Market sizing: top-down vs bottom-up

Estimate the same market two independent ways — when both routes land in the same ballpark, you can trust the number.

The idea

There are two honest ways to size a market. Top-down starts from the whole population and filters down: households × the share that fits × what each one spends. Bottom-up starts from a single unit and builds up: providers × volume × price.

Neither is “the” answer. The credibility move — the thing interviewers actually listen for — is running both and reconciling the gap. Two routes that agree within roughly 2× is a real signal; a 10× gap means one assumption is wrong, and finding it is the point.

How it works

Both routes are just a chain of multiplications — but they lean on different facts, so agreement is meaningful. Here are the defaults worked through:

top-down  (start big, filter down)
  130M households x 45% with a dog        = 58.5M dog homes
             x 35% who pay a groomer      = 20.5M grooming homes
             x $250 spend / yr            = $5.1B

bottom-up (start with one unit, scale up)
  100k groomers x 6 dogs/day              = 600k grooms/day
             x 250 working days           = 150M grooms/yr
             x $55 per groom              = $8.3B

reconcile:  $8.3B / $5.1B = 1.6x apart  ->  within 2x, they agree
credible number ~= geometric middle       ~= $6.5B

TAM / SAM / SOM are the same funnel, one level up. TAM is the whole market (all US dog-grooming spend, the number above). SAM is the slice your product can actually serve — say mobile grooming in the 30 largest metros. SOM is what you could realistically capture in a few years given competition and reach. Each is a filter on the last, never the other way around.

When to use it

Route	Lean on it when…	Its blind spot
Top-down	You know population and rough penetration rates.	Small errors in the percentages compound fast.
Bottom-up	You can picture one provider or customer’s economics.	Easy to under- or over-count the number of units.
Both, reconciled	Always, in an interview — it’s the credibility move.	Takes a minute more; worth it every time.

Watch out for

• Reporting one route only. A lone estimate can’t be checked. Two independent routes that agree is the whole point — if you only have time for one number, at least sanity-check it against the other lens out loud.
• Unit mismatches. Households vs individuals vs dogs; per-visit price vs annual spend. Mixing them silently is the most common way an estimate ends up 5× off. State your unit at every step.
• False precision. Inputs are guesses, so $5,118,750,000 is a tell. Round to ~$5B. The interviewer is grading your structure, not your long division.
• Reconciling by splitting the difference. If the two routes are 8× apart, don’t just average them. Hunt for the wrong assumption — usually the unit count or a penetration rate — and fix it.
• Quoting TAM as if it were revenue. TAM is everyone who could buy. Your achievable share (SOM) is a small fraction. Confusing them makes a business plan look naive.

Worked example

An interviewer asks: “How big is the US dog-grooming market?” You say it out loud both ways. Top-down: ~130M households, call it ~45% with a dog, maybe a third pay a groomer, at a couple hundred dollars a year — that lands around $5B. Bottom-up: roughly 100k working groomers, ~6 dogs a day, ~250 days, ~$55 a groom — that’s about $8B. You note they’re only ~1.6× apart, so you’re confident calling it “somewhere around $5–8B, I’d anchor near $6.5B.” Then, if it’s a startup case, you filter once more: your SAM might be app-booked mobile grooming in big metros (a few hundred million), and your 3-year SOM a slice of that. Naming the gap and the filters is what separates a memorized number from real reasoning.