Our framework
WEDGE: where to put the budget when there isn’t much of it
Five steps we use to answer the most awkward question in search marketing: where do you start when the money will not stretch to everything? The mechanics, with real examples — including the directions we turned down.
12 minutes · Applied to more than 20 companies over six years
Almost every client arrives with the same request: we need more good clients. Almost every agency answers the same way: we propose comprehensive promotion. And right there the budget starts getting smeared thinly over everything at once.
The trouble is not that broad promotion is bad. The trouble is that it isn’t an answer. It is a list of tasks. The question was different: where exactly do you put what you have so that it works?
A client always has less money and less time than the leader in the niche. Always. Try to win on volume and you have lost before you start, because the other side has more of it by definition. Precision is what is left.
Hence the name. A wedge is a narrow point where limited force splits something much larger. That is not a nice metaphor for a deck. It is a literal description of the job.
W
Where
Where you can win
The first step is yours, not ours. The client fills in a competence map — a spreadsheet where the company describes itself. Five tabs:
- Solutions — products and projects you have built: CRM, a dating app, a corporate messenger
- Verticals — industries you have worked in and are competent in: healthcare, agrotech, fintech
- Languages, APIs, frameworks — Java, Angular, Laravel, Postman
- Technologies — AR, VR, blockchain, AI
- Services — native iOS development, cross-platform, Big Data consulting
Plus a column for “what we are good at here” — the results that matter most in that direction. It is not there for scoring. It is there so that later there is something to write from.
What gets scored in each row
Every direction is scored 1 to 3, where 1 is worst and 3 is best. Four scores, and different people put them in.
- Expertise — the client: from “we stood next to it once” to “we will build anything here, and build it well”
- Sellability — the client: how good the clients you have already won in this direction were
- Margin — the client: what the company earns on a project like this
- Market growth — us: CAGR, the rate of change rather than the absolute size
Market growth deserves a pause, because it is the usual misunderstanding. What matters is the first derivative, not the magnitude. A market can be enormous and flat, in which case it is already carved up and you are late. A market can be ten times smaller and growing — and that is where there is room for a newcomer.
How the rows get ranked
Then a formula runs. What rises to the top is where everything lines up at once: the market is growing, the expertise is real, the clients can pay, the margin is decent. We don’t publish the coefficients — but we also don’t call this a secret method, because that would be a lie. The logic is open; the exact weights simply don’t matter for understanding the result.
One technical detail that does matter: the tabs are ranked separately, not merged into one list. We take the top of each, work it, and only then go to the intersections — more on that at the last step.
Why the top row is almost never the expected one
Here is how the Verticals tab might look for an IT outsourcer. The numbers are invented, to show the mechanics:
| Direction | Expertise | Sellability | Margin | Market growth | Rank |
|---|---|---|---|---|---|
| Healthcare | 3 | 3 | 3 | 3 | 1 |
| Fintech | 3 | 2 | 3 | 2 | 4 |
| E-commerce | 3 | 3 | 1 | 1 | 9 |
| Agrotech | 2 | 1 | 2 | 3 | 12 |
| Education platforms | 1 | 2 | 2 | 3 | 15 |
Look at the third row. E-commerce is what the company does most often, sells most easily, and names first when asked what it does. And it sits ninth. The margin is low and the market is flat: it is a commodity, the client picks on price, and there are hundreds of competitors.
Nearly every map has a row like that. The direction a company treats as its face turns out to be the worst place to put a marketing budget. That is the awkward answer the whole exercise exists to produce.
There is a worn-out phrase, “we will uncover your unique advantages”. We don’t like phrases like that, so, concretely: we find the row where you earn the most against the least competition — and very often it is not the row you would have named.
From practice: a direction the client closed himself
At Aimprosoft the map was built as its own task, and markets were then analysed monthly against it. One direction — Alfresco — was deliberately dropped from the priorities, and marketing had nothing to do with it. The client’s wording was short: it sells badly, and has done for a long time. The sellability score closed the direction before we ever reached competitive analysis.
That is the first screen doing its job. It is supposed to filter cheaply, before anyone has spent money on research.
Who scores what, and why it is arranged that way
The sharpest objection to the map goes like this: why take the client’s word for it? He will inflate his own scores.
