Skip to content
PMInterviewKit
In this module6 lessons

How to practice so it sticks

Run practice sessions that build real interview skill, using out-loud timed reps, immediate grading, one focus per session and a simple log.

5 min readFree lesson

Why most practice doesn't work

Most candidates prepare by reading. They read frameworks, sample answers and company guides, then think through a few questions in their head on the train. It feels like progress. Then, in the interview, the words come out slower and messier than the thoughts did, and the structure they knew perfectly well disappears under pressure.

The problem isn't effort. It's that reading and thinking train a different skill from the one an interview tests. An interview tests whether you can produce a clear, structured, choiceful answer, out loud, in real time, while someone watches. The only way to train that is to do it.

This lesson sets out six principles for practice that transfers to the real thing, then gives you a session template you can run tonight.

Principle 1: Out loud, every time

Say every practice answer out loud, even alone. Speaking exposes gaps that thinking hides: the sentence you can't finish, the transition you don't have, the point that takes three minutes instead of thirty seconds.

Weak practice: "I thought through how I'd design a meal-planning app on my walk." Strong practice: "I answered a meal-planning prompt out loud for 25 minutes and recorded it."

Principle 2: Timed, like the real thing

Set a timer that matches the round. Rough guides that work for most loops:

Round typePractice time
Product sense25 to 30 minutes
Metrics, diagnosis or execution20 to 25 minutes
Estimation10 to 15 minutes
Strategy20 to 25 minutes
Behavioral story2 to 3 minutes per story
Technical or AI product20 to 25 minutes

These are practice targets, not any company's rule. Real interviews leave time for introductions and your own questions, so a 45-minute slot often holds a 25 to 30 minute case. Time pressure changes how you think. Practice without it and you'll learn to give answers that only work when nobody's counting.

Principle 3: Record and grade right away

Record on your phone. Then, within the hour, listen back and score yourself with the self-grading rubric.

Listening back is uncomfortable. Do it anyway. You'll hear things you didn't notice while speaking: the two-minute preamble, the "um, so, yeah" filler, the moment you listed four users and never picked one. Grading while it's fresh means you still remember what you were thinking and can compare it with what you actually said.

When you grade, quote yourself. Instead of "structure: 3", write "structure: 3, announced a plan at 1:10 but skipped the prioritization step". Scores without quotes drift upward over time.

Key idea A practice answer you don't grade is mostly wasted. The grading is where you learn what to change.

Principle 4: One focus per session

Don't try to fix everything at once. After each graded answer, pick one thing to change in the next one: the lowest-scoring rubric dimension, or a specific habit you keep skipping.

Write it at the top of your next session: "Focus: pick one user and say why within the first five minutes." Then judge that session mainly on whether you did that one thing. This is deliberate practice: narrow, measurable, repeated until it sticks. Drills by weakness has a short drill for each common gap, so you can do extra reps on just that skill without running a full answer.

Principle 5: Space it and mix it

Two patterns help skills last.

Spacing. Practicing a round type a little on several days beats cramming it all in one day. If you do product sense on Monday, come back to it on Wednesday and Saturday rather than doing three answers on Monday.

Mixing. Once you've learned each round's method, stop practicing in blocks. Pick questions from the question bank across types, so you have to recognize the round and choose the method each time, just like the real loop. Blocked practice feels smoother. Mixed practice prepares you better for a loop where the type changes every 45 minutes.

Principle 6: Mix solo and partner practice

Solo practice builds your method. Partner practice builds what solo can't: handling interruptions, pushback and follow-up questions, and talking to a real face.

A good rhythm for most people is three or four solo answers for every partner mock. Your partner doesn't need to be a PM. Mock interview scripts gives a non-PM friend word-for-word lines to read, plus follow-up probes and a feedback form.

How many reps?

There's no magic number, and anyone who gives you one is guessing. A workable rule of thumb, not a research finding: plan on about ten graded answers in a round type before you judge your progress, and don't be surprised if you need twice that. Watch your rubric scores rather than counting: when your last three answers in a type all score in the interview-ready band, shift your time to the next weakest type.

A 45-minute session template

Use this for a normal practice day.

  1. Two minutes: set the focus. Read the "next time" note from your last log entry. Write today's focus at the top of the page.
  2. One minute: pick a question. Choose at random from the question bank, filtered to today's type. Don't preview several and pick your favorite.
  3. 25 minutes: answer it. Out loud, timed, recorded. Have paper for notes, as you would in a real interview.
  4. 12 minutes: grade. Listen back at 1.5x speed if you like. Score each rubric dimension with a quote as evidence.
  5. Five minutes: log and decide. Write the score, the question and one specific thing to change next time.

For estimation or behavioral days, swap step 3 for two or three shorter answers and grade each.

Try it Run this template once today with any product sense question from the bank. Don't aim for a good answer. Aim for a completed log entry with one specific thing to fix.

The practice log

Keep a running log in a document or spreadsheet. Columns:

DateQuestionTypeScore (of 20)Lowest dimensionNext time I will…

The last column is the one that matters. "Be better at metrics" isn't useful. "Name a guardrail metric before I summarize" is.

Every week, read the log top to bottom. You'll see patterns you can't see answer by answer: the dimension that's stuck at 2, the round type you've been avoiding, the improvement that came from one specific change.

Recap

  • Reading builds knowledge; interviews test performance, so practice out loud, timed and recorded.
  • Grade every answer within the hour, using the rubric and quoting yourself as evidence.
  • Focus each session on one change, and drill that skill on its own when it's stuck.
  • Space practice across days and mix round types once you know each method.
  • Keep a log with a specific "next time" line, and read it back every week.

You’ve just read one of 15 free lessons.

The other 73, including 66 worked answers written out in full, are in the full kit: $129 once.

Get one email when the full kit opens

No spam: just one launch email. See our privacy policy.