Market Sizing: From Top-Down to Bottom-Up in 90 Seconds
Most candidates pick top-down or bottom-up and commit. The strong move is to know when each one is structurally right — and to triangulate when neither is.
Market sizing is one of the most common case types and one of the most misunderstood. Candidates often default to one approach — usually top-down because it feels structured — and end up anchoring on a number that is plausibly off by 10x. The discipline is in choosing the right approach for the actual question.
Top-down: when the population is well-known
Top-down works when you can quickly state a population number with confidence. Number of US households (~130M). Number of US adults aged 25–54 (~130M). Number of cars on the road in the US (~280M). When the denominator is known, you walk down: % who qualify, % who buy, average spend, frequency.
The risk is the funnel multiplies error. Each step is "about 40%, maybe 30%, let's say 35%" and four steps in your range is ±60% of the true number. That is a directional answer, not a useful one.
Bottom-up: when the unit economics are tractable
Bottom-up starts with one buyer or one transaction and scales up. How many supermarkets are there per zip code, how many zip codes per city, how many cities. How many widgets does a typical commercial kitchen go through per week, how many commercial kitchens.
Bottom-up tends to be more defensible because each input is anchored in something the interviewer can validate. The risk is you can spend three minutes trying to remember whether there are 15,000 or 50,000 dentists in the US.
Pick top-down when the customer base is roughly the general population and the question is about purchase volume. Pick bottom-up when the customer base is an institutional segment (hospitals, restaurants, factories) and the per-unit usage is easier to estimate than the unit count.
The strong move: triangulate
At partner round, the candidates who stand out do both. They size top-down, get a number, then quickly check it bottom-up. If the two come out within 30%, you have a confident answer. If they come out 5x apart, you have just discovered something interesting about the market structure — and you should call that out.
The triangulation does not have to be slow. "$10B top-down, let me sanity-check: if there are 100,000 of these and each spends $100K, that is $10B — consistent." 20 seconds, and you have demonstrated a level of analytical rigor that scores at partner level.
Three traps to avoid
- Sizing the wrong market. If the question is "what is the addressable market for our SaaS product", do not size the entire industry. Size the segment that could plausibly buy your product at your price point.
- Forgetting frequency. A market sizing for coffee is not just coffee drinkers × cups per day. It is drinkers × cups × days per year. The annual factor is where most off-by-an-order-of-magnitude errors happen.
- Mixing units. If you start in dollars, end in dollars. If you start in units, decide whether the answer is asked in units or in dollars and convert once at the end.
Drill: 90-second sizes
Pick five questions and force yourself to land an answer within 90 seconds, with stated assumptions. Not "I would need more time" — a number, with the assumption tree exposed:
- How many electric vehicles will be sold in the US in 2030?
- What is the annual market for diabetes test strips?
- How many haircuts are given in NYC per year?
- What is the global market for commercial drones?
- How many gallons of paint are used in US residential remodels per year?
After ten of these, the 90-second constraint stops feeling artificial. That is the bar for casing fluency: not getting the right answer, but getting an answer with defensible assumptions in the time the interviewer expects.
Mental Math for Consulting: 12 Shortcuts You Can't Afford to Miss
Sizing math falls apart without fluent arithmetic. Twelve shortcuts that compound across an hour of casing.
One short, useful email per week.
New case frameworks, mock interview tactics, and recruiting intel from the CaseGrade team. Built for serious candidates — no fluff, no upsell every paragraph.
Unsubscribe anytime. We never share your address.
Keep reading
All articles →The State of Consulting in 2026: AI Delivery, Flat Entry Pay, and What Candidates Should Care About
Six things are shifting at once inside the consulting industry. Most of them are being talked about loudly. A few are being talked about quietly. Candidates who only track the loud ones are missing the more important story.
McKinsey Solve in 2026: Red Rock, Sustainable Future Lab, and What Actually Improves Your Score
McKinsey's digital assessment got a third module in late 2025. The good news: the underlying skills haven't changed. The bad news: most of the prep advice on the internet hasn't caught up.
What AI Case Interviewers Can — and Can't — Evaluate
Several AI case-prep tools have launched in 2026. Some are genuinely useful. Some are thin wrappers. The difference is what they can score honestly and what they have to hand-wave through.