The answer: because we physically cannot verify them, and pretending otherwise would be a con. Establishing how strong a company really is in fintech means going into the code and running a full technical audit. A marketing agency will not be let in there, and rightly so.
The second reason is simpler — the presumption that people are sane. Someone paying for promotion usually knows what their company earns from. It is not a thing worth lying to yourself about.
We do not audit your engineering and we do not pretend we can. Your read on yourself × our data on the market. Each side brings what it actually knows.
What we bring is what the client does not have: market movement, who already owns the results page, how much real traffic sits there, and whether you can get in with the site as it stands. That is not a compromise and not a weakness. It is a division of labour.
E
Evaluate
Whether you can get in
The map produced a ranked list. But it knows nothing about what is already happening in the results. So a second screen runs over it, and for each direction from the top we look at three things:
- competition in classic search
- competition in LLMs — who gets cited when the question is put to a model
- potential traffic: whether there is anyone there to catch at all
This is the reachability axis, and it is orthogonal to attractiveness: a direction can rank first on the map and still be unreachable with the current site. Two axes, two coordinates. Hence the matrix.
Built from a real 3D visualisation project. The dashed line is one direction moving between zones over four years.
What comes out of the second screen is a queue, not a single direction: what we take now, what opens in six months once authority has grown, and what we leave alone entirely.
A direction that waited four years
At Omega Render the 2020 map produced an interesting picture. Property developers and general contractors scored highest on margin — the most expensive projects, the longest contracts. And in the final ranking they landed somewhere in the twenties.
The client put the reason in words worth quoting: you cannot reach them through outreach or meetings, you have to establish yourself in the industry first. That is a reachability assessment, just made by a person rather than a tool.
So we went where we could: virtual staging, real estate agencies, the events side. The developers stayed in the queue. In April 2024, once the site had weight, the direction opened and we moved into it. Four years — and not one unit of budget wasted trying to break through something that would not break.
A direction that did not pass the screen at all
Successful examples convince nobody, so here is the reverse. On an education platform we looked at whether to run part of the directions in Ukrainian alongside English. The tools showed no demand.
But the conclusion in the second-screen document was not “we are not doing it”. It was more precise: the tools do not show demand for equivalent topics in Ukrainian — but that demand may not be absent so much as low, and hidden because it is low.
The distinction matters. “There is no demand” and “the tools cannot see demand” are different claims, and confusing them is expensive. The decision was a compromise: run it paired with English, where the English version carries the load and Ukrainian rides along cheaply. If the demand is hidden, we catch it. If it is not, we lost very little.
That is also the answer to why we trust no traffic tool blindly. In narrow B2B niches they are wrong regularly, and they are wrong mostly on the low side.
If a direction does not work out, we take the next one in the queue — that is what the queue is for. We walk away from a client when the whole map is closed, not one branch of it.
D
Decode
Who signs the contract, and what stops them
The direction is chosen. Now: who decides in that direction? In B2B it is almost never the person who ran the search. A CTO, a head of procurement, an owner — each has their own criteria, their own fears and their own vocabulary.
We research who that person is, what they already know, what they are afraid of, and what they need to see to take the next step. The result does not go into a deck. It goes straight into the content brief.
Here is what gets added beyond a standard SEO brief:
- which specific objection this page removes
- what proof it has to carry: numbers, diagrams, comparisons, other people’s cases
- the language this person speaks, and the words that will put them off
- what they should do next, and why now
The persona does not sit in a document
The usual fate of persona research is a PDF opened twice. Ours is different, and it is not a matter of discipline but of tooling.
Each project has its own GPT agent with that project’s personas built into it. Every brief and every copy check goes through it. That has been the case since 2023 — we put it in almost as soon as it was technically possible. On one project the agent also accounts for the axis from the competence map, which means step D is wired directly to step W.
There is a worn-out phrase, “we write for people, not search engines”. We don’t like phrases like that, so, concretely: the brief is written for a named persona — their pains, objections and selection criteria — at the moment it is created. Not for an algorithm.
G
Guide
Guide them to a decision
A B2B buyer reads around thirteen pieces of content before committing to a supplier. Thirteen. That is an industry estimate rather than our own metric, but it explains neatly why a single service page decides nothing.
