McKinsey Solve Preparation: A 4-Week Plan to Pass the Assessment

McKinsey’s Solve assessment eliminates roughly 70% of candidates before they ever reach a case interview. Since around 2020, this gamified digital test has replaced the paper-based Problem Solving Test (PST) as McKinsey’s global default — if you’re applying in 2026, Solve is almost certainly what stands between you and the interview rounds.

Unlike the old PST, you cannot cram for Solve by memorizing a question bank, because there isn’t one. Solve scores your decision-making process — the sequence of choices you make, how you allocate your time, whether you adjust when conditions change — not simply whether you land on a correct final answer. That means the highest-leverage preparation is building genuine, repeatable analytical habits rather than drilling practice questions.

This guide gives you exactly that: a 4-week plan organized around the skills Solve actually measures — data analysis, constrained optimization and systems thinking, and mental math — plus the diagnostic, simulation, and test-day routines that turn those skills into a passing performance.

What Solve Actually Tests

A Solve session assigns you two mini-games back to back, running roughly 60–81 minutes in total. The overwhelming majority of candidates get Ecosystem Building paired with the Red Rock Study; a minority encounter other games such as Sea Wolf or Plant Defense.

• Ecosystem Building drops you into an environment with a list of candidate species and a set of rules (altitude, terrain, temperature, and similar constraints). Your task is to select a set of species that forms a viable food chain under those rules. It is fundamentally a constrained-optimization exercise, and the algorithm watches how systematically you explore the constraint space, not just whether you eventually land on a working chain.
• Red Rock Study hands you a packet of ecological data — charts, tables, and short text — and asks you to extract, combine, and apply it to answer a sequence of questions. It is a data-analysis and investigation task that rewards structured note-taking and precise reading of scales and units.

Scoring is algorithmic and behavioral: Solve logs your decision sequences, click patterns, and time allocation, and it can distinguish a candidate who reached an answer through systematic exploration from one who arrived there by trial and error. Because scoring is based on process, undirected clicking and restarting from scratch are actively penalized — there is no equivalent of “guess and move on” the way there was on the old PST.

For the full breakdown of both games, the scoring model, and a historical note on the legacy PST, see our McKinsey Solve assessment hub guide. This article focuses on the week-by-week plan for getting ready.

Before Week 1: Run a Skills Diagnostic

Because there is no official Solve practice test with a published score, your diagnostic is a self-assessment rather than a graded exam. Spend a single session — about 45 minutes — running three short exercises and paying attention to how you work, not just whether you get the “right” answer:

1. A mental math speed check. Work through ten mixed percentage, division, and growth-rate problems without a calculator and time yourself. If most take longer than 20–25 seconds, calculation speed is a priority.
2. A small constrained-selection exercise. Take any small set of items with a few compatibility rules (for example, seating people around a table with “cannot sit next to” constraints, or grouping objects that must share at least one property) and solve it. Notice whether you mapped the constraints before touching anything, or started guessing and adjusting reactively.
3. A data-reading exercise. Pull a chart-heavy exhibit from a real business case and give yourself 90 seconds to answer three questions about it (a specific value, a trend direction, a comparison). Notice how much of the 90 seconds went to finding the right part of the exhibit versus reading everything.

Diagnostic Signal	What It Reveals	Where to Focus
Calculations take 20+ seconds or need scratch work	Mental math is a bottleneck	Week 2 mental math track
You started selecting/clicking before mapping constraints	Approach is reactive, not systematic	Week 2 constrained-optimization track
You reread the same chart section more than once	Data extraction lacks structure	Week 2 data-analysis track
You solved the puzzle but couldn’t explain your method afterward	Process isn’t yet repeatable	Week 3 process-review emphasis
Everything above felt comfortable and fast	You are close to test-ready	Compress to a 2-week version of this plan

Whatever the diagnostic reveals, the goal is the same: identify one or two skills to weight more heavily in Week 2, without skipping the others entirely.

The 4-Week Solve Preparation Roadmap

flowchart LR
    A["Week 1\nFormat & Diagnostic"] --> B["Week 2\nCore Skill Building"]
    B --> C["Week 3\nFull Simulations"]
    C --> D["Week 4\nRefinement & Rest"]
    style A fill:#e8f4fd,stroke:#2196F3
    style B fill:#e8f4fd,stroke:#2196F3
    style C fill:#fff3e0,stroke:#FF9800
    style D fill:#e8f5e9,stroke:#4CAF50

Week 1: Format Mastery and Diagnostic

Objective: Understand the shape of a Solve session, run your diagnostic, and identify the one or two skills that need the most work.

