Estimation question

Estimate the number of vacuum sold in the US.

Practice this question out loud. An AI interviewer asks it, follows up like a real interviewer would, and scores your answer. Type or speak.

What this question tests

Tests structured Fermi estimation using household counts and replacement cycles for a durable goods product.

Frame the formula: number of US households times the share that buy a vacuum in a given year, driven mostly by replacement, not first purchase.
Assume roughly 130 million households in the US.
Assume a vacuum lasts on average about 8 years before replacement, so roughly 1 in 8 households buys a replacement vacuum in a given year, about 16 million.
Add new-household formation and gift or secondary-vacuum purchases, roughly another 2 to 4 million units, for a combined estimate around 18 to 20 million.
State a final estimate of roughly 18 to 20 million vacuums sold per year in the US, clearly marked as assumption-based.

Uses a replacement-cycle model, the correct approach for a durable good, rather than treating this like a consumable product.
States the average product lifespan assumption explicitly, the key driver of the whole estimate.
Adds a secondary factor, new households and gifting, to avoid understating the total from replacement alone.

Treating vacuums like a fast-consumable product bought frequently, ignoring the multi-year replacement cycle.
Giving a single guessed number without decomposing it into households and replacement rate.

How would this estimate change for robot vacuums specifically versus traditional ones?
How would you sanity-check this against retail sales data if you had access to it?