McKinsey & Company Psychometric Tests Prep

History at McKinsey & Company. Developed by Imbellus (founded by Rebecca Kantar); McKinsey acquired the company and brought the testing infrastructure fully in-house.

Candidate reputation. A sophisticated digital assessment that bypasses rote drilling: it tracks not just your final choice but your behavioural approach (every click, filter and correction), measuring pure fluid intelligence rather than acquired business knowledge.

• McKinsey Solve (Imbellus, in-house)-calibrated practice. Not a generic stand-in. The exact provider and section structure McKinsey & Company uses on the day.
• Adaptive question difficulty. Questions get harder when you nail them, easier when you struggle. Walk away knowing where you actually stand.
• Coaching, not just a score. "You are spending too long on table-data questions" beats "you scored 68%". Specific advice per weak section.
• Percentile benchmarking. Your score compared to the UK candidate pool, so you know if 70% is excellent or worrying for that test.

Free practice section, scored. Full report unlocks with the Pack.

1. 1

   Poor cross-module time management

   Over-investing in the Redrock investigation phase leaves too little for the actual calculations.

2. 2

   The click-happy process profile

   Trial-and-error microbe swapping logs as erratic and tanks the process score, even if a valid combination is found.

3. 3

   Confusing averaged and absolute constraints

   Treating an averaged threshold as an additive cap, eliminating valid configurations.

4. 4

   Information overload in data review

   Reading the Redrock narrative instead of isolating the functional charts and tables for the math.

5. 5

   Inconsistent behavioural logic

   Switching between collaborative and autocratic styles across the Sustainable Futures Lab questions.

6. 6

   Bypassing the tutorials

   Rushing the untimed rules screens and misunderstanding the filter and drag-and-drop mechanics under time.

• Filter-first elimination

  On Sea Wolf, use the characteristic filters to cross out 60% of options before evaluating any combination.

• 70/30 pacing on Redrock

  No more than 30% of the time on data gathering, 70% on building, validating and writing the recommendation.

• A clean physical scratchpad

  A matrix grid to track variables, formulae and elimination states rather than holding them in your head.

• Hypothesis-first on charts

  Form a hypothesis from the axes before deep calculation, isolating the independent variable driving the shift.

• Calibration simulations

  Build muscle memory for the filter configurations and data layouts on realistic interactive practice tools.

• An external mouse

  Far faster than a trackpad for the rapid filtering and drag-and-drop steps.

• High-fidelity simulations

  Use updated interactive tools that replicate the Redrock and Sea Wolf interfaces, filters and data layouts.

• GMAT Integrated Reasoning

  Drill graph-description and table-analysis questions for chart speed, compound growth and weighted averages without a spreadsheet.

• Free practice on Intervyo

  Run psychometric and numerical practice in the real formats to build pacing under cognitive load.

Time investment. Successful candidates typically invest 15-25 hours of focused prep over two weeks, weighted to simulation runs, quantitative refinement and constraint-logic drills.

1. 01Set your own clock per question. Divide section time by question count. Move on when you hit your per-question budget, even if you are mid-thought.
2. 02Read the question before the chart. Half the work in numerical is finding the right number in a table. Knowing what you are looking for cuts the time in half.
3. 03Use "Cannot Say" generously. Verbal reasoning rewards strict reading. If the passage does not say it, the answer is "Cannot Say", not your own inference.
4. 04Skip the impossible ones. Most tests do not penalise wrong answers more than skips. If you cannot see it in 20 seconds, flag and move on.
5. 05Practise the exact format, not a generic stand-in. McKinsey Solve (Imbellus, in-house) has its own rhythm, and a generic reasoning test is not the same. Intervyo's simulation is calibrated to this format, so you rehearse the real thing under real timing rather than a generic aptitude set.