McKinsey Solve: Complete Guide to the Problem-Solving Game 2026

McKinsey Solve—the firm’s digital assessment game—has replaced traditional screening tests and now stands as the first major hurdle in the McKinsey interview process. Unlike resumes or case interviews, Solve evaluates how you think through complex problems in real time by tracking every mouse movement, click pattern, and decision sequence.

Over 70% of candidates do not pass Solve on their first attempt. The game is not about memorizing answers or business frameworks—it tests raw cognitive abilities through simulated ecosystems and data challenges that change for every candidate.

What McKinsey Solve Tests

Solve measures five cognitive abilities that McKinsey considers essential for consulting work:

McKinsey calculates a “process score” by analyzing your approach—not just your final answers. The algorithm detects whether you followed a systematic reasoning process or simply guessed. A candidate who uses disciplined logic but runs out of time will score higher than someone who stumbles upon correct answers through trial and error.

Game Format and Structure

You have 71 minutes total to complete:

• An untimed tutorial (review as long as needed)
• Two mini-games from three possible categories

Each mini-game takes 30-45 minutes. The scenarios are randomized, so your version will differ from other candidates’. This makes it impossible to copy strategies from others—you must understand the underlying principles.

Three Types of Mini-Games

McKinsey rotates between three mini-game types, and you’ll face two of them:

mindmap
  root((Solve Mini-Games))
    Constrained Optimization
      Ecosystem Building
        Choose species
        Select location
        Balance food chains
      Disaster Management
        Identify disaster type
        Move animals to safety
    Data Interpretation
      Red Rock Study
        Analyze research data
        Forecast populations
        Adapt to new info
    Cause vs Effect
      Disease Diagnosis
        Identify symptoms
        Reverse engineer cause
        Recommend treatment

1. Constrained Optimization (most common)

You build a sustainable ecosystem by selecting species and a location. Each species has specific environmental needs (temperature, rainfall, elevation) and a role in the food chain. Your goal: maximize species survival while respecting constraints.

The challenge is balancing competing requirements. An apex predator needs prey to survive, but if you add too many predators, they’ll deplete the prey population. If your primary producers can’t survive in the chosen location, the entire food chain collapses.

2. Data Interpretation + Critical Numerical Reasoning (introduced 2024)

The Red Rock Study presents a research scenario—typically ecological forecasting based on provided data. This replaced McKinsey’s discontinued Problem Solving Test and focuses on skills that standard math tests miss: reading charts correctly, distinguishing conclusive evidence from suggestive patterns, and stating logically sound conclusions.

3. Cause vs. Effect (less common now)

Given an outcome (like animal population decline) and historical data, you reverse-engineer the cause. Or, given a cause (like a price reduction), you predict effects using pattern analysis. This tests whether you can identify relevant data signals among noise.

Proven Strategies by Mini-Game Type

For Constrained Optimization (Ecosystem Building)

Start at one end of the food chain, not the middle. Begin with primary producers (plants that need no prey) or apex predators (animals that have no predators), then work linearly. Starting in the middle forces you to guess dependencies in both directions simultaneously.

Identify and ignore irrelevant variables. You might see 8-10 environmental factors, but often only 3-4 actually constrain your choices. If all available species can tolerate the full range of wind speeds in your location options, wind speed is irrelevant—ignore it.

Track your trials on paper. Each iteration of trial-and-error consumes time. Write down what you’ve tested and why it failed. The game allows pen and paper—use them to avoid retesting the same hypothesis.

Note maximums and minimums. Rules like “temperature must be between 15-25°C” or “at least 500 calories of prey per predator” create the constraints. Quickly identify which metrics have thresholds and design your system around them.

Here’s a systematic approach to ecosystem building:

flowchart TD
    A[Review All Species] --> B{Choose Approach}
    B -->|Top-Down| C[Select Apex Predator]
    B -->|Bottom-Up| D[Select Primary Producers]
    C --> E[Identify Prey Requirements]
    D --> F[Identify Environmental Needs]
    E --> G[Add Supporting Species]
    F --> G
    G --> H[Check Constraints]
    H -->|Violation| I[Adjust Selection]
    H -->|Pass| J[Choose Location]
    I --> H
    J --> K[Verify All Species Survive]
    K -->|Some Die| I
    K -->|All Survive| L[Done]

For Data Interpretation (Red Rock Study)

Go slow and read new information carefully. Each turn introduces new data that may contradict your initial hypothesis. Rushing through without absorbing updates leads to persistent errors.

