Updated September 2026. Originally published June 2024.
Here is a problem I have been working on recently, and it explains why most go to market plans fail before anyone writes a single piece of copy.
A buyer has virtualisation infrastructure. Broadcom’s acquisition of VMware has reset the licensing model to per core, and the increases being reported across the market are not small. Depending on environment size and who you read, customers describe multiples rather than percentages.
That buyer now believes something very specific: that their costs are going up several times over and they have to absorb it. They have usually confirmed it before speaking to anyone, by asking a language model, which returned a confident number.
They are not wrong about the increase. They are wrong that absorbing it is the only option. Staying put, switching hypervisor, moving to a managed private cloud, going public cloud, or running a hybrid all produce very different three year totals.
Now the go to market question. Do you tell that buyer their mental model is wrong?
If you do, you lose. Not because the argument is weak, but because you have made them do work before you have given them anything. That is the whole subject of this article.
What a go to market strategy actually is
You are planning how to present a solution to a market that will test it, question it, and judge it. The difficulty is almost never execution. It is preparation, and specifically it is how much of the buyer’s existing world you understand before you start talking.
The version of this article I published in 2024 had four steps and stopped, unfinished, in the middle of pricing. This version finishes it, and the step that was missing turns out to be the one that decides whether any of the others mattered.
Step 1: Market research, and the word that matters
Market research rests on two pillars: what your competition proposes, and what customers think they need.
I use the word think deliberately. At this stage, what customers believe they need outranks what they actually need, because their belief is what your messaging has to enter. Widely adopted products create assumptions about what a category is and what it should cost, even when your solution would redefine the need entirely.
You will get to redefine it. Just not in the first sentence.
Step 2: Define the audience
Who are your buyers? What has to happen inside an organization before they become a customer? Which roles are involved in the decision? Is there a size, a sector, a geography?
If you are new in a category you will not have enough of a sample to make confident assumptions, but there are two reliable ways to build one.
Read the demand your competitors are paying for
Keyword research is not only an SEO exercise. What people search before they land on a competitor tells you how the market describes its own problem, and paid keywords tell you what a competitor considers valuable enough to buy.
- Enter a competitor domain in a keyword tool
- Open their top paid keywords
- Sort by cost or by share of spend to see what they are protecting
- Filter out anything containing their brand name, so you are left with demand for the problem rather than for them
What this gives you is vocabulary and priority, not a target list. The distinction matters, and I set out why at length in the piece on keyword research and validated intent. Volume tells you how many people are curious. It does not tell you which of them can sign.
Where I no longer get personas from
The 2024 version of this article recommended pulling five profiles from a professional network, feeding them to a model, and asking it to merge them into a composite persona.
That is the output, and it reads well. That is the problem with it.
A composite built from public profiles tells you what people put on their profile. It does not tell you what they said when something broke, what they were told by their board, or which words they used to describe the problem before they knew your category existed. It produces a persona that is plausible, internally consistent, and untested.
What I do now is slower and worth it. Ask existing customers how they described the problem before they found you. Have whoever makes first contact record the language people actually use. Then use models for the part they are genuinely good at: prompting as the buyer, seeing what comes back, and reading what the answer cites. That last step tells you which sources your market’s understanding is currently being built from, which is more useful than any composite.
Use models to widen the search. Do not use them to invent the buyer.
Step 3: The value proposition, and the meeting point
This is the part of the original article I would keep unchanged if I could only keep one thing, and it has become more important rather than less.
When you are going to market, perceived value decides everything, and perception starts from where the buyer already stands. If your positioning requires them to first accept that they have been thinking about the category incorrectly, you have charged them for the privilege of listening to you.
The structure I use looks like this. Back to virtualisation:
- What the customer thinks they need: to keep paying for the platform they standardized on, at the new price
- What competitors propose: the same platform, resold, with the increase passed through
- What your solution actually does: removes the dependency entirely, with a different cost base
- Your original value proposition: you should not be paying per core for a hypervisor at all
- The meeting point: here is what the next three years cost if you stay, and here is what they cost on each alternative
The meeting point is where you communicate your value without disconnecting from what your buyer already has in mind, and without contradicting what you actually believe.
Notice what the meeting point does not do. It does not tell them they are wrong. It accepts their frame, which is cost over a period they already care about, and then puts the alternatives inside that frame. The comparison does the arguing. You do not have to.
Why prompting made this matter more
In 2024 a buyer’s mental model was usually vague. They had a general sense of the category, assembled from vendors, peers, and whatever they had read.
Now they ask a model first, and they arrive with something much harder to move: a specific answer, often a number, delivered confidently, that they did not have to work for.
Two things follow. Models synthesise consensus, which means the answer a buyer arrives with usually reflects the dominant framing of the category, which is usually the incumbent’s framing. And a position that arrives pre-formed and externally validated costs more to dislodge than a vague impression ever did.