Daily schedule (60 minutes):

• 15 min: Read through how Ecosystem Building and Red Rock actually work — mechanics, what is scored, common structural traps (covered in full in the Solve hub guide)
• 20 min: Mental math warm-up (percentages, growth rates, quick division)
• 20 min: Practice one constrained-selection puzzle or one data-heavy exhibit, alternating by day
• 5 min: Log what felt slow or uncertain in a short study journal

Week 1 milestones:

• Complete the skills diagnostic (day 1 or 2)
• Be able to describe, in your own words, what each mini-game is actually measuring
• Identify your one or two weakest skills from the diagnostic table above
• Establish a mental math baseline you can compare against in Week 3

Why this week matters: Candidates who skip straight to “practicing the games” often spend their first attempts figuring out the interface and the rules instead of demonstrating a method. Understanding the constraint structure and scoring logic in advance means every subsequent practice session is spent building skill, not orientation.

Week 2: Core Skill Building

Objective: Drill the three skills Solve actually measures until each becomes closer to automatic. Weight your weakest track from the diagnostic more heavily, but touch all three every day.

Split each 90-minute session across three tracks:

Track 1 — Data analysis (for Red Rock), 30 min: Practice extracting, organizing, and cross-referencing data from chart-heavy exhibits. Take structured notes as you read (data source, key numbers, units, relationships), then answer a handful of questions against your notes rather than the raw exhibit. Real case exhibits work well for this — our Red Rock data-interpretation guide walks through the specific reading habits that transfer directly.

Track 2 — Constrained optimization and systems thinking (for Ecosystem Building), 30 min: Practice problems where you must select or arrange a set of items under multiple simultaneous rules — food-chain and habitat puzzles, logic-grid puzzles, or resource-allocation problems. The skill you are building is not puzzle-specific; it is mapping every constraint before you commit to a choice, working systematically from one end of the problem (for example, from the most-constrained item), and backtracking one step at a time instead of restarting when a choice fails.

Track 3 — Mental math, 30 min: Percentages, division shortcuts, and compound-growth estimation, done without a calculator. Our mental math guide sequences these drills with progressive difficulty so speed builds week over week.

Week 2 milestones:

• Map constraints before acting on at least 8 of 10 practice puzzles, rather than starting with trial and error
• Extract a specific data point from a standard exhibit in under 30 seconds
• Complete basic percentage and division problems without scratch work
• Write down 8–10 “rules” from your own recurring mistakes across all three tracks (a habit you’ll keep refining through Week 4)

A habit log, not a rule book: Unlike PST prep, where a “rule” often meant a shortcut for a specific question type, your Week 2 rules should describe behaviors — “map every constraint before selecting anything,” “write down units before comparing two numbers,” “if a choice fails, swap one item before rebuilding.” These transfer across games and are exactly what the algorithm is designed to detect.

Week 3: Full Timed Simulations — Review Process, Not Just Answers

Objective: Practice under conditions that resemble an actual Solve session, and train yourself to review how you solved a problem, not only whether you reached a workable outcome.

Daily schedule (75–90 minutes):

• 60–80 min: A full simulated session — one constrained-selection exercise and one data-analysis exercise, back to back, timed, with no pauses between them
• 20–30 min: Process review (see the log below), not a score check

Because Solve has no released practice test with a numeric score, “reviewing your results” means reconstructing your decision path, not comparing a percentage to a passing line. After each simulation, walk back through what you did using a simple log:

Decision Point	What I Did	Was It Systematic?	Time Spent	What I’d Do Differently
First move on the constraint puzzle	Picked an item that looked promising	No — hadn’t mapped constraints yet	3 min	Spend 2 min mapping constraints first
Stuck mid-puzzle	Restarted from scratch	No — wasted prior progress	6 min	Backtrack one step instead of rebuilding
Reading the data exhibit	Reread the same chart twice	Partially — notes weren’t structured	4 min	Take notes by data source, not by question

Week 3 milestones:

• Complete a full 60–81 minute simulated session without needing a break
• Identify at least one recurring process failure (not a knowledge gap) from your logs — for example, restarting instead of backtracking, or reading exhibits without taking notes
• Reduce the number of “not systematic” entries in your log across successive simulations
• Notice whether you are spending disproportionate time on the easy parts of a puzzle at the expense of the harder ones — a pacing signal worth correcting before Week 4

If timing consistently feels tight, our Solve and PST time-management guide covers pacing checkpoints you can build into a simulation.

Week 4: Refinement and Test-Day Execution

Objective: Tighten your process into a small number of dependable habits, confirm your technical setup, and arrive rested rather than freshly drilled.

Daily schedule (45–60 minutes):

• Day 1–2: One more full simulation, focused specifically on the process failure you identified in Week 3
• Day 3: Consolidate your habit log into 4–6 rules you can genuinely recall under pressure — more than that and you will not remember them mid-session
• Day 4: Light mental math and one short constrained-selection puzzle only — no full simulation
• Day 5: Confirm your technical setup (see checklist below) and do a light review of your consolidated rules
• Day 6–7: Rest. Fatigue measurably degrades the consistency Solve rewards, and no amount of last-minute drilling offsets a poor night’s sleep.

