How to prepare for McKinsey Solve

1. What McKinsey Solve actually is

McKinsey Solve — historically called the Problem Solving Game (PSG) and sometimes the Imbellus assessment — is a ~70-minute, gamified screen sent to most McKinsey candidates before the case interview. It currently presents two scenarios: Ecosystem Building and Redrock Study.

Despite the wildlife dressing, it is not a biology test. It measures five cognitive traits McKinsey cares about: critical thinking, decision making, meta-cognition, situational awareness, and systems thinking. Every click is instrumented; the process matters as much as the answer.

2. The Ecosystem scenario

You are shown a terrain (mountain, reef, jungle, etc.) with environmental parameters — temperature, depth, sunlight, salinity. You must build a sustainable food chain by selecting eight species from roughly forty candidates, subject to two hard constraints:

• Every species' environmental tolerances must overlap the location.
• The food chain must produce a net calorie surplus across producers, herbivores, and carnivores.

The trap is treating it as a sorting exercise. Strong candidates start from the calorie equation, work backwards to required producer biomass, then filter for tolerance — not the other way round.

3. The Redrock scenario

Redrock replaced the older Plant Defense and Disaster Management scenarios. You are a researcher studying an animal population in a national park, working through four chapters: investigation, analysis, report, and cases. Each chapter feeds the next — your selected data points in chapter one constrain what you can compute in chapter two, and so on.

It is mostly quantitative reasoning over messy data: weighted averages, growth rates, sampling. The hard part is not the maths but resisting the urge to grab every interesting fact instead of the few that answer the actual question.

4. How scoring really works

The exact algorithm is proprietary, but the structure is well-understood. Two parallel scores are computed for each scenario:

• Product score — did your final answer satisfy the constraints?
• Process score — did your clicks, time spent, and revisions look like a structured problem-solver's?

You can get the right answer with a poor process and still fail. You can also get parts of the answer wrong and pass, if your process was clean. This is why "memorise the optimal food chain" advice fails — McKinsey watches the path, not just the destination.

5. The five traps that sink candidates

1. Treating it as a quiz. It rewards a hypothesis-first loop, not pattern matching.
2. Ignoring time discipline. Allocate ~35 minutes per scenario and stick to it; running out on Redrock chapter four is fatal.
3. Hoarding evidence. Every irrelevant data point you collect is a process-score penalty in disguise.
4. Forgetting the constraint. In Ecosystem, the calorie equation is the constraint — most failures violate it.
5. Not revising. Refusing to overturn an early hypothesis when new evidence arrives is the single biggest predictor of a low meta-cognition score.

6. A 3-week practice plan

7. How to practise without the official game

McKinsey does not release the official Solve game for practice, and most prep providers sell videos and PDFs — useful for mechanics, useless for building the actual reasoning loop. The only way to train the loop is to do scenario-based modules with debriefs that explain why each decision worked.

STRATAGEM was built for this. Two modules mirror the Solve scenarios:

• Ecosystem — a Solve-style food-chain construction module with explicit calorie and tolerance constraints, a hypothesis prompt before you build, and a causal debrief that traces every species choice through to your final score.
• Wildfire — a Redrock-style multi-chapter investigation under a 35-minute clock, with chapter-to-chapter dependencies and a debrief that flags every irrelevant data point you collected.

Both produce a single SAR score so you can track real improvement across sessions instead of guessing whether you're getting better.

8. FAQ