Write down assumptions and test them. When you form a hypothesis about population trends, note it on paper. When new data arrives, explicitly check if your assumption still holds. Candidates who adapt quickly score higher than those who cling to disproven strategies.

Distinguish “suggests” from “proves.” Data may indicate a correlation without proving causation. Your conclusions must match the strength of evidence. Overstating findings (claiming certainty from suggestive data) is a critical error.

For Cause vs. Effect (Disease Diagnosis)

Map relationships between variables first. Before diving into data, understand the logical connections: Does rainfall affect temperature? Does elevation constrain species distribution? Sketch these relationships to guide your analysis.

Separate relevant from irrelevant data. You’ll receive far more information than needed. In a real consulting project, clients often provide hundreds of documents when only one category matters. Practice filtering noise.

Common Mistakes That Lower Your Score

Based on our analysis of 835+ case interview patterns and candidate feedback:

Mistake 1: Optimizing locally instead of globally. You might perfect one part of the ecosystem (e.g., herbivore-plant balance) while ignoring how it affects the entire food chain. Always step back and check system-wide implications.

Mistake 2: Ignoring the clock. While finishing early isn’t required, spending 35 minutes on iteration #1 leaves too little time for adjustments. Aim for rough correctness quickly, then refine.

Mistake 3: Not using pen and paper. Mental tracking fails under time pressure. Candidates who take structured notes avoid repeating failed attempts and spot patterns faster.

Mistake 4: Treating it like a traditional game. Solve rewards systematic thinking, not gaming instincts. Your process matters more than your final score. Show disciplined reasoning through your click patterns.

Mistake 5: Trying to memorize other candidates’ solutions. The randomization makes this useless. Two candidates might face different species sets, location variables, and constraints. Focus on principles, not specific answers.

How to Prepare for McKinsey Solve

McKinsey does not allow practice access to the actual game. However, you can prepare by:

1. Practice systems thinking. Work through problems involving multiple interacting variables—supply chain optimization, ecological simulations, or resource allocation under constraints. Train your ability to see second-order effects.

2. Improve data interpretation speed. Review charts and datasets, then practice stating precise conclusions. What does the data definitively show? What does it merely suggest? Where is the evidence insufficient?

3. Simulate time pressure. Set 35-minute limits for complex problem-solving exercises. Practice making “good enough” decisions quickly rather than perfect decisions slowly.

4. Use the untimed tutorial strategically. When you start Solve, the tutorial does not count against your time. Read instructions twice. Take notes. Understand the objective before the clock starts.

Solve’s Role in McKinsey Screening

McKinsey uses Solve scores in two ways:

• Eliminate low scorers with strong resumes: A poor Solve performance can override excellent academics
• Identify high scorers with weaker backgrounds: A strong Solve result can compensate for less prestigious schools

The game does not replace resume screening or case interviews—it supplements them. Even a perfect Solve score requires solid academic credentials and interview performance to secure an offer.

Candidates who pass Solve proceed to first-round interviews, typically one case interview and one experience interview. Preparing for both is essential, as Solve alone does not guarantee success.

Key Takeaways

• McKinsey Solve tests cognitive abilities through randomized simulations—memorizing answers does not work
• The “process score” evaluates your reasoning approach, not just final answers, by tracking every click
• Start ecosystem problems at one end of the food chain (top or bottom), never the middle
• Use pen and paper to track trials, note constraints, and avoid repeating failed hypotheses
• Data interpretation requires distinguishing what data proves versus what it merely suggests
• Over 70% fail on first attempt; systematic preparation improves success rates significantly

Practice Real Case Scenarios with AI Feedback

While McKinsey Solve tests cognitive abilities through simulations, real case interviews require structured frameworks and business judgment. CasesCoach offers two complementary ways to build these skills:

Guided Mode provides interviewer hints and direction as you work through cases—ideal for learning framework application and building confidence with structured problem-solving. Our AI coach adapts to your experience level and guides you toward complete answers.

Expert Mode simulates the pressure of real McKinsey interviews with minimal guidance and time constraints. This mode tests whether you can independently structure problems, prioritize issues, and communicate insights clearly—exactly what you’ll face after passing Solve.

Both modes provide instant AI feedback across five dimensions: Structure, Analysis, Communication, Business Acumen, and Quantitative Skills. You can also use Chat with Case to explore real business problems through our library of 835+ consulting cases, asking questions and testing approaches as you would with a mentor.

Start your free AI Mock Interview with 3 complete case practice sessions, or explore our full case library to see the types of problems top-tier consultants solve daily. Passing Solve opens the door—mastering case interviews gets you the offer.