Private equity buys businesses that already work. Revenue, customers, a pipeline. The thesis is usually that it can grow faster, wider, or through more people than it is today.
That assumes the reason customers buy is available to those people. It generally sits with a founder and one or two sellers who have had the conversation enough times to know what moves a decision.
Commercial transfer is whether that reason has made it out of those people and onto the surfaces a buyer meets: the website, the deck, the proposal, the whitepaper, the sales conversation. When it has, those surfaces do some of the selling. When it has not, every deal needs one of the few in the room.
The website is the only one of those a stranger can read without being invited. That is what the Commercial Transfer Index holds: 607 business-to-business homepages, read the way a buyer reads them. Four hundred and seventy-six belong to private-equity-owned companies, across 72 funds.
Each homepage is scored out of thirty, in three bands: Fragile below twelve, Partial from twelve to twenty-one, Strong at twenty-two and above. Strong is a score across all six conditions, and it means a buyer arriving cold can act on the page.
Two of the 476 score in the Strong band. One in 238, across seventy-two houses. Three across the whole calibrated index of 624.
Nineteen funds hold eight or more companies in the index. Their averages run from 10.0 to 15.8. The typical spread inside a single fund is 10.9 points, and twelve of the nineteen span more than eight.
The portfolio average is usually set by two or three companies.
Our hypothesis is that this leads to a shortfall of two or three quarters. A homepage that fails to help a buyer recognise themselves leaks opportunity from the day it goes live. The loss reaches a board pack a quarter or two later, by which point it presents as a market problem.
We have not tested it. The index holds no pipeline data, so it cannot.
The instrument scores six conditions a homepage has to meet. Whether it names the buyer's situation. Whether it uses their words. Whether it says who it is for. Whether it shows the work has worked before. Whether there is a way in. Whether it says why this company rather than another.
Across the private-equity population:
Average, out of 5:
The six do not fail evenly. Who it is for, the way in, and why this company average 2.39. The buyer's situation, the buyer's words, and proof it works average 1.81.
Our reading of that split is that the first three can be assembled inside the building, from a roadmap, a client list and a competitor's website, and the second three cannot. The index records what appears on a homepage. It does not record how the company arrived at it, so the explanation is ours and the gap is the measurement.
Eighty-nine per cent score two or below out of five on whether the homepage names a situation the buyer recognises. Eleven per cent score zero. Only twelve per cent produce a single proof point carrying a metric, a timeframe and a named client.
So the homepage is not doing the work of recognition. Whatever persuasion happens, happens somewhere else.
Usually in a conversation with a founder, or with one of the one or two sellers who know the answer. If that is right, the ceiling on growth is how many conversations those individuals can hold in a week, and new sellers will not reproduce the result because they were given a product to describe rather than a reason a customer buys.
That is the implication we are testing. It is the opposite of the thesis most of these companies were bought on.
We expected to find companies that had not done the work. What we found, again and again, was companies that had done it and put it in the wrong place.
A data infrastructure business runs a second version of its homepage at a URL nobody lands on, carrying a sharper opening line than the live one. A process orchestration company names four situations its buyers arrive with, five screens below a metaphor. A location intelligence firm places four buyer situations in a closing block under the footer navigation. A public sector software company makes an exclusivity claim — the only platform dedicated to — in white text on a black panel, two screens down, while the hero says amplify your impact.
That material is already written and signed off. It is simply not where a buyer arrives.
A company that has never articulated why customers choose it has a research problem lasting months. A company that has articulated it and buried it has a decision to make on a Tuesday afternoon.
Conversion rate. Win rate. Cycle length. Rep attainment. Every one of them is denominated on existing pipeline.
A buyer who scans a homepage, fails to recognise their own situation, and leaves is not a lost deal. They are not a deal. They never enter the denominator. So a weak commercial surface is indistinguishable in a board pack from a market that is smaller than hoped.
It measures what the commercial surface says. It does not measure the business, and a company can run well and score low.
Anthropic's homepage scores 7 out of 30. That tells you almost nothing worth acting on because people do not arrive asking whether Anthropic can solve their problem. They already know what it is.
A mid-market portfolio company cannot rely on that. Its homepage has to establish recognition before it can create preference.
We hold no evidence that a higher score causes better financial performance. We have run no controlled comparison and observed no company without the work. The corpus records no acquisition history, product count, or service count, so it cannot establish that a company's capability has outgrown its explanation. That is an evidence gap, and it stays one.
Eighty-three per cent of the index was assembled from published fund portfolio lists, with no reference to whether the company had a commercial problem. Three per cent approached us. That is unusual for a dataset in this field, and it is the reason we are willing to publish distributions from it at all.
This article summarises one finding from the Commercial Transfer Index.
The full Private Equity Briefing sets out the fund-level distribution, the six-dimension profile, the page constructions that hide evidence from buyers and machines alike, and the questions a partner can answer about a portfolio company in a single afternoon.
It also explains why we believe commercial transfer can be measured before it appears in pipeline or revenue, and how that hypothesis can now be tested.
Mark Gibson
Download the Private Equity Briefing →
Or, if you'd like to test the hypothesis on your own portfolio:
Pick one company.
Within forty-eight hours, we will tell you whether its commercial surface explains one business or a collection of products and services, and establish a measured baseline against which future change can be observed.
Shift90 Partners · Commercial Transfer Index · August 2026
Index composition: 607 European B2B commercial surfaces, of which 476 are private-equity owned across 72 funds. Eighty-six rows are certified or human-read; the remainder are machine-scored and are quoted with that composition attached. Figures current at 21 August 2026.