McKinsey Solve Assessment Practice: What the Ecosystem and Red Rock Games Test

The McKinsey Solve assessment (McKinsey's own name for what candidates still call the Problem Solving Game or the Imbellus test) is a set of ecology-themed simulation games that measure how you solve problems, not what you already know. As of 2026, most candidates play a build-an-ecosystem game and a data-analysis case game (Red Rock), often with a third microbe-selection game (Sea Wolf) in the mix, over a window that commonly runs 65 to 105 minutes depending on the version you are assigned. It is scored two ways at once: a product score for the quality of your final answers, and a process score for how you got there, meaning McKinsey tracks your clicks, your sequence, and how long you spend before acting. Roughly 70% of candidates are screened out at this stage, so it is a real gate, not a formality.

This guide explains each game, how the dual scoring actually works, what the current 2026 format looks like (it has been shifting), and how to practice for it when your assessment invite is already in your inbox. Every format claim here is attributed, because McKinsey changes this test regularly and confident-but-stale detail is the most common way prep pages mislead you.

Quick takeaways

• Solve is a gamified assessment created by McKinsey to showcase problem-solving ability, played on a computer through realistic ecology scenarios, per McKinsey's own careers page.
• The current 2026 lineup centers on Ecosystem Building and Red Rock Study, with Sea Wolf (a newer microbe-selection game) appearing in some versions and older mini-games largely retired.
• Total time varies by version: reported windows include roughly 65 minutes, 85 minutes, and up to 105 to 110 minutes, so confirm what your invite says.
• You get two scores: a product score (were your answers right) and a process score (how efficiently and sensibly you worked). Both matter.
• About 70% of candidates do not pass this stage, based on aggregated candidate reporting, so it screens hard.
• You can absolutely practice. The games reward a repeatable method (read the whole scenario first, organize your data, watch the clock), and that method is trainable with realistic simulations.

What "Solve" is, in McKinsey's own words

McKinsey describes Solve as "a gamified assessment created to showcase your problem-solving abilities," and it is now a standard step in the consultant recruitment process. The important design choice is that Solve is deliberately not a business-knowledge test. There are no case-math templates to memorize and no framework to recite. You are dropped into an unfamiliar ecological world and asked to reason your way through it.

That design is intentional. McKinsey wants to see raw problem-solving under time pressure without the advantage that goes to candidates who have crammed case frameworks. It is also why the test tracks your process: the firm cares how you approach an unfamiliar problem, because that is the actual job.

The assessment has gone through several names and versions. It was built with Imbellus, which is why older resources call it the Imbellus game; McKinsey's own term is Solve. Treat "Solve," "Problem Solving Game," "PSG," and "Imbellus test" as the same thing.

The games, one by one (2026)

McKinsey rotates and pilots games, so the exact set you face depends on your version. Here is the current lineup and what each game demands.

Game	What you do	Skills it measures	Reported time
Ecosystem Building	Pick and place a set of species so the food chain sustains itself	Systems thinking, constraint balancing, calorie/food-chain logic	~35 min
Red Rock Study	Investigate data, run calculations, build a short report, then answer mini-cases	Data filtering, numerical reasoning, prioritization	~35 min
Sea Wolf	Select microbes from a catalog to clean specific pollution types	Classification, analytical reasoning, matching under constraints	~30 min

Ecosystem Building

You are given a location with terrain data (elevation, temperature, humidity, and for reef versions, depth) and a pool of species. You choose a set (commonly eight) and place them so that every species' calorie needs are met and the whole system stays stable. Species eat in order of the calories they provide, food sources deplete, and a species dies if its needs go unmet. It rewards thinking in a system, not picking your favorite animals.

Red Rock Study

This is the analytical heart of Solve. It runs in phases: an investigation phase where you filter raw data and save the relevant pieces to a research journal; an analysis phase where you answer numerical questions using an in-game calculator; a report phase where you assemble findings and choose the right chart; and a case phase with a set of additional mini-cases (often around 10) using the same investigate-analyze-answer loop. The single most common Red Rock mistake is spending too long up front and running out of time for the case questions, so budget your minutes deliberately.

Sea Wolf

The newer game. You review a catalog of microbes with different properties and select a set (commonly ten) that can clean specific types of ocean pollution, sometimes needing microbes that combine into chains to remove a pollutant. It is a matching-under-constraints task: classify, filter, and pick the set that satisfies the requirements.

Older mini-games (Plant Defense, Disaster Management, Disease Management, Migration) show up in legacy prep material but are largely no longer the standard, per current candidate reporting. Do not spend your limited prep time on Plant Defense if your invite points to the current lineup.

How Solve is scored: product and process

This is the part that separates Solve from a normal test and the part most guides gloss over.

You are scored on two dimensions at once:

• Product score: the quality and correctness of your final outputs. Did your ecosystem survive? Were your Red Rock calculations right? Did you pick a valid microbe set?
• Process score: how you worked to get there. As one detailed breakdown puts it, "every move you make is tracked, from the data you choose to analyze to the sequence in which you take actions." The process score reflects things like where you click, how long you spend on each part, and how long you take between reading information and acting on it.

Practically, that means two candidates with the same final answers can score differently based on how they got there. A candidate who flailed, backtracked constantly, and only stumbled into the right answer looks worse than one who moved deliberately. You cannot game the process score by faking efficiency, but you can protect it by working methodically instead of frantically.

