What is a denominator problem in market share analysis?

Old tractors parked together in a snowy field, machinery that stays in the field long after it ships.

A denominator problem shows up the moment someone asks "what's our market share in this region" and the honest answer is "share of what, exactly."

Market share is a fraction. Your units sold (or your fleet in the field) sit on top. Some measure of the total market sits on the bottom. Everyone argues about the top number constantly, because it's your own sales data and you trust it. Nobody argues about the bottom number, because almost nobody has a good way to check it. That's the denominator problem: the number doing half the work in your share calculation is usually the number you've done the least to verify.

It gets worse in equipment markets specifically. A car OEM can lean on registration data. A construction, ag, or mining equipment OEM usually can't, not at the regional level that matters for a sales territory or a dealer footprint decision. So the denominator gets built from whatever's lying around: an industry association shipment estimate, a dealer's gut feel about "what's out there," a consultant's regional multiplier applied to national totals. Each of those is a reasonable starting point and a weak foundation for a number your board is going to see.

Why TAM and installed base answer different questions

This is where a lot of share analysis goes wrong: TAM and installed base get used interchangeably, but they measure different things.

TAM (total addressable market), sized the usual way, is a demand estimate. Analysts build it from macro inputs: construction spend, ag acreage, mining capex, whatever drives purchases in the category, run through a model that converts that spending into unit volume. The number is forward-looking, useful for sizing an opportunity or justifying a budget. As a denominator for a share calculation it falls short, because it isn't counting anything that exists yet. It estimates what the market will buy, not what's already running in the field.

Installed base is a stock number: it counts how many units of your competitive set are physically in operation in a region right now, regardless of when they were bought or what shipment data says. For a share calculation, that's the number you want, because your share is a function of what's deployed, not what's projected. If you want to know whether your fleet is gaining or losing ground against a competitor in a specific territory, the denominator has to be "units in the field in that territory," not "units the market could theoretically absorb this year."

Swap TAM in as your denominator and you'll get a share figure that moves around for reasons that have nothing to do with your sales performance, like a revision to a macro input three steps upstream. That's a hard number to defend in front of a board, and it's a worse number to plan dealer coverage or service network investment against.

Where the installed-base number usually comes from, and why it's shaky

Most teams get their installed-base denominator one of three ways. Industry association shipment data, cumulated over some assumed fleet life, which ignores retirements, re-exports, and regional reallocation. Dealer or field-rep surveys, which top out at whatever your own dealers can see and miss every unit sold through a channel you don't touch. Or a straight national-to-regional split based on population or GDP share, which assumes equipment density is uniform across a region when it almost never is.

All three produce a number. None of them produce a number you'd want to defend line by line if a competitor's strategy team challenged your market share claim in a board deck.

A denominator built from an actual count of machines on the ground, scaled up with a correction for what a single image or survey pass can't see, gets you closer to a number you can walk someone through step by step instead of citing a source and hoping nobody asks further. That's the gap Installed Fleet Estimation is built around: turning site-level machine counts from high-resolution satellite imagery into a regional installed-base figure, with coverage and occlusion correction built into the scaling, so the denominator in your share math has a methodology behind it instead of a footnote.

If your next board deck needs a share number you can defend past the first follow-up question, that's worth a look.

Request early access

← Back to the blog