McKinsey Problem Solving Game (Solve): The 2026 Guide

The McKinsey Problem Solving Game (branded Solve, and built by Imbellus, which McKinsey acquired) is a 35-minute digital assessment that replaces the old pen-and-paper Problem Solving Test. It looks like a video game and scores like a cognitive assessment. Candidates work through scenario-based tasks, typically an ecosystem-building exercise and a plant-defence exercise, in which they gather information, form a rule, apply it under constraints and adapt when the environment changes. It measures systems thinking, decision-making under uncertainty and cognitive flexibility. McKinsey's own position is that it cannot be practised in the way an aptitude test can.

What Is the McKinsey Problem Solving Game?

Solve is a 35-minute digital assessment disguised as a game. Instead of asking you multiple-choice questions about a business case, it drops you into an unfamiliar simulated environment with rules you have to discover, and watches what you do. The design goal is to measure raw cognitive ability in a way that is harder to coach than a traditional test, which is precisely why the honest advice about it is so different from the advice about an SHL numerical test.

That distinction matters for how you spend your preparation time. With a numerical reasoning test, format-specific practice reliably moves your percentile. With Solve, the returns come from being in good cognitive condition and from not wasting the clock on the interface. The rest of your energy should go into the stages that genuinely reward preparation: the case interviews and the Personal Experience Interview.

Where the Game Sits in McKinsey's Process

1. 1Online application (about a week to hear back). CV, cover letter and academic transcripts. Academic performance carries real weight at McKinsey.
2. 2McKinsey Problem Solving Game / Solve (35 minutes). The digital cognitive assessment described in this guide, built by Imbellus. It measures systems thinking and decision-making under uncertainty.
3. 3First-round interviews (one day, typically two interviews). Two case interviews with associates or engagement managers. Each includes about five minutes of Personal Experience Interview and a 25-minute business case.
4. 4Final-round interviews (one day, typically two to three interviews). Case interviews with partners and senior partners: more ambiguous cases and deeper PEI probing.

The Two Scenarios: Ecosystem Building and Plant Defence

The two scenarios most consistently reported are ecosystem building and plant defence. Neither requires any biology knowledge. Both are dressed-up constraint problems.

• Ecosystem building. You are given a location (a mountain, a reef) with environmental conditions, and a library of species with their own requirements and food relationships. You must place a set of species so that the ecosystem is self-sustaining: every species must have its conditions met and a viable food source. The work is filtering candidates against hard constraints, then building a food chain that closes.
• Plant defence. You must protect a plant from waves of incoming pathogens or invaders, choosing defences and placements under resource limits, while the threat pattern changes between waves. The work is reading a pattern, allocating scarce resources against it, and adapting when the pattern shifts.

The common structure is worth naming, because it is the thing you can actually internalise: gather information, identify the binding constraint, form a rule, apply it consistently, and revise the rule when the environment gives you new evidence. That is also, not coincidentally, what consulting work looks like.

What the Game Actually Measures

• Systems thinking: whether you can see how elements interact rather than optimising one element in isolation. Placing the strongest individual species is not the same as building an ecosystem that survives.
• Decision-making under uncertainty: whether you act sensibly when you do not have complete information, rather than freezing or guessing wildly.
• Cognitive flexibility: whether you can drop a rule that stops working. Candidates who commit early to a strategy and defend it against the evidence do poorly.
• Process discipline: the assessment observes how you work, not only what you produce: the order you inspect information in, whether you check constraints before committing, whether your actions are consistent.
• Working under a clock: 35 minutes for two substantial scenarios is not generous. Time spent fighting the interface is time not spent reasoning.

How It Is Scored

Solve produces a profile rather than a simple mark, and the distinguishing feature of the Imbellus design is that it captures process data (the sequence and timing of your actions) alongside the outcome. In practice this means two candidates who both build a viable ecosystem can score differently if one arrived there systematically and the other by trial and error. McKinsey does not publish a pass mark or a cut-off, and the bar is not a single global number: it varies by office, role and intake, and the game result is read alongside the rest of your application rather than in isolation. Treat any specific threshold you find online as a guess.

