Start market sizing by defining exactly what you are estimating: geography, customer, time period and unit. Many weak answers come from solving the wrong market rather than from bad arithmetic.

Then choose a sensible driver tree. A demand-side approach might use population, penetration and usage, while a supply-side approach might use locations, capacity and throughput.

Make assumptions explicit and keep the math simple enough to audit. The interviewer cares about whether your assumptions are reasonable and your logic is coherent, not whether you know every real-world number.

After calculating, sanity-check the result. Compare it with something familiar, test one key assumption and ask whether the order of magnitude makes sense before presenting the answer. For more detail, see the market sizing guide.

I always get stuck deciding whether to estimate from households or from individual buyers. Is there a good way to choose?