The approach in five steps

Estimation questions sound like trivia. "How many haircuts happen in the US each year?" "What is the annual market for a dog-walking app?" "How much storage does a photo app need?" Nobody expects you to know the answer. The interviewer wants to watch you get from a vague question to a defensible number, with every step visible.

That is good news. You cannot prepare for every question, but you can prepare one method and use it on all of them. This lesson gives you that method. The rest of the module gives you the round numbers, the two ways to build a structure, eight full worked cases and a list of the slips that cost people points.

#What the interviewer is scoring

Most interviewers who ask estimation questions are looking at the same handful of things:

• Structure. Did you break the question into parts that multiply or add up to the answer?
• Assumptions. Did you say each number out loud, with a reason, so it can be challenged?
• Arithmetic. Can you multiply big numbers without getting lost or dropping a zero?
• Judgment. Did you notice when a number looked wrong, and did you know which assumption mattered most?
• Communication. Could the interviewer follow you without asking "wait, where did that come from?"

Notice what is missing: being right. Real answers to these questions are often unknown, disputed or private.

Key idea You are graded on how you reach the number, not on the number. A clear structure that lands within a factor of two beats a lucky guess with no working.

#The five-step estimate

The kit's method has five steps. You will see it used the same way in every worked case in this module.

A typical answer takes about ten minutes. As a rough budget: one minute to clarify, two to structure, two to assume, three to calculate and two to check.

Here is the method on one question, start to finish.

Interviewer How many haircuts happen in the US each year?

#Step 1: Clarify

Spend about a minute making the question concrete. You want to know what counts and what does not.

You A couple of quick checks. Do we mean professional haircuts, at a barber or salon, or do haircuts at home count too? And is this one recent year, not a trend?

Interviewer Professional only. One recent year.

Good clarifying questions change the answer. "Do you mean haircuts for dogs?" does not help you. "Do home haircuts count?" does, because it could move the answer by a large share.

#Step 2: Structure

Say the equation before you touch a number. This is the step interviewers remember, because it shows you can turn a fuzzy question into a model.

You I'll start from the population and split it by how often people get a haircut, because that varies a lot. Haircuts per year equals, for each group, the number of people times haircuts per person per year. Then I'll take out the share done at home, since we only want professional cuts.

Write it where the interviewer can see it, on a whiteboard or a shared doc:

Professional haircuts = (short-hair people × cuts per year + longer-hair people × cuts per year) × share done professionally

Segmenting by behavior (how often people cut their hair) is better than segmenting by something that only loosely predicts it. Pick the split that most changes the number.

#Step 3: Assume

Now put a round, stated number on each part. Every number gets a reason, even a short one.

You These are all round numbers so the math stays easy. Tell me if any of them looks off to you.

That last line matters. It invites the interviewer to correct you early, which is far better than finding out at the end.

#Step 4: Calculate

Multiply out loud, one line at a time, and round as you go.

• Short hair: 170 million × 9 = 1.53 billion
• Longer hair: 170 million × 4 = 0.68 billion
• Total: 1.53 + 0.68 = 2.21 billion
• Professional: 2.21 billion × 0.75 ≈ 1.66 billion

You So roughly 1.7 billion professional haircuts a year.

Round the final answer to one or two significant figures. "1,657,500,000" signals false precision. "About 1.7 billion" signals judgment.

#Step 5: Check

Now test the answer from the other side. If the demand side says 1.7 billion haircuts, the supply side should be able to deliver them.

You Let me sanity-check from the supply side. A barber or stylist might do about 8 cuts a day, 250 working days a year, so about 2,000 a year. 1.66 billion divided by 2,000 is about 830,000 full-time barbers and stylists. That's one for every 410 or so people. A town of 10,000 would have about 24 stylists, so maybe six salons with four chairs each. That sounds plausible. The assumption I'm least sure about is the 9 cuts a year for short hair. If it's really 6, the answer drops to about 1.3 billion.

Two things happened there. You checked the answer against a different structure, and you named the assumption that moves the answer most. Both show judgment.

Finish with a one-line summary:

You So my estimate is about 1.7 billion professional haircuts a year, driven mostly by how often short-hair customers come back.

#What a weaker answer does

Compare the strong version with the answers that lose points:

• Jumps straight to a number. "I'd guess about a billion." There is nothing to give credit for, and nothing to discuss.
• Starts calculating before structuring. The candidate multiplies numbers as they occur to them, then loses track of which ones are already in.
• Hides assumptions. Numbers appear with no reason, so the interviewer cannot tell a considered guess from a random one.
• Uses awkward numbers. "341.8 million times 8.67" costs time and invites arithmetic slips with no gain in accuracy.
• Skips the check. The answer might be ten times too big and the candidate never notices.

#The habits that make the method work

A few habits apply in every step:

• Talk the whole time. Silence while you think reads as being stuck. Say what you are doing, even when it is "let me just multiply this".
• Write as you go. A visible equation and a list of assumptions let the interviewer follow you and let you recover if you lose your place.
• Use round numbers on purpose. Accuracy comes from good structure, not from precise inputs. The next lesson, Round numbers worth knowing, gives you a table to learn.
• Welcome pushback. When the interviewer says "that seems high", treat it as a gift. Change the assumption, rerun the line, move on.

These habits are the estimation version of the ones in Habits every strong answer shares.

Try it Set a timer for ten minutes and estimate how many birthday cakes are bought in the US each year. Use all five steps, out loud. Then ask yourself: did you clarify whether supermarket cakes count, did you say the equation before the numbers, and did you check the answer from another side?

#Where this module goes next

• Round numbers worth knowing: the population, household, time and data figures that make the arithmetic easy.
• Top-down and bottom-up: the two ways to build Step 2, and how to use one to check the other.
• Worked cases 1 to 4 and Worked cases 5 to 8: eight full answers using this exact method.
• Sanity checks and common slips: how to catch a bad number before the interviewer does.

#Recap

• Estimation questions score your structure, stated assumptions, arithmetic, judgment and communication, not whether you hit a true number.
• The five-step estimate: Clarify, Structure, Assume, Calculate, Check.
• Say the equation before any numbers, and give every number a short reason.
• Round as you go and give the final answer to one or two significant figures.
• Always check from a second angle and name the assumption that moves the answer most.