Worked Example: Structuring an Ambiguous Problem

The habit Solve rewards (structure before action) is the same habit McKinsey's case interviews reward, which is the best argument for practising cases even though you cannot practise the game. The example below is a McKinsey-style case opening.

What You Can Prepare, and What You Cannot

Being honest about this is more useful than selling you a shortcut. You cannot practise Solve the way you practise a numerical reasoning test: there is no question bank of real items, the scenarios evolve, and the assessment is deliberately built to resist memorised patterns. Products that claim to replicate it exactly are claiming more than they can deliver.

• Can prepare: the interface. Understand what each scenario asks so you are not learning the controls while the clock runs. Read McKinsey's own candidate materials, which they provide.
• Can prepare: your condition. Sleep, food, a quiet room, a stable connection, no interruptions. McKinsey's own guidance to candidates is essentially this: the game measures cognition, and cognition degrades sharply when you are tired or rushed.
• Can prepare: general reasoning fitness. Constraint problems, logical and inductive reasoning, and pattern work keep the underlying machinery warm even though they do not replicate the game. Our logical reasoning and free cognitive ability test are useful for that, not as a simulation of Solve.
• Cannot prepare: the answers. There is no correct build to memorise, and someone else's ecosystem from a YouTube video will not be your ecosystem.
• Cannot prepare: your way around process tracking. Faking a systematic process you are not actually running tends to make you slower, not better-scored.

Your Preparation Plan

1. 1Two weeks out: understand the scenarios. Read McKinsey's candidate materials until the ecosystem and plant-defence tasks are familiar as concepts. Aim to remove surprise, not to memorise solutions.
2. 2Two weeks out: start cases. The case interview is the stage that genuinely rewards practice, and it is right behind the game. Begin working through cases with a partner now rather than after the game result arrives.
3. 3One week out: reasoning fitness. Short daily sessions of logical and inductive reasoning. Twenty minutes, not two hours.
4. 4One week out: build the PEI stories. McKinsey probes a single experience extremely deeply. Pick your three or four strongest stories and prepare them with unusual depth, because you will get ten or more follow-up questions on one of them.
5. 5The day before: stop. No cramming. Confirm your device, your browser and your connection. Book a quiet room.
6. 6On the day: read everything before you act. Instructions first, then plan, then move. Time spent understanding the constraint is not time lost.

Common Mistakes

• Trying to game the game. Copying a strategy from a video for a scenario version you will not sit wastes preparation time you could have spent on cases.
• Acting before reading. The instructions contain the constraints. Skimming them and starting to place species is the fastest way to build an ecosystem that collapses.
• Optimising one element. The strongest single species does not make a self-sustaining system. Think in relationships.
• Refusing to revise. When the evidence says your rule is wrong, change the rule. Cognitive flexibility is explicitly what is being measured.
• Neglecting the case interviews. The game is one stage; four or five case interviews follow it. McKinsey cases are more quantitative than most, so practise mental maths to two significant figures.
• Preparing shallow PEI stories. 'Tell me about a time you led a team' will be followed by ten or more probing questions. Depth beats breadth.

After the Game: Cases and the PEI

The Personal Experience Interview is McKinsey's behavioural stage and it takes about five minutes inside each case interview. It looks for personal impact, entrepreneurial drive and inclusive leadership, and it probes one story in extreme depth rather than sampling many. Prepare three or four experiences you can talk about for fifteen minutes each without running out of specifics: what you actually did, what you decided, what you would do differently.

• For the case side, our McKinsey case interview guide covers structuring, maths and the closing recommendation in depth.
• For the PEI side, the STAR method guide gives you a structure that survives ten follow-up questions.
• Process, roles and reported assessments for the firm sit on the McKinsey company page.

Frequently Asked Questions