Top 10 McKinsey Solve Assessment Questions and Answers 2026: Beat the Game-Based Screen and Reach Your Case Interview

If you just got an email inviting you to McKinsey Solve, congratulations. You cleared the resume screen, and now you’re at the real gate between applying and ever talking to a human.

Here’s the first thing to get straight. McKinsey Solve is a browser-based, game-style assessment, not a recorded video interview. Plenty of people search for a McKinsey HireVue, but it doesn’t exist in McKinsey’s standard consulting process. There are no webcam questions, no one-way video, no “tell me about yourself” on camera. If you want that kind of prep for later rounds, our guide to McKinsey interview questions covers the human stage.

What trips people up is the scoring. Solve grades you twice: a product score for whether your answers are right, and a process score built from how you got there, meaning your clicks, your scrolls, your calculator use, and how you move through the screen. You can land the correct answer and still look sloppy to the system.

This guide walks you through all three modules, Redrock Study, Sea Wolf, and the new Sustainable Futures Lab, with 10 of the most-reported question types and a specific way to approach each one. No fluff, just what actually moves your score.

Understanding McKinsey’s Solve Assessment Process

McKinsey Solve sits right after the resume screen and before any first-round case interview. It’s the primary filter between your application and a human conversation, so treat it like the real round it is.

The format is two or three interactive simulations taken remotely with no live interviewer. Redrock Study is a data-analysis and mini-case world. Sea Wolf is an optimization puzzle where you pick microbes under constraints. Sustainable Futures Lab, added in March 2026, is a 20-minute scenario-judgment module with 13 questions, per Hacking the Case Interview.

Your invitation email tells you which version you have. The 65-minute version is Redrock plus Sea Wolf; the 85-minute version adds Sustainable Futures Lab, according to Road to Offer. You usually get a 5 to 7 day window to complete it.

Nearly all generalist consulting roles get Solve: Business Analyst for undergrads, Associate for MBAs, and most generalist hires globally. QuantumBlack data science roles substitute a coding test instead, and some specialist roles get a role-specific assessment. If you’re on the data track, the QuantumBlack guide is the one to read.

One thing to retire from your prep: the old Imbellus ecosystem-building game is gone for nearly all candidates as of mid-2025. If a guide you’re reading leads with coral reefs and food chains, it’s out of date.

• Modules: Redrock Study, Sea Wolf, and (85-min version) Sustainable Futures Lab
• Duration: 65 minutes (two games) or 85 minutes (three games)
• Redrock Study: ~35 min: multi-part data analysis, 6 standalone mini-cases, 1 written report
• Sea Wolf: ~30 min: 3 optimization sites, pick 3 microbes each
• Sustainable Futures Lab: ~20 min: 13 questions (1 drag-and-rank, 12 scenario multiple-choice)
• Retakes: None; one practice tutorial opens each module
• Scoring: Dual: product score (accuracy) + process score (behavioral telemetry)
• Deadline: Typically 5 to 7 days from the invitation email

Top 10 McKinsey HireVue Questions With Sample Answers

1. Redrock Study, Investigate Phase: Read a multi-exhibit article and drag the most relevant data points into your research journal.

What they’re really testing: They’re testing hypothesis-driven reading and whether you can filter signal from noise fast, without bouncing between exhibits in a way the system flags.

Interview Guys Tip: Situation: I open Redrock’s investigate phase with [a multi-exhibit article on a geological or business scenario] and one report question to answer. Obstacle: there’s far more data on screen than I need, and every jump between exhibits gets logged against my process score. Action: I read the report question first and turn it into a working hypothesis, something like [the decline is driven by factor X, not factor Y]. Then I scan each exhibit once, dragging only the facts that confirm or kill that hypothesis into my journal, and I label every entry with [the variable it supports]. Result: my journal stays clean and organized, I avoid repeated back-and-forth between exhibits, and I enter the analysis phase already holding the three or four numbers I actually need.

Avoid this: Reading everything twice and dragging in facts “just in case” bloats your journal and racks up the backtracking the telemetry penalizes.

2. Redrock Study, Analysis Phase: Use the digital calculator to compute the percentage change between two values from the exhibits.

What they’re really testing: This checks quantitative accuracy under time pressure, and the calculator use itself is tracked, so skipping it looks suspicious to the process score.

Interview Guys Tip: Situation: I’m asked to calculate [the year-over-year change between two revenue or resource figures] from the exhibits. Obstacle: the math is easy enough that I’d be tempted to do it in my head, but the system logs no calculator activity if I do, which can flag even a correct answer. Action: I pull the two exact values straight from [the relevant exhibit], type both into the built-in calculator, and run the full percentage-change formula there, (new minus old) divided by old, times 100. I never use a personal calculator or a spreadsheet, since that’s grounds for disqualification. Result: I get a clean, accurate number with calculator activity logged, so the answer is both correct and backed by visible, methodical work.

