McKinsey Solve runs two timed mini-games (usually Ecosystem Building plus Red Rock Study) over roughly 60–81 minutes. The algorithm scores your decision process — the sequence of your moves, your click patterns, and how you allocate time — not just your final answer. Panicked, random clicking under the countdown is actively penalized. This guide covers budgeting time inside each game, resisting the urge to over-optimize early moves, recovering from a bad start, and keeping your nervous system calm enough to think while the clock is visible.
Pressure kills more McKinsey Solve candidates than raw difficulty does. The moment a countdown appears on screen, otherwise capable people start clicking faster, abandon their plan, and produce a chaotic decision trail — and Solve’s algorithm is watching that trail, not just the answer at the end. With roughly 70% of candidates eliminated, composure under a visible clock is not a soft skill here. It is scored.
This guide breaks down how to allocate time inside each Solve mini-game, why over-optimizing your opening moves is a trap, how to recover when you start badly, and how to keep your nervous system quiet enough to reason while the timer runs down.
What Solve Actually Is Now
McKinsey retired the paper Problem Solving Test in most markets around 2020 and replaced it with Solve, a game-based digital assessment that is now the global default for first-round screening. (A few offices may still administer the legacy PST, so confirm your specific format with your recruiter.)
A Solve session is two mini-games running roughly 60–81 minutes total. The two you will almost certainly see:
- Ecosystem Building — you are given a fixed time window to build a viable food chain or ecosystem from a set of species, respecting rules about habitat, food sources, and calories. It rewards systematic constraint-checking over guesswork.
- Red Rock Study — a longer, multi-phase timed investigation where you collect data, run analyses, and answer questions across sequential stages.
Occasionally candidates see Sea Wolf or Plant Defense instead. Whatever the pairing, the structure is the same: each game is timed, and you cannot bank time from one game to spend on another the way you could skip between PST questions.
The Scoring Shift That Changes Everything
The single most important thing to understand about Solve is that it is algorithmic and process-based. It does not simply mark a final answer right or wrong. It tracks:
- The sequence of your decisions — did you gather information before acting, or act first and correct later?
- Your click patterns — deliberate, purposeful selections versus rapid, scattered, exploratory clicking
- Your time allocation — where you spent your minutes, and whether you slowed down where the problem was genuinely hard
This is why panic is so costly. On the old PST, a wild guess on a hard question was neutral — there was no penalty for a wrong answer. On Solve, random or frantic clicking is a signal, and it reads as a weak decision process even if you happen to land on something workable. The assessment is trying to infer how you think under pressure. Chaotic behavior answers that question for you, in the wrong direction.
The practical takeaway: your job is not to look fast. It is to look deliberate. A calm candidate who makes eight considered moves usually outscores a frantic one who makes thirty.
Budgeting Time Inside a Single Game
Because you cannot move time between games, each one needs its own internal budget. The mistake most candidates make is treating the whole window as “solving time” and starting to act immediately. Split it instead.
flowchart TD
A[Game starts: countdown visible] --> B[Orient: read rules and constraints]
B --> C[Plan: identify the real constraints]
C --> D{Do I understand what a valid solution needs?}
D -->|Yes| E[Execute deliberate moves]
D -->|No| F[Re-read the binding rule, do not guess]
F --> C
E --> G{Time checkpoint: past halfway?}
G -->|On track| H[Continue, verify each move]
G -->|Behind| I[Lock in a valid-but-imperfect solution]
H --> J[Reserve final minutes to review and confirm]
I --> J
A workable mental split for any timed game is roughly:
- Orient and plan first. Read every rule before you touch anything. In Ecosystem, that means understanding the calorie, habitat, and food-source constraints before placing a single species. Rushing this is the most common self-inflicted wound — you cannot recover a plan you never made.
- Execute in the long middle. Make deliberate moves, checking each against the constraints as you go. This is where most of your window lives.
- Reserve the end to verify. Keep a genuine buffer to confirm your solution is valid, not just complete. A valid ecosystem submitted with time to spare beats a “better” one that breaks a rule at the buzzer.
Notice these are proportions, not fixed minute counts. Solve’s timings vary by game and by cohort, so anchor to phases, not to a stopwatch number someone quoted you online.
The Over-Optimization Trap
The most seductive Solve mistake is trying to find the perfect opening. Candidates freeze on the first Ecosystem placement, hunting for the theoretically optimal species, and burn a third of their window before making a single move.
Solve does not reward a perfect first move. It rewards a complete, valid, well-reasoned solution arrived at through a sound process. A good-enough opening that you can build on is worth far more than a perfect opening you never get to because you ran out of time deliberating.