On outcomes: no official pass rate is published, but aggregated candidate reporting consistently points to roughly 70% of candidates being screened out at this stage, and performance is weighed alongside your CV and application, not in isolation. If you do not pass, McKinsey typically enforces a waiting period (commonly around a year) before you can reapply.

Worked example: budgeting your time in Red Rock

Time management is where Red Rock is won or lost, so here is a concrete way to run it.

Say your Red Rock block is 35 minutes and ends with 10 case questions. The trap is treating the investigation and analysis phases as open-ended and arriving at the case questions with five minutes left. The fix is to split your time before you start:

• First few minutes: skim the whole scenario and the research question so you know what data actually matters. Do not start filtering blindly.
• Investigation and analysis: filter to the relevant data, save it to your journal, and answer the numerical questions. Keep moving; a "good enough, correct" answer now beats a perfect answer you never reach.
• Reserve roughly half your total block for the case questions. They carry real weight and there are many of them, so protect that time hard.

The candidates who clear Red Rock are almost never the ones who computed the most beautiful answer in phase one. They are the ones who paced themselves so they actually finished. That is a process-score win and a product-score win at the same time.

Which employers use Solve, and what that tells you

Solve is McKinsey's assessment, used in its consultant recruitment funnel worldwide. BCG and Bain run their own comparable gamified or written screens (BCG's Casey/online case and Bain's tests), so if you are recruiting across the top consulting firms, expect a different assessment at each, not one shared test. That is worth knowing because prep that transfers on principles (structured problem-solving, data prioritization, time discipline) does not transfer as a memorized game walkthrough. Practice the method, not one firm's exact levels.

Because Solve is scored against a strong applicant pool and weighed alongside your resume, the realistic goal is not just to "pass" but to post a clean, deliberate performance that holds up next to a competitive CV. That is the employer-data reality of this test: it is one filter in a stack, and it screens out most people who take it.

FAQ

How do you pass the McKinsey Solve assessment?

Read the entire scenario before acting, because Solve rewards understanding the whole system first. Work methodically rather than frantically, since your process (clicks, sequence, timing) is scored alongside your answers. Budget your time hard, especially in Red Rock, and reserve roughly half your block for the case questions. Practice with realistic simulations so the interface and the ecology logic are familiar, which frees your attention for the actual problem on test day.

How hard is the McKinsey Solve test?

Hard enough that aggregated candidate reporting points to roughly 70% being screened out. The difficulty is not business knowledge, since there is none to memorize; it is solving unfamiliar problems under a tight clock while a dual scoring system watches how you work. Candidates who practice the games and build a time-budgeting habit generally find it manageable; those who go in cold usually run out of time in Red Rock.

What is the 80/20 rule at McKinsey?

The 80/20 rule (the Pareto principle) is the idea that about 80% of results come from about 20% of causes. At McKinsey it is shorthand for focusing on the small number of factors that drive most of the outcome instead of analyzing everything equally. It is directly useful in Solve: in Red Rock especially, identify the few data points that actually answer the question and prioritize those, rather than filtering every variable evenly and running out of time.

Can you practice for McKinsey Solve?

Yes. Despite the "you can't prepare" myth, the games reward a repeatable method (understand the system first, organize your data, watch the clock, prioritize the high-impact factors), and that method is trainable. Realistic simulations of Ecosystem Building, Red Rock, and Sea Wolf let you learn the interface and the underlying logic in advance so that on test day your effort goes into the problem, not into figuring out the controls.

How long is the McKinsey Solve assessment?

It depends on your version. Reported windows include roughly 65 minutes (Red Rock plus Sea Wolf), around 85 minutes (adding another game), and up to 105 to 110 minutes for versions that still include Ecosystem Building. Individual games run about 30 to 35 minutes each. Because McKinsey pilots and rotates versions, confirm the exact length in your assessment invitation.

What happens if I fail the McKinsey Solve assessment?

You are typically not advanced to the interview stage, and McKinsey commonly enforces a waiting period (often around a year) before you can reapply. Your Solve performance is weighed together with your CV and application rather than in pure isolation, but a weak score usually ends that cycle. That waiting period is the practical reason to prepare properly before you sit it, even on short notice.

Is McKinsey Solve the same as the Imbellus test?

Yes. Solve was built with the assessment company Imbellus, so older resources call it the Imbellus game or the Problem Solving Game (PSG). McKinsey's current name is Solve. The names refer to the same gamified, ecology-themed assessment; only the branding and the specific games in rotation have changed over time.

Related on PrepClubs

• McKinsey Solve assessment overview and format
• Consulting case interview prep guide
• McKinsey Solve vs the BCG assessment

Ready to walk in prepared?

PrepClubs gives you realistic McKinsey Solve simulations plus targeted drills for Ecosystem Building, Red Rock Study, and Sea Wolf, so the interface, the ecology logic, and the time budget are familiar before it counts, not discovered live while the clock runs. Practice the method that protects both your product score and your process score. The 30-day Pass Guarantee is simple: if you prepare with PrepClubs and do not pass your real assessment, we extend your access at no extra cost. No cash-back hedge, just more time with the material if you need it. Get McKinsey Solve access

Junaid Khalid runs PrepClubs, a practice-test platform with 1,600+ students who have prepped for cognitive and aptitude tests.