So the meeting point is no longer a courtesy. It is the entry condition. You either enter the frame your buyer already holds, or you spend your whole first interaction arguing with an answer they trust more than they trust you.
I have written more about how positioning gets expressed across different levels in the value proposition article.
Step 4: Marketing and sales, which are the same thing
Webinars, events, collateral, content, social, email, search, paid. Everyone knows the channel list, and listing it again is not useful. What matters is that your strategy is downstream of the three steps above. Your marketing strategy is, in the simplest terms, how you communicate the results of your research, your positioning, and your audience work.
Sales is not a department
I am not a fan of the sales floor and marketing floor distinction. Treating them as two departments with separate goals is an outdated way to run a company. Sales is not the name of a team. It is what every function is trying to produce. When that is understood across an organization, the output of every role becomes more purposeful, because everyone can see which end they are serving.
The clearest evidence that the split is outdated is account based marketing. Most companies still assign marketing the early touches and bring sales in further down. That confuses the buyer and produces an inconsistent voice, because two groups with different targets are talking to the same person about the same decision. Assign accounts to a group of people responsible for the outcome and the KPI argument disappears, because there is nothing left to divide.
Choosing channels
Being everywhere at once would be lovely. Budget, timeline and goals will not allow it. The most practical prioritization I have used is to go back to the audience work: once you genuinely know who you are talking to, the shortlist of places to reach them is usually obvious and usually shorter than the plan you started with.
Pricing
Pricing is the most argued over part of go to market. The standard approaches are well known: cost plus, competitive, penetration, skimming, dynamic, value based.
My disagreement is not with any of them individually. It is with picking one before analyzing the specific circumstances of the offering. You can run several as hypotheses and find out which holds. The failure is applying a named strategy without the research underneath it.
While cheap and expensive are subjective and depend on your buyer’s means, cheaper and more expensive are not.
There is an inverse relationship between how much cost matters and how strong your perceived value is. The better the value your market perceives, the more room you have to price where you want to.
Four questions to settle it:
- Does my solution solve a significant problem my competitors cannot? If so, do I have the budget and the team to communicate that clearly?
- What actually drives my competitors’ customers, cost or value?
- Do I have better brand credibility than my competitors?
- Are any of my competitors using prestige pricing successfully, and if so, on what basis?
One more, which the virtualisation example makes obvious: what unit is my buyer already measuring in? If they are thinking in three year totals, a monthly price is not a lower price. It is a harder comparison, and harder comparisons lose.
Step 5: Measurement, which is the step I left out
The original version of this article stopped at pricing. That was the real gap, because without this step a go to market plan cannot be judged, only defended.
Two things have to be settled before launch, not after.
An ICP that sales, marketing and the CEO have all signed
Write down exactly who counts. Company size, sector, geography, exclusions, and who has to be present in the meeting for it to be a real meeting. Get it agreed by everyone whose numbers depend on it.
This looks like bureaucracy and it is the single highest leverage thing in a launch. Once the definition is agreed, the expected close rate on a qualified meeting becomes a constant. If the rate is not achieved, that is a sales question, not a negotiation about whether marketing sent the right leads. Without the agreement, every disappointing quarter becomes an argument about definitions, and those arguments are unwinnable because both sides are right under their own version.
Numbers that exist in a P&L
Not impressions, not MQLs. Cost per appointment, cost per appointment held, cost per customer acquired, recurring revenue added, and contribution after delivery cost. Those are the numbers a CFO recognizes, and I have set out how I build and defend them in what one sales appointment is worth.
Decide them before launch. If you choose your measures afterwards you will choose the ones that flatter the result, and so will everyone else in the room.
If you are building one now
- Find out what your buyer already believes. Not what they should believe. What they will arrive holding, including whatever a model told them last week.
- Enter their frame before you challenge it. The meeting point is the entry condition, not a softening of your position.
- Take vocabulary from customers, not composites. Models are for widening the search, not for inventing the buyer.
- Price in the unit your buyer is already using. Their comparison period, not yours.
- Agree the ICP in writing before launch. With sales and the CEO, not just within marketing.
- Choose the measures before you can see the result. Otherwise you are not measuring, you are narrating.
Go to market is not a document. It is a set of decisions about whose reality you are willing to start from, and most plans fail because they start from the seller’s.
This article was substantially rewritten in September 2026. The measurement step is new, as is the material on why buyers now arrive with pre-formed positions. The persona method from the original, which merged public profiles into a composite using AI, is no longer recommended and the reasoning is given above. The meeting point framework and the position on sales and marketing alignment are unchanged from the original. The virtualisation example refers to widely reported market conditions following Broadcom’s acquisition of VMware; no client, employer or vendor engagement is described.