Every market study starts with the same three reports
An investor in the middle of diligence needed a number no report had: how many of one class of asset sit in American utilities, where, and how many are due for servicing. The market study I would have done for that question a few years ago would have started with the same three top-down reports. This is the workflow that answers it from the bottom, one utility at a time — and since that gap is why this company exists, it is partly an origin story.
In February an investor that backs companies selling into water came to us in the middle of diligence. The business they were looking at inspects and services a particular class of utility asset, and the investment memo needed a number nobody had: how many of those assets are there across American utilities, where are they, how big are they, and how many are due for servicing. The company had a rough idea. The investor wanted an accurate one, so that the sales team could be handed the list and go.
They had the usual options. The two or three top-down market reports everybody in the sector buys, which would give them a segment size and a growth rate to 2030 and not one named utility. Or the person inside the company whose job, it turned out, this already was — weeks of pulling it by hand from databases that are public and nearly impossible to export. The investor's own comment was that in 2026 that should not be anybody's job.
We built it in three lenses: the federal dataset, which is easy to get and incomplete; the state datasets, which have to be read one state at a time; and the utility level, which is where the count becomes accurate and where the time goes. What came back was a list rather than a figure — every utility, how many of the asset, what size — which is what a sales team can act on and what no market report can give.
I spent years in strategy consulting doing market studies for exactly that kind of client, and every one of them started the same way: with those two or three reports.
They are good reports. They will tell you the size of a segment, a growth rate, a forecast to 2030. What they will not tell you — because it is not what they are for — is which specific organizations make up that number, which of them are about to spend, and which of them you could sell to next year.
I had come from a consulting practice covering mining, where we held our own entity-level data on essentially every mine and steel plant in the world. That made a different class of answer possible. Not just how big is this market but which hundred customers should you be selling to, and why those hundred.
Coming into water and finding that view simply did not exist was the thing that eventually became this company. So this is the one workflow where the work and the origin are the same story.
The job, as it exists today
Somebody has to make a decision with money attached and a deadline on it.
An investor is looking at a business that sells into water utilities and needs to know whether its addressable market is what the deck claims. An acquirer wants to know whether a target's customer base is concentrated in places that are growing or places that are quietly shrinking. A lender or an underwriter needs a view of a utility's capacity to pay that does not rest entirely on a rating and a set of audited accounts. A strategy team is being asked which segment to enter first.
All of those are the same underlying question — what does this market look like from the bottom, one organization at a time — and all of them currently get answered with a top-down report plus a fortnight of somebody's time on the phone.
Why it is the same machinery
This is the part worth explaining, because "our platform also does M&A" is the sort of claim that should make a reader suspicious.
The market screening workflow already had to answer which utilities have this characteristic, across the whole market, from documents. The account profile already had to answer what is happening inside this one organization and what is it committed to spend. The installed-base workflow already had to answer what is in the ground and when does it come up.
A market-sizing question is those three, aggregated instead of ranked. A diligence question is the same evidence assembled around a company rather than a utility: where its customers are, what those customers are committed to, who else is selling to them, and what happens to that base over the next five years.
Nothing new has to be believed for this to work. If the underlying record supports telling a vendor which twenty accounts to pursue, it supports telling an investor what those accounts are collectively going to spend. It is the same corpus, asked a question with a different unit of analysis.
Where it stops being the same is the second half of any diligence job — the private half. Contracts, churn, pricing, the actual quality of the customer relationships. None of that is in the public record and none of it is what this is for. This produces the market-side view, which is the half most often taken on faith.
What the workflow does
It builds the universe from the bottom. Not a segment estimate but the actual set of organizations, each with the attributes the question depends on — size, treatment type, compliance position, capital commitments, funding capacity, what they already own.
It applies the buyer's definition of the market, not ours. A market is not a fact; it is a definition, and two investors looking at the same company will draw it differently depending on what they think the company is. The workflow takes the definition as an input, the same way screening takes the customer's criteria rather than imposing ours.
It shows the composition, not just the total. A number is easy to challenge and impossible to act on. The useful output is the number and the list behind it — so that a partner who does not believe the figure can open it, disagree with four hundred of the entries, and get a revised figure rather than an argument.
And it dates everything. A market view assembled from a public record is a view as of a date, and the underlying documents each have their own date. That matters more in diligence than anywhere else we apply it, because the output goes into a paper that somebody signs.
Market / Surface-water treatment upgrades / Composition
The number is not the deliverable
Everything we build makes a version of this argument, and here it has a specific edge.
A market size is the most quotable and least useful thing a study produces. It gets lifted into a slide, loses its definition and its date on the way, and is then repeated for three years by people who have no idea which organizations it counted.
So the deliverable is the composition and the reasoning: this is the definition used, this is the set it produces, these are the assumptions that move it most, and here is what happens to the total if you disagree with any of them. A number that cannot be opened is a rumor with a decimal point.
What it cannot do
It cannot see the private half, and diligence lives there. Contracts, terms, concentration, churn, the reason the last three deals were lost. Public-record analysis answers whether the market is real and where it is going. It cannot tell you whether this company is good at selling into it. Anyone offering both from public data is selling you the first and implying the second.
And it is directional in exactly the places people most want precision. This is worth being concrete about, because we have run into it. In some segments the public record is dense and the analysis is close to definitive. In others — particularly outside the municipal market, where buyers are private companies protective of what they have installed and vendors rarely announce wins — the signal is fragmentary: a patent filing, a conference sponsorship, a named customer on a slide in somebody's investor deck. The honest promise there is a directional map, not a comprehensive one, and we now say so before the work starts rather than in the final report.
The distinction is not cosmetic. A directional map is genuinely useful for deciding where to look. It is not a basis for a valuation, and the failure mode of this whole category is letting a client mistake one for the other because nobody wanted to say it in the first meeting.
Related
- Fifty thousand is not a market — the same machinery, screening for a vendor instead of sizing for a buyer.
- One utility, all the way down — the entity-level record a market total is built up from.