Two rules protect you here:
- Satisfice early, optimize late. Make a defensible move that respects the constraints, then refine if time allows. Progress creates information; paralysis creates nothing.
- Treat the first phase of Red Rock as scaffolding, not the exam. Early data-gathering exists to set up later decisions. Do not agonize over early steps as if each were the final grade — they inform what comes next, and the algorithm cares about the arc, not one click.
Recovering From a Bad Start
Everyone makes an early wrong move. What separates passing candidates is what they do in the sixty seconds after realizing it.
The instinct is to spiral: one bad placement triggers faster, less careful clicking to “make up time,” which produces exactly the scattered decision trail the algorithm penalizes. A bad start does not fail you. A bad start plus panic fails you.
When you notice you have gone wrong:
- Stop for one breath. Literally pause. A single deliberate breath costs you two seconds and prevents a two-minute cascade.
- Diagnose, do not thrash. Identify which constraint you violated or which assumption was wrong. Fix that one thing rather than restarting everything.
- Return to deliberate pace. Resume making considered moves. The algorithm reads recovery — a candidate who stabilizes after an error demonstrates exactly the judgment McKinsey is screening for.
A clean recovery can score better than a flawless run, because it shows how you behave when things go wrong, which is the more informative signal.
Managing Anxiety Under a Visible Countdown
The on-screen timer is a psychological instrument as much as a functional one. It exists partly to see who cracks. A few techniques keep the countdown from hijacking your thinking:
- Glance, don’t stare. Check the clock at natural decision points, not continuously. A timer watched every three seconds steals the attention you need for the actual problem.
- Convert time into phases, not seconds. Instead of “I have 4 minutes 12 seconds,” think “I am still in my execution phase, with buffer to verify.” Phase-based awareness is calmer and more actionable than a ticking number.
- Pre-commit to a pace before you start. Decide in advance that the first stretch is for reading and planning. When you have already granted yourself planning time, the early countdown stops feeling like time you are “losing.”
- Slow down exactly where it is hard. Counterintuitively, deliberately slowing on the genuinely difficult step is what the algorithm rewards — it mirrors how strong problem-solvers actually allocate attention.
A Pacing Framework for Any Solve Game
You will not know the exact minute budget in advance, so pace by phase and by internal checkpoints rather than by absolute time.
| Phase | What you should be doing | Red flag |
|---|---|---|
| Opening third | Reading rules, identifying binding constraints, forming a plan | Already placing pieces before you understand the rules |
| Around halfway | Executing deliberate moves, verifying each against constraints | Still hunting for the “perfect” first move |
| Final stretch | Reviewing for validity, confirming no rule is broken | Frantic clicking to fill remaining slots |
| Any point | Steady, purposeful decisions | Speed spikes after a mistake |
The one-breath rule: any time you feel your clicking accelerate out of anxiety rather than confidence, take a single breath and return to deliberate pace. Speed born of panic is the behavior Solve is specifically built to catch.
Where This Fits in Your Preparation
Solve screens for the same composure that carries the live McKinsey case interview: reasoning clearly while a clock and an interviewer both apply pressure. The candidate who can stay deliberate under Solve’s countdown is building the exact nervous-system control that later separates strong case performers from those who unravel mid-interview.
For the full picture of both assessments and how to prepare for each, start with our McKinsey Solve & PST preparation hub. If your office still runs the legacy paper test, our McKinsey Problem Solving Test strategy guide covers question-type tactics, and the most common Solve and PST mistakes shows you the traps to avoid before test day.
Build Composure Under a Clock
The only reliable way to stay calm under a countdown is to have felt one before. Timed CasesCoach AI mock sessions put you under a real clock so the pressure of decision-making against a timer becomes familiar instead of destabilizing — the same self-regulation Solve is scoring. Our free tier includes 3 practice cases plus AI Mock; the Pro plan opens the full 835+ case library and multiple AI mock sessions. See pricing for details.
Key Takeaways
- Solve is two timed mini-games (usually Ecosystem Building and Red Rock Study) over roughly 60–81 minutes; you cannot move time between them, so budget each game internally
- The algorithm scores your decision process — sequence, click patterns, and time allocation — so deliberate beats fast, and random panic clicking is actively penalized
- Split every game into orient-and-plan, execute, and verify phases; read all the rules before you touch anything
- Do not over-optimize your opening — a valid, well-reasoned solution beats a perfect first move you never finish
- A bad start does not fail you; a bad start plus panic does. Take one breath, fix the one thing that went wrong, and return to a deliberate pace
- Pace by phase, not by staring at the timer; slow down deliberately where the problem is genuinely hard