3f. Sizing questions

How do I tackle an estimation question?

During product interviews, you may get a “sizing question”, where you’re asked to estimate some business-related number. Examples of these:

  • “Estimate the number of apps being downloaded by Americans each day”
  • “Estimate the number of fridges being sold around the world each day”
  • “Estimate the total number of packages being delivered by Amazon each day”

A couple of things to note about these questions. First, it is extremely unlikely that you’ve actually heard the statistic before – don’t panic. Second, they are questions that your interviewer also probably doesn’t know the real answer to. Third, the point of the question is not to test your knowledge of random facts, but to get you to demonstrate analytical ability to estimate something.

This brings us to probably the most fatal mistake you can make on this type of question: making too few assumptions and jumping too quickly to a conclusion. For example, if you’re estimating the number of apps being downloaded by Americans each day, it’s tempting to go down this route:

  • There are around 300M Americans
  • Assume 80% of them have smartphones → 240M
  • Assume 10% of them download 1 new app a day → 24M apps being downloaded each day

Sure, you arrived at an answer, but you made very few assumptions, which really doesn’t go a long way towards demonstrating your analytical abilities. Also, with fewer assumptions, they are likely to be bigger, unfounded assumptions, so it’s harder for you to justify why you made them.

So in other words, here’s some general advice for these estimation questions:

  • Complicate your answer a bit – but not so complicated that you confuse yourself or get lost in your own numbers
  • Pick round numbers to work with, and make sure your arithmetic is right
  • It is likely worthwhile to split off the calculation by different user segments (product is all about the user!)
  • Make slightly differentiated assumptions for each of the user segments (e.g. “millennials probably use 10 apps per day because of X reason, middle-aged adults use 5 apps per day because of Y reason”)
  • Try to justify each assumption, and note how strong your justification feels

Once you’ve arrived at a final estimate, your job isn’t done yet! Without skipping a beat, you should then transition to the following:

  • Immediately do a “gut check” of your answer – “This feels low / high to me because of X, Y, Z reasons”
  • As you’re writing things down, put a little mark next to all numbers that are assumptions; at the end, mentally review all your assumptions
  • Point out the assumptions that you think are either most unfounded, or which appear to make the biggest difference to your final estimate (arithmetically)

Estimation questions are a bit more notorious as parts of the consulting interview process. There is a lot of good advice out there about estimation questions in that context, so do a quick Google of those resources and study those – there is too much good material out there already.

A product estimation question is usually less rigorous than a consulting estimation question (so no need to be as stiff as consultants want you to be). Estimation questions also tend to be a bit less common in product interviews, so as you’re prepping for interviews, devote proportionately a bit less time to studying for these.