Avoid this: Doing mental arithmetic to save a few seconds. A right answer with zero calculator keystrokes can read as a guess.

3. Redrock Study, Report Phase: Based on your journal and analysis, write a brief recommendation identifying the primary driver and the action you’d take.

What they’re really testing: This is the deliverable McKinsey consultants actually produce. They want structured synthesis and a clear, defensible answer, not a wall of observations.

Interview Guys Tip: Situation: After the analysis phase, I have to write [a short recommendation on the core problem] using my journal notes. Obstacle: it’s easy to dump everything I found and bury the point. Action: I lead with the answer first, stating [the recommended action] in one line, then name [the primary driver] the data supports, then back it with the two or three specific numbers I tagged earlier. I close by flagging [the one assumption or risk] worth noting. Result: the report is answer-first, tightly structured, and traceable straight back to my labeled journal entries, which is exactly how a McKinsey deck lands with a client.

Avoid this: Writing a narrative that lists what you noticed and makes the reader hunt for your recommendation. Put the conclusion in the first sentence.

4. Redrock Study, Case Section: Six standalone mini-cases, each asking which option is most cost-effective, highest-impact, or most consistent with the exhibit data. Multiple choice.

What they’re really testing: This measures speed, accuracy, and decisiveness. Each case is short and must be answered quickly, so hesitation and backtracking cost you.

Interview Guys Tip: Situation: I hit six short cases, each with its own exhibit and a multiple-choice question like [which option is most cost-effective]. Obstacle: there’s real time pressure, and overthinking one case eats into the next. Action: for each, I read the question stem first so I know what I’m solving for, then I find [the single metric the answer hinges on] in the exhibit, run a quick calculator check if numbers are involved, eliminate the two weakest options, and commit. I don’t revisit a case once I’ve answered it. Result: I move at a steady, decisive pace, keep my navigation clean, and leave margin for Sea Wolf instead of burning time double-checking cases I already solved.

Avoid this: Treating each mini-case like a full framework exercise. These reward one sharp calculation and a committed choice, not a structured walkthrough.

5. Sea Wolf, Site Setup: For Site #1 you have 10 microbes with 5 attributes each. The site needs an average resilience between 4.0 and 6.0, average growth above 3.5, and no ‘toxic’ trait. Pick exactly 3 microbes that satisfy every constraint at once.

What they’re really testing: This is pure optimization and constraint management. They want systematic set-level thinking, not one-microbe-at-a-time filtering.

Interview Guys Tip: Situation: Site #1 gives me 10 microbes and a set of averaged constraints, like [resilience averaging 4.0 to 6.0 and growth above 3.5], plus a hard exclusion on the toxic trait. Obstacle: the trap is filtering microbes individually, which breaks because the rules apply to the average of my chosen three, not to each one. Action: first I drop every microbe carrying [the excluded trait]. Then I think in set terms: for a three-microbe average, I need the group totals to land in range, so I target combinations where a lower-growth microbe is balanced by two stronger ones. I test one or two candidate trios against all constraints simultaneously before committing. Result: I find a valid set quickly, without brute-forcing every combination, and I’ve built a method I can reuse on the next site.

Avoid this: Rejecting individual microbes because one of their stats looks low. A “weak” microbe can be perfect once you average it with two strong ones.

6. Sea Wolf, Across Three Sites: Each site has different attribute ranges and trait rules. Can you keep a consistent, efficient process as the constraints change, without running out of time?

What they’re really testing: This tests process discipline and time management. The scoring rewards candidates who apply the same method regardless of the specific constraints.

Interview Guys Tip: Situation: I have three sites, each with its own attribute ranges and trait exclusions, inside a single 30-minute budget with no per-site clock. Obstacle: there’s no timer per site, so it’s easy to over-invest in Site 1 and get squeezed on Sites 2 and 3. Action: I give myself a rough pace of about [10 minutes per site] and run the exact same four-step routine each time: cut excluded traits, read the averaged constraints, target a balanced trio, verify against all rules, commit. I don’t reinvent my approach when the numbers change. Result: my process stays identical and efficient across all three sites, which is exactly what the telemetry rewards, and I finish with time to spare instead of panic-guessing the last site.

Avoid this: Treating Site 1 like a puzzle you must perfect. Slow, luxurious solving early leaves you guessing later, and guessing shows up in the process score.

7. Sustainable Futures Lab, Prioritization (Question 1): A sustainability project has four possible first actions. Drag and rank them from highest to lowest priority based on the scenario.