Technical and environment checklist:

• Wired internet connection if available — lag can distort behavioral telemetry
• Laptop or desktop with a large screen, not a phone or tablet
• Quiet, private room with no expected interruptions for the full session length
• Browser extensions disabled or an incognito/private window, in case any extension interferes with game elements
• Full charge or a reliable power source; do not risk a mid-session low-battery warning
• Scheduled at your personal peak-alertness time of day, not squeezed into a tired evening slot

The night before and morning of: Review your 4–6 consolidated rules once, lightly. Do not attempt a full simulation. Sleep 7–8 hours — cognitive processing speed and consistency both drop measurably on inadequate sleep, and consistency is one of the things Solve’s algorithm is specifically built to detect. In the morning, eat normally, arrive at your setup with time to spare, and do two or three quick mental math problems as a warm-up rather than a test.

From Solve Skills to Interview Performance

flowchart TD
    A[Solve Skills] --> B[Interview Application]

    A1[Data Analysis] --> B1[Exhibit Analysis in Cases]
    A2[Constrained Optimization] --> B2[Structuring Under Multiple Constraints]
    A3[Mental Math] --> B3[Real-Time Calculations for the Interviewer]
    A4[Systematic, Adaptive Process] --> B4[Hypothesis-Driven Case Structure]

    A1 --- A
    A2 --- A
    A3 --- A
    A4 --- A
    B1 --- B
    B2 --- B
    B3 --- B
    B4 --- B

Solve Skill	Interview Application	How It Transfers
Data analysis	Reading exhibits during a live case	Same extraction habits, now applied out loud in conversation
Constrained optimization	Structuring a problem with multiple business constraints	Same habit of mapping constraints before committing to a path
Mental math	Back-of-envelope calculations	Identical arithmetic, now verbalized for the interviewer
Systematic, adaptive process	Hypothesis-driven problem solving	Same discipline of testing, adjusting, and explaining your reasoning

Candidates who build these habits deliberately during Solve preparation tend to walk into their first case interview practice with a structure they already trust, rather than improvising one for the first time under pressure.

What Comes After Solve

Clearing Solve moves you into McKinsey’s interview rounds, which have two parts: the Personal Experience Interview (PEI), a set of structured behavioral questions about leadership, impact, and conflict; and the interviewer-led case, a live business problem where structured, hypothesis-driven thinking decides the outcome.

Shift your focus as soon as you advance. Our McKinsey case interview guide covers that next stage, and practicing with real McKinsey-style cases — particularly profitability, market sizing, and financial analysis cases — builds directly on the analytical habits this plan develops.

Common Mistakes That Undermine Solve Prep

Mistake	Why It Hurts	How to Fix
Starting without reading the rules	You miss constraints and look disorganized to the algorithm	Force a read-everything-first habit, even in practice
Random trial and error	Solve’s behavioral scoring detects non-systematic exploration	Map constraints before making any selection
Restarting from scratch when stuck	Wastes time and signals an inability to recover	Backtrack one step at a time instead
Reviewing only whether the puzzle “worked”	Misses the process failures the algorithm actually scores	Use a decision-path log after every simulation, not just a pass/fail check
Over-optimizing one dimension	The games track multiple criteria at once	Balance speed, accuracy, and adaptability across a session
Testing on an unstable connection or noisy room	Lag and interruptions fragment your behavioral signal	Confirm your setup during Week 4, not on test day
Cramming a full simulation the night before	Fatigue undermines the consistency being measured	Rest in the final 48 hours instead

Key Takeaways

• Solve has replaced the paper PST as McKinsey’s global default; a session is two mini-games running roughly 60–81 minutes
• There is no question bank to memorize — Solve scores your decision process, so preparation means building trainable habits
• Run a skills diagnostic before Week 1 to identify whether data analysis, constrained optimization, or mental math needs the most attention
• Week 2 drills all three skills in parallel, weighted toward your weakest track
• Week 3 simulations should be reviewed for process — decision path, backtracking, time allocation — not just whether you reached a workable outcome
• Week 4 consolidates your habits into a small, memorable rule set and prioritizes rest over last-minute drilling
• The habits this plan builds transfer directly into the PEI and interviewer-led case that follow Solve

Put These Skills Into Practice

The timed-simulation weeks in this plan pair naturally with structured case practice. CasesCoach’s AI Mock Interview lets you rehearse hypothesis-driven problem solving and real-time calculation under pressure — the same systematic, adaptive process Solve measures, and exactly what the PEI and interviewer-led case demand next. The free tier includes three full cases plus AI Mock; Pro unlocks 835+ real cases and multiple AI mock sessions so you can build the habit until it holds under pressure. For the full picture of both mini-games and how Solve is scored, start with our Solve assessment hub guide.