Someone who has just searched the problem for the first time and someone already comparing three vendors are different readers. They need different pieces. Most corporate blogs write only for the first, then wonder why there is traffic but no enquiries.
So content is built across the whole funnel — Awareness → Consideration → Decision — and each stage removes its own objection:
- Awareness. They don’t yet know the problem is solved this way at all. Objection: “is this even about me?”
- Consideration. They know the approach and are choosing between options. Objection: “why this way and not another?”
- Decision. They are choosing a supplier. Objection: “why you, and why now?”
There is only one way to check whether this works: tie enquiries to the funnel stage of the page they came from. On projects where we have the data, we do exactly that and look at which stages are under-covered.
E
Expand
Widen the coverage
Once a direction is worked through inside the site, the external part begins: links, mentions on other people’s sites, topical authority. The aim is presence in classic search and in model answers at the same time, because the sources behind them overlap more than people assume.
The clearest example from our own work is OroCommerce. For four years, from 2019 to 2023, we worked every month on getting the platform into other people’s round-ups — not articles on its own site, but reviews and comparisons on third-party resources. The tasks were split by region: US, UK and FR separately. When a model lists B2B eCommerce platforms today, those texts are part of what it is drawing on.
Intersections
And here is the payoff for ranking the tabs separately. Once the axes have been worked individually, we go to where they cross.
Not “fintech” and not “Node.js”, but fintech on Node.js. That is where competition is thinnest and traffic is warmest: someone searching for that combination already knows what they want. At EPAM it looked like re-clustering pages onto a vertical and a service together — Retail and Big Data as one thing, not two.
The fintech page and the Node.js page both stay where they are, and both link to cases. An intersection does not replace the axes. It stands on them.
What it costs to start
Briefly, because the numbers are small.
- The competence map is free. You fill it in yourselves; it is a few hours of your team’s time. Homework we go through at the first meeting.
- Market analysis is $30–40 per market, depending on how many. That is billed.
- CAGR data comes from open sources, not paid industry reports. Which is why it is cheap.
- The filled-in map stays with you whether or not we end up working together.
That is the entire up-front cost of finding out whether this is worth continuing. After that it is Time & Materials, you approve the monthly plan, and external costs sit on their own line away from team hours.
The short version
- WEDGE is a five-step framework answering one question: where to put a limited budget for the most potential profit.
- W (Where) — the competence map: the client describes themselves across five tabs, scores 1–3, rows get ranked.
- E (Evaluate) — the second screen: competition in search and LLMs, potential traffic. The output is a queue of directions, not one item.
- D (Decode) — research into the people who decide; the result goes straight into content briefs.
- G (Guide) — content across Awareness → Consideration → Decision.
- E (Expand) — external mentions, topical authority, then the intersections between directions.
- Expertise, sellability and margin are scored by the client. Market movement and reachability are scored by us.
- The map is free; market analysis is $30–40 per market. That is the whole up-front cost before a decision.
Questions
The questions we get most
How is this different from an ordinary SEO audit?
An audit answers “what is wrong with the site”. WEDGE answers the question before that one: which way should this site be growing at all. Different jobs, and the second matters more — a perfectly tuned site in a direction where you earn little remains a site where you earn little.
Do you do this for any kind of business?
The five-tab configuration is built for IT outsourcing. For product companies, industrial manufacturers or education platforms the set of tabs differs — some tabs go, some get replaced. What does not change is the logic: the client scores themselves, we score the market, and all of it then passes through the reachability screen.
What if the map surfaces a direction we do not want?
Then we do not go there. The map is an instrument for a conversation, not a verdict. But the fact that it rose to the top is worth discussing: resistance usually traces back to someone inside the company having tried it once, badly, under different conditions.
How long does the whole cycle take before work starts?
The map is a few hours of your team’s time. Market analysis and the second screen run from a few days to two weeks, depending on how many markets and directions there are. After that you have the matrix, the queue and a plan for the first period.
Why not publish the ranking formula?
Because coefficients without context explain nothing, and publishing them would turn a working instrument into a form people fill in mechanically. The logic itself is stated openly: what rises is where market growth, expertise, client solvency and margin meet. We do not call it a secret method — that would contradict everything else we do.
Start with the competence map
You fill it in, we add the market data. What comes out is a ranked list of your directions and an answer to where the budget should go first. Free, and yours whether or not we work together.