McKinsey Solve Red Rock Study: Data Analysis Guide

Since around 2020, McKinsey has replaced the paper Problem Solving Test (PST) with the digital Solve assessment, now the global default (a few offices may still run the legacy PST). A Solve session is two mini-games over roughly 60–81 minutes, and the two you are most likely to see are Ecosystem Building and the Red Rock Study. Occasionally candidates draw variants such as Sea Wolf or Plant Defense.

The Red Rock Study is the data game. It presents an investigation scenario through tables, charts, and short case documents spread across several phases, then asks you to draw the right numbers out, reason from them, and answer accurately — all while a clock runs and the platform records how you move. Scoring is algorithmic and process-based: the system evaluates your decision path and timing telemetry, not just your final answers, and roughly 70% of candidates are eliminated at this stage. This guide breaks down the exhibit-reading and fast-calculation skills that let strong candidates finish each data step with time to spare.

What the Red Rock Study Actually Tests

The Red Rock Study feels like a research assignment. You work through phases of a scenario, and each phase surfaces its own evidence: bar charts, line graphs, tables, and blocks of case text mixed together. Your job is to pull specific data points, perform calculations, and combine information across sources to reach a defensible conclusion — then do it again in the next phase before the timer forces you on.

In our experience coaching candidates through Solve preparation, the data work rewards three distinct sub-skills:

Sub-Skill	What It Looks Like	Where It Bites
Data extraction	Pulling one exact figure from a dense table or chart	Reading the wrong row, cell, or segment
Calculation	Percentage change, ratios, growth, breakeven	Slow arithmetic that eats the clock
Inference	Choosing the conclusion the data actually supports	Over-reading a distractor exhibit

Each sub-skill needs a different reading strategy. Applying a generic “study the exhibit carefully” approach burns seconds on data you do not need — and because timing telemetry is scored, wasted seconds cost you twice.

The 3-Pass Exhibit Reading Method

Most candidates read an exhibit linearly — title, axes, every data point, then the question. That costs 45–60 seconds per exhibit. Because the Red Rock Study buries the numbers you need inside data you don’t, a question-first method cuts exhibit reading to 15–25 seconds.

flowchart TD
    A[Read the Question First] --> B[Identify What Data You Need]
    B --> C{Single Source or Multiple?}
    C -->|Single| D[Locate Specific Data Points]
    C -->|Multiple| E[Map Which Source Has What]
    D --> F[Extract & Calculate]
    E --> F
    F --> G[Verify Against Answer Choices]
    G --> H{Answer Matches?}
    H -->|Yes| I[Lock In and Move On]
    H -->|No| J[Re-read Question for Missed Detail]

Pass 1: Read the Question (5 seconds)

Before touching the exhibit, read the prompt and identify exactly what it wants. A specific year’s figure? A percentage change between two periods? A comparison across segments or across phases of the study? Fix the key parameters in your mind first.

Pass 2: Targeted Scan (10–15 seconds)

Go straight to the data you need. For bar charts, jump to the specific bars. For tables, scan row and column headers first, then locate the intersection. Ignore the surrounding numbers — in Red Rock, a lot of the on-screen data exists precisely to pull you off course.

Pass 3: Calculate and Verify (15–30 seconds)

Run the calculation, then check it against the available answers. If your result matches no option cleanly, you probably misread a data point. Return to Pass 2 and re-locate the single figure — do not re-read the whole exhibit.

Chart-Specific Reading Strategies

Each exhibit type has predictable traps. Knowing them in advance saves 5–10 seconds per question by preventing re-reads.

Bar Charts

Red Rock uses stacked and grouped bar charts to test whether you can isolate one segment correctly rather than grabbing the total.

Trap	Example	Fix
Reading the total instead of a segment	“Population of species A” when bars are stacked	Read the segment boundaries, not the bar top
Confusing axis scales	Left axis = count, right axis = rate	Check the axis label before reading any value
Misreading grouped bars	Comparing Phase 1 of group A to Phase 2 of group B	Verify the legend color mapping first

Tables

Tables are the workhorse exhibit. They test how fast you can find an intersection and calculate between cells.

Speed technique: Trace with your cursor or a finger — row horizontally, column vertically, to the intersection. This physical tracking prevents the single most common error, reading from the wrong row.

Line Graphs

Line graphs usually test trend identification and percentage change. The main trap is confusing absolute change with percentage change.

• Absolute change: end value minus start value
• Percentage change: (end − start) / start × 100

A line rising from 50 to 75 is a 50% increase, not 25%. In our coaching, this one error accounts for roughly 15% of the wrong answers we see on line-graph questions.

Calculation Shortcuts for Red Rock Speed

The Red Rock clock is shared across the whole study, so a calculation that drags past 30 seconds is quietly stealing time from later phases you have not seen yet. These shortcuts keep most data calculations under 20 seconds.

Percentage Calculations

The anchor method: instead of computing an exact percentage, stack known benchmarks.

• 10% = move the decimal one place left
• 5% = half of 10%
• 1% = move the decimal two places left
• 25% = divide by 4
• 33% = divide by 3
• 75% = subtract 25% from the whole

Example: What is 17% of 4,200?

• 10% = 420
• 5% = 210
• 2% = 84
• 17% = 420 + 210 + 84 = 714

Ratios and Breakeven

Red Rock questions often ask which option “pays back” or which allocation is most efficient — ratio and breakeven work.