What they’re really testing: They want to see you set strategic direction early, spot the primary impact driver, and make a call with incomplete information.

Interview Guys Tip: Situation: I’m handed [a sustainability consulting scenario] and four candidate first actions to rank. Obstacle: the information is incomplete and more than one action looks reasonable. Action: I anchor on [the project’s stated primary objective], then rank each action by how directly it moves that objective and how much it unblocks the others. The action that drives the biggest impact or removes a dependency goes to the top; the nice-to-have that depends on earlier steps goes to the bottom. I commit rather than agonize over the middle two. Result: my ranking reflects a clear logic (highest impact and fewest dependencies first), which is the strategic judgment the module is built to detect.

Avoid this: Ranking by what feels urgent or easy. Prioritize by impact on the core objective, not by what’s quickest to do.

8. Sustainable Futures Lab, Scenario Judgment: A stakeholder pushes back on your recommended approach during an early meeting. Which response best balances keeping your analytical recommendation and preserving the client relationship?

What they’re really testing: This is stakeholder judgment under pushback. They’re testing whether you can hold your evidence and the relationship at the same time.

Interview Guys Tip: Situation: A stakeholder challenges [my recommended approach] in an early project meeting. Obstacle: I can either cave on my analysis or dig in and bruise the relationship, and both are wrong. Action: I pick the response that acknowledges their concern specifically, restates the evidence behind my recommendation in plain terms, and proposes a concrete way to test the disagreement together, like [reviewing the underlying data in the next working session]. That keeps the recommendation intact while treating the stakeholder as a partner, not an obstacle. Result: I preserve both my analytical position and the working relationship, which is exactly the balance McKinsey looks for. Our leadership interview guide with SOAR answers drills this same balance for the human rounds.

Avoid this: Choosing the option that abandons your recommendation to keep the peace, or the one that overrides the stakeholder. The best answer does both at once.

9. Sustainable Futures Lab, Trade-Off Judgment: Option A has higher expected value, Option B has lower variance, and the client is risk-averse. Which do you recommend, and why?

What they’re really testing: This checks whether you fold client context into a quantitative recommendation instead of reflexively chasing the bigger number.

Interview Guys Tip: Situation: The data shows [Option A has higher expected value] while [Option B carries lower variance], and the scenario states the client is risk-averse. Obstacle: the tempting move is to pick the higher expected value and ignore the stated risk preference. Action: I weigh the client context as a real constraint, not a footnote. Because the client prioritizes predictability, I recommend [Option B], and my rationale names the reason explicitly: the lower variance better matches their risk tolerance even though it sacrifices some expected upside. Result: my choice fits the client’s actual situation, which is the central consulting skill here, matching the recommendation to the decision-maker in the room, not to the spreadsheet in a vacuum.

Avoid this: Defaulting to the highest expected value every time. The scenario hands you the client’s risk appetite on purpose, so use it.

10. Sustainable Futures Lab, Uncertainty Handling: Two data sources in the scenario conflict. What’s your next step before making a recommendation? Options range from ‘proceed with best estimate’ to ‘gather more data’ to ‘flag the conflict to the team.’

What they’re really testing: This is metacognition and intellectual humility. McKinsey values candidates who recognize the limits of their evidence instead of forcing a conclusion.

Interview Guys Tip: Situation: Two sources in the scenario give me [conflicting signals about the same variable]. Obstacle: I could barrel ahead with a best guess, but a recommendation built on an unresolved contradiction is fragile. Action: rather than proceed blindly or stall completely, I pick the response that surfaces the conflict, flagging it and, where the options allow, reconciling which source is more reliable before committing. I show I noticed the discrepancy instead of papering over it. Result: I demonstrate the intellectual honesty the module rewards, recognizing when the evidence isn’t settled and handling it deliberately. That awareness carries into panel and group rounds too, which our panel interview guide gets into.

Avoid this: Picking ‘proceed with best estimate’ to look decisive. Forcing a conclusion over a known contradiction reads as low self-awareness, not confidence.

5 Insider Tips for Acing It

Interview Guys Tip: Let the calculator do everything

The system logs calculator keystrokes as part of your process score, and a correct answer with no calculator activity can get flagged as a guess. Use the built-in tool for every single calculation, even the ones you could do in your head. Never touch a personal calculator or spreadsheet; that’s a disqualification line, not a gray area.

Stop backtracking in Redrock

Every time you flip between exhibits, the telemetry sees it, and excessive back-and-forth drags your process score down. Build a hypothesis from the report question first, then make one clean pass, dragging only what matters. A disciplined path beats a thorough-but-frantic one.

Think in sets, not singles, on Sea Wolf

The biggest Sea Wolf mistake is filtering microbes one by one by their individual stats. The constraints apply to the average of your three picks, so compute what the group needs to total and build toward that. A “weak” microbe is often the perfect third piece.

