Market share is money in the category, not a headcount — the definition every finance page uses. Penetration is how many buyers you win. They're separate axes: hold one still and the other still moves. So penetration is a pathway into revenue, not the objective — and the goal was never a headcount. It's profitable revenue.
Wins a set of buyers and takes a thin slice of each one's category spend.
Wins the exact same buyers — same people, same count — but takes most of each one's spend.
Both brands win the identical set of buyers — same people, same count, 60% penetration each. One captures 5% of the category's money, the other 50%. Same penetration, 10× the share — so penetration doesn't determine the money; depth of capture did the work here. Penetration is one route into revenue, not the thing you're after. And it's worse for “grow penetration” than this: B could win fewer buyers than A and still take more of the money.
You've won 60% of buyers and 7% of the money. Most of the market, almost none of the revenue — the headcount didn't pay.
Profitable revenue — contribution, not vanity volume. Capture more of every buyer's category spend, durably rather than by renting it with discounts — from the ones you already have (their next purchase) and the ones you don't yet. “Loyalty” is the wrong word for winning the next purchase: it drags in arguments about whether brand love is real. It isn't a feeling — it's a transaction you win or lose. Call it next-purchase capture.
Penetration is a pathway metric, not the objective function. Chase profitable revenue, then find the route — more buyers, more occasions, more wallet, higher realized price, better mix — whichever builds the most value in your category. Sometimes that lifts penetration; sometimes it doesn't, and that's fine, because penetration was never the point. Set penetration as the goal and you privilege one route and collapse marketing into buying reach. Chase value and you're back to the full toolkit — price, product, availability, communication. The 4 Ps. That's the job.
Share = your dollars ÷ the category's dollars. Penetration is one term in an exact identity — share = penetration × share-of-wallet × how heavy your buyers are — so it's in the maths, but you can hold it fixed and move share across its whole range. That separability isn't a model. It's arithmetic; try to break the proof above and you can’t.
The shape of category spend — a few heavy buyers, a long tail of light ones. Realistic for most categories, but illustrative, not your curve. Flip “which buyers” to see how much of the result rides on it.
Whether, in your category, chasing revenue shows up as more buyers or heavier ones. That's an empirical fact about your market — not a universal law. It's the one thing here you actually have to measure, and the thing “grow penetration” assumes it already knows.
Partly because they're built from the same transactions — you can't add a buyer's spend without moving both, so acquiring buyers genuinely can lift revenue. The causal path isn't the problem. What equilibrium panel data can't tell you is whether maximizing penetration beats the alternatives — deepening wallet, raising price, improving mix — on incremental value. Revenue moves through several of those at once, so co-movement can't isolate what penetration alone contributed. So the question was never whether penetration matters. It's when adding buyers produces more value than the other routes — a question about decisions that this kind of data can’t answer on its own, and one “grow penetration” tends to skip. The same gap this whole site keeps finding: a readout is never the mechanism.
Adjacent, and a module of its own: the Double Jeopardy pattern this argument circles — bigger brands earn slightly higher repeat — is itself causally ambiguous. Wider distribution alone would generate the same curve, so it may be measuring shelf presence as much as memory. The data shows the pattern; which mechanism you credit is a choice it doesn't force.
Numbers illustrative