• Ratio comparison: reduce both sides to a common base before comparing (e.g., cost per unit, output per resource) rather than eyeballing raw totals.
• Breakeven: divide the fixed commitment by the per-unit contribution. If a plan costs 1,200 and each unit returns a net 40, breakeven is 1,200 ÷ 40 = 30 units. Anything the exhibit shows above 30 clears the bar.

Growth Rate Calculations

The Rule of 72: to estimate doubling time, divide 72 by the growth rate. At 8% growth per period, a quantity doubles in about 9 periods (72 ÷ 8 = 9).

Compound vs. simple: for spans under 3 periods, simple approximation is close enough. Over longer spans, compounding pulls the result noticeably higher — do not approximate a long horizon linearly.

Division Shortcuts

Many questions require dividing large numbers to find a share or a per-unit value.

Division	Shortcut
÷ 5	× 2, then ÷ 10
÷ 25	× 4, then ÷ 100
÷ 8	÷ 2, ÷ 2, ÷ 2
÷ 15	÷ 3, then ÷ 5
÷ 50	× 2, then ÷ 100

Approximation Strategy

When the answer options are spread 10%+ apart, rough approximation beats exact calculation. Scan the options first — if they read 12%, 28%, 43%, 61%, you only need to land within 5% to pick the right one.

Multi-Source Synthesis

The hardest Red Rock questions make you combine a table, a chart, and a line of case text into one answer. These carry the same weight as simpler questions but take 2–3x longer, so they are where candidates lose the clock.

flowchart LR
    A[Data Table] --> D[Combine]
    B[Chart / Trend] --> D
    C[Case Document Detail] --> D
    D --> E[Calculate Answer]

Strategy for multi-source questions:

1. Identify which sources you actually need — the prompt usually points to them.
2. Extract one number (or fact) from each source before attempting any calculation.
3. Note the extracted values on your scratch paper — do not hold them in memory.
4. Calculate from the written values.

The error to avoid: flipping between sources mid-calculation. Each flip costs 5–8 seconds of re-orientation and increases misread risk — and in a telemetry-scored game, that thrashing is visible.

Watch for Distractor Data

Red Rock deliberately shows more data than any single question needs. A phase might display five metrics when the question turns on one. Before calculating, ask: which of these figures does the prompt reference? If a number is not tied to the question, it is scenery. Confident candidates ignore most of what is on screen.

Practice Drill: Build Your Data Speed

Data speed comes from deliberate, timed practice, not from reading more theory. Candidates who complete 50+ timed exhibit readings before test day consistently outperform those who only run full simulations.

Drill Structure (15 minutes daily, 2 weeks)

Week	Focus	Drill
Week 1, Days 1–3	Bar charts	Read 10 exhibits, 15 seconds each. Extract one value per exhibit
Week 1, Days 4–5	Tables	Read 10 tables, 20 seconds each. Find specific intersections
Week 1, Days 6–7	Line graphs	Read 10 graphs, calculate % change in 25 seconds
Week 2, Days 1–3	Ratios & breakeven	10 prompts, compute per-unit / breakeven in 30 seconds
Week 2, Days 4–5	Multi-source	Table + chart + text detail, synthesize in 90 seconds
Week 2, Days 6–7	Full simulation	8 mixed questions, timed, at Red Rock pace

Where to Find Practice Exhibits

McKinsey Global Institute reports and McKinsey Quarterly articles use the same chart styles you will meet in Solve. Pull a chart, write your own question against it, and time your reading. You can also practice with exhibits from our McKinsey case library — the case data tables mirror Red Rock exhibit formats closely.

Common Data Analysis Errors

Across the debriefs we review, five errors account for the large majority of Red Rock data mistakes:

1. Unit confusion: reading millions as thousands, or vice versa. Always check the axis title and any footnote for the unit.
2. Wrong time period or phase: the prompt asks about one window and you compute another. Re-read the boundary the question names before you calculate.
3. Percentage vs. percentage point: “rose from 20% to 25%” is a 5 percentage-point increase but a 25% relative increase — and Solve will offer both as options.
4. Ignoring footnotes and case-text caveats: an exhibit or document may exclude certain data or redefine a category. A 3-second check prevents the misread.
5. Rounding too early: on multi-step calculations, round only at the final step. Early rounding compounds into 5–10% errors.

Key Takeaways

• The Red Rock Study is the data-analysis mini-game in McKinsey Solve — one of the two core games alongside Ecosystem Building.
• Read the question before the exhibit; most wasted time comes from reading data the prompt never needed.
• Use the 3-Pass method: question first, targeted scan, calculate and verify.
• Learn the chart traps: stacked-bar boundaries, dual axes, grouped-bar color mapping.
• Drill calculation shortcuts daily — the anchor method, ratio/breakeven, and the division table cover most Red Rock math.
• Treat surplus figures as distractors; confident candidates ignore most of what is on screen.
• Scoring is process- and timing-based, so speed and a clean decision path are scored, not just correctness.

What Comes Next

The Red Rock Study is only half of the assessment. Pair this guide with the Ecosystem Building and two-module strategy overview so you prepare for both games, and use our Solve time management strategies to build a complete speed toolkit. For the full picture — format, scoring, and a study plan — start from our McKinsey Solve preparation hub.

Red Rock rewards fast, accurate arithmetic under pressure, and that is exactly what deliberate case-math practice builds. CasesCoach AI practice includes case-math drills with instant feedback, so you can see where your speed breaks down and fix it before test day. Free access includes 3 cases plus AI Mock Interview; Pro unlocks 835+ cases and multiple AI mock sessions — see pricing or jump straight into an AI Mock Interview.