Interview Guys Tip: Budget your time across games, not within one

Redrock feels the most familiar, so people overspend there and get crushed on Sea Wolf and Sustainable Futures Lab. Decide your rough time budget per module before you start and hold the line. Finishing all games steadily beats acing one and guessing the rest.

Run the tutorial like it’s real

Each module opens with a short practice so you can learn the drag-and-drop and microbe-selection interface. Use it. Meeting the UI for the first time during the scored portion wastes seconds you can’t get back, and no outside AI tool is allowed to bail you out mid-game. For the AI-era rules more broadly, our AI interview questions guide is worth a skim.

Your Technical Setup Checklist

• Use a reliable laptop or desktop with an up-to-date browser; Solve is browser-based, so a stable setup matters more than any webcam.
• Get on the most stable internet connection you have, ideally wired, so a dropout doesn’t eat your timed module.
• Clear your desk of personal calculators, phones, and notes. The only calculator you may use is the one built into the assessment.
• Do not use any outside tools, spreadsheets, or AI assistance; McKinsey’s rules treat this as grounds for disqualification.
• Pick a quiet block of 65 to 85 uninterrupted minutes, since there are no retakes and no pausing mid-game.
• Have scratch paper and a pen ready for your own rough notes, and keep a glass of water nearby so you don’t break focus.

What Happens After You Submit

Once you submit, there’s nothing to do but wait. McKinsey doesn’t publish score thresholds or the exact weighting between product and process, so you won’t get a breakdown, just a decision on whether you advance to first-round cases.

Turnaround varies by office and recruiting cycle, from a few days to a few weeks. If a recruiter gave you a timeline, hold them to it with one polite, brief follow-up after that date passes. A short, warm nudge is fine, and the same etiquette applies to any recruiter screen; our phone interview questions guide covers that conversation.

If you pass, start case prep immediately, because the human rounds move quickly after Solve. If you don’t, remember roughly 70 to 80 percent of candidates get cut here, so it’s a brutal filter, not a verdict on your worth.

Frequently Asked Questions

Is McKinsey Solve a HireVue video interview?

No. There’s no verified evidence McKinsey uses HireVue or any one-way video platform in its standard consulting process. Solve is a game-based, browser assessment with simulation modules, no webcam questions and no recorded answers. The video-style prep comes later, in your live case and behavioral rounds.

Can I retake McKinsey Solve if I do badly?

No, there are no retakes. Each module does start with a short practice tutorial so you can learn the interface, but once you begin the scored portion it counts. Treat the tutorial seriously and go in already familiar with the question types.

How long does the assessment take and how many games are there?

Either 65 minutes with two games (Redrock Study plus Sea Wolf) or 85 minutes with three games (adding Sustainable Futures Lab). Your invitation email states which version you have. Redrock runs about 35 minutes, Sea Wolf about 30, and Sustainable Futures Lab about 20 with 13 questions.

What’s the pass rate, and what gets measured?

Prep providers and candidate reports converge on roughly 20 to 30 percent passing, per Hacking the Case Interview, though McKinsey doesn’t publish an official figure. You’re scored twice: a product score for accuracy and a process score built from behavioral telemetry like clicks, scrolls, calculator use, and navigation.

How is this different from a behavioral interview?

Solve is a timed simulation that measures how you solve problems through your actions, not how you tell stories. There’s no SOAR answer to speak aloud here. Save your structured storytelling for the human rounds, which look more like the behavioral questions top employers ask, and prep those separately.

Here’s the honest version: McKinsey Solve isn’t about being a genius, it’s about being disciplined. Use the calculator for everything, build a hypothesis before you read, think in sets on Sea Wolf, pace yourself across all the games, and commit to well-reasoned choices instead of thrashing. The process score is watching your habits, so make your habits clean.

Clear this gate and the human rounds come fast, so don’t let momentum stall. Line up your case practice and your story prep now, starting with our guide to answering ‘tell me about yourself’, so when the recruiter calls you’re already two steps ahead. You’ve got this.

Sources and Further Reading

• McKinsey Official Solve Assessment Page (careers.mckinsey.com)
• Hacking the Case Interview: McKinsey Solve Complete Guide (2026)
• My Consulting Coach: McKinsey Solve 2026, Redrock, Sea Wolf & SFL Prep
• Road to Offer: McKinsey Solve Assessment 2026 Format, Modules & Strategy
• PSG Secrets: What Reddit Really Says About the McKinsey Solve Game
• Case Interview Hub: McKinsey Solve Game in 2026 (former McKinsey interviewer breakdown)
• McKinsey QuantumBlack Interview Guide, coding assessment format for data science roles