slice of the market can you actually win?
The most useful thing I kept from that session was a framework that felt plain enough to survive contact with reality:
Ease of win × Size of prize
It is not telling you to choose the easiest customer. It is not telling you to lunge at the largest one. It is forcing a harder optimisation problem: which segment is relatively winnable, and still meaningful if you do win it.
That logic is not far from how serious GTM operators talk about segmentation. OpenView has long argued that useful customer segmentation is not a branding exercise. It is about grouping prospects whose needs and buying motions are similar enough that one message and one motion can work on them efficiently. Bessemer’s founder guidance pushes in the same direction from another angle: before you expand, define an ICP that feels almost uncomfortably narrow. The common thread is simple. You are not looking for everyone who could buy you. You are looking for the group that is worth concentrating your company around first. (openviewpartners.com) (openviewpartners.com) (bvp.com)
My own mistake for a long time was treating GTM as something that comes after product-market fit.
Build the product. Then work out how to sell it.
That sounds reasonable. In practice it is often too late. The first customer segment shapes the product more than founders like to admit. It shapes the kind of value claim you can make, the sort of evidence you need, the sales motion you can afford, the onboarding burden you inherit, and the cadence of cash into the business.
The framework from the session is useful precisely because it keeps those decisions linked.
First, choose who. That clarifies whether you are really selling a feature, a workflow improvement, a solution, or a measurable business outcome.
Then decide what. Only then does it become clearer whether the right motion is product-led, founder-led, assisted sales, or direct enterprise sales.
Plenty of startups do the reverse.
They build features first and hope a market will emerge. Or they imagine themselves selling to enterprise buyers and start speaking in enterprise language long before the product or proof structure can support it.
The result is usually familiar: the message gets softer, the feature set gets broader, and eventually the company starts telling itself that lots of customer types could use it.
That is not breadth. It is blur.
What I find more practical now is to think about GTM as three linked decisions:
claim, proof, and channel.
What are you claiming to the customer. What evidence is required to make that claim credible. What selling motion is realistic given the cost of generating that evidence.
Those three things cannot contradict each other.
You cannot sell a simple feature and pair it with a heavy enterprise motion unless the economics somehow justify it. You cannot sell ROI with almost no customer-specific evidence and expect a light-touch motion to carry it. Value claims move upwards in abstraction, but so do proof demands and sales complexity.
That was one of the sharper points in the session. The stronger and more strategic your value proposition becomes, the harder it usually is to prove and therefore the harder it is to sell cheaply.
That is not a messaging problem. It is a transaction-structure problem.
Once you are selling ROI, you need customer data. You need buy-in from more stakeholders. You need to survive more objections. You often need to carry a longer sales cycle. And you need to be able to deliver what you promised once the deal closes.
This is where the phrase right to win becomes more useful than TAM.
Not whether you could theoretically build for that market. Not whether it is a prestigious segment. Not whether the logo would look good in fundraising.
The question is whether, given your current product maturity, competitive landscape, buyer inertia, brand position, and delivery capability, you have enough reasons to win that segment now.
That is a harsher standard than market size. It is also more honest.
The speaker also made another point I have kept: markets are not abstract pools of demand. They are wallets. The money already exists somewhere. If you are going to win a budget, you need to know which budget it is.
Are you taking spend from headcount. From agency work. From incumbent software. From error correction and rework. From a broader efficiency or revenue target.
If you cannot answer that, your GTM usually becomes fluffy very quickly.
This is one reason Bessemer’s framing of ICP tends to hold up. Your ideal customer profile is not simply the customer you most want. It is often the customer you can already make successful, who will actually stay, expand, and become a source of signal rather than just short-term revenue. (bvp.com) (bvp.com)
That is a less romantic definition than founders usually prefer.
It often means your best current market is not the fanciest one. Not the segment with the prettiest logos. Not necessarily the highest ACV.
It may simply be the set of customers you can genuinely win and serve without distorting the whole company.
That last part matters. One of the more dangerous early-stage self-deceptions is to treat any large custom deal as strategic validation. Sometimes it is. Quite often it is just expensive confusion. The roadmap bends, support gets heavier, the sales motion warps, and the company spends a year serving a handful of customers that look impressive from the outside while quietly teaching the product the wrong lessons.
This is why OpenView’s older work on segmentation still feels relevant. Segmentation is not just market research. It is a growth-efficiency discipline. Sharper segmentation improves message, pricing, motion, and eventually product focus as well. (openviewpartners.com)
None of this makes the framework universal.
Some categories do require hard customers first. Heavy implementation, regulatory products, design-partner-heavy enterprise software, and serious change-management tools often cannot be learned properly from the easiest segment. In those cases, pure ease-of-win logic can send you into a shallow market that teaches you the wrong thing.
So the real question is not whether you should always start with SMB or whether enterprise is always a mistake.
The more useful question is:
given the product, resources, brand, and proof you have today, which segment gives you the highest right to win now and a believable path into the next one later.
That question is irritating, because it forces founders to admit that some ambition may still be early.
But it is better than the alternative.
If I had to reduce that whole session into a working test I would actually use, it would be this:
Before asking how large the market is, ask:
- what value this segment is actually willing to buy on
- what level of proof that claim requires
- whether you can afford the selling motion that proof implies
- whether winning this slice makes the next slice easier or just messier
If those questions are still fuzzy, the problem usually is not that the market is too hard. It is that the GTM is still underthought.
That was the correction for me.
I used to think GTM was what happened after the product was built. Now I think it starts much earlier than that.
Early enough that if you choose the wrong segment first, a lot of later execution is just high-quality service for the wrong market.
Series | What Reality Corrected