Reading Charts and Exhibits in Case Interviews

Your client's market share has been declining. The interviewer hands you a bar chart. You have 60 seconds to find the insight, explain what it means, and connect it to a recommendation.

This is the data interpretation moment that over 70% of consulting case interviews include — and it's where many candidates lose significant ground. Not because they can't read charts in general, but because they haven't practiced the specific skill of reading an exhibit quickly, accurately, and at the level of insight the interviewer expects. Describing a chart earns you nothing. Interpreting it — leading with what the data means for the business — is what separates candidates who advance from those who don't.

Why Exhibit Reading Is Its Own Skill

Reading a chart under pressure in a high-stakes interview is fundamentally different from reading a chart at your desk with unlimited time. Three things change:

Stakes create cognitive load. You're simultaneously managing anxiety, thinking about the case, and formulating a response. This makes it easy to read too quickly, miss details, and anchor on the first pattern you see rather than the most important one.

Interviewers evaluate your process, not just your conclusion. A partner watching you interpret an exhibit is assessing: Did you read the axes first? Did you notice the footnote? Did you identify the key exception, or just the obvious trend? The interpretation you deliver out loud is as important as whether you got it right. MBB interviewers have noted that outstanding candidates deliver "Level 2 insights" — the non-obvious interpretation that goes beyond the primary trend — while average candidates stop at restating what the chart shows.

Different firms weight exhibit reading differently. BCG cases are among the most exhibit-heavy in consulting interviews — a single BCG case may include 4-6 charts across the interview. If your exhibit reading is slow or error-prone, you fall behind the case narrative and lose time you can't recover. McKinsey's candidate-led format uses exhibits when you request specific data, which means you need to interpret them in the context of your own analytical plan. Bain typically provides 2-3 exhibits per case. The BCG Case Interview Guide covers exhibit reading in the context of BCG's full interviewer-led format.

The 4-Step Exhibit Reading Method

This method works for any chart type. Practice it until it's automatic.

Step 1: Read the title, axes, and units (10 seconds)

Before looking at the data, read the chart title and the axis labels including units. This sounds obvious, but candidates routinely skip it when nervous.

Why it matters: The title tells you what the chart is measuring. The units tell you whether you're looking at absolute numbers, percentages, indices, or growth rates. Confusing an index with an absolute value — a common exhibit design choice in case interviews — will produce a completely wrong interpretation.

What to check:

• Titles that include qualifiers ("Top 5 Markets Only," "Excludes Asia-Pacific")
• Non-zero baselines that visually exaggerate small differences
• Logarithmic scales that compress large ranges
• Secondary y-axes on dual-axis charts
• Units: millions vs. billions, percentage vs. percentage points, index vs. absolute

Step 2: Identify the trend or pattern (15 seconds)

Look at the data as a whole. What is the main story? Is this a chart that shows a declining trend? A comparison across categories? A distribution? A correlation?

Don't try to read specific values yet. Get the shape of the story first.

Common patterns and what they suggest:

• Steady decline: Something structural is changing — look for inflection points that indicate when the shift started
• Accelerating decline: The rate of change is worsening, which is a more urgent problem than a linear decline and suggests a compounding factor
• Category comparison: You're identifying which category is highest, lowest, growing fastest, or declining fastest
• Divergence: Two lines or bars that started similar and are now far apart — the divergence point is often the key event to explain
• Convergence: Previously separated metrics moving toward each other, which can signal commoditization or competitive catch-up

Step 3: Find the exception or outlier (10 seconds)

The main trend is usually not the most interesting thing about a chart in a case interview. Case interviewers design exhibits with a key exception or outlier embedded in them — one data point or one category that breaks the pattern.

After identifying the trend, actively look for what doesn't fit. A market that's growing when all others are declining. A cost category that's flat when all others are rising. A time period where the trend reverses. A single product line that deviates from the pack.

The exception is usually where the insight lives. In BCG's interviewer-led format especially, the key finding is often in the anomaly, not the overall pattern.

Step 4: Connect to the case and state an implication (15-20 seconds)

Now you can speak. Lead with your implication — the "so what" — not with a description of the chart.

Weak: "This chart shows that revenue has been declining since 2022, dropping from $450 million to $380 million over three years."

Strong: "The key insight here is that revenue decline accelerated sharply in 2024 — the rate of decline doubled compared to 2022-2023. That suggests a structural shift happened in 2024, not a gradual deterioration. I'd want to understand what changed that year — a new competitor, a pricing change, or a product issue."

The strong version still contains the facts, but it leads with interpretation and ends with the next analytical question. This is what a consultant does with data — they don't recite it, they advance the thinking.

Seven Common Exhibit Types and Their Traps

Most case interview guides cover bar charts and line graphs. Here are the seven chart types you'll actually encounter, including the less common ones that trip candidates up the most.

Bar Charts (Simple, Grouped, and Stacked)

What they show: Comparisons across categories or a single metric over time.

Common traps:

• Non-zero baseline. A bar chart with a y-axis starting at 80 makes a 2-point difference look dramatic. Always check whether the baseline is zero.
• Percentage vs. absolute. A chart showing 20% decline in a small market versus 5% decline in a large market has very different strategic implications. Know which metric is plotted.
• Stacked vs. grouped confusion. Stacked bar charts compare totals and composition simultaneously. Candidates frequently misread individual segments of a stacked bar because they compare segment heights rather than segment sizes (which requires subtracting the bottom boundary from the top).

How to read a stacked bar correctly: Look at total bar height first for the overall comparison, then examine individual segments by estimating their width (top boundary minus bottom boundary), not their absolute position.

Line Graphs

What they show: Trends over time, often with multiple series on the same chart.

Common traps:

• Different scales on multiple series. Two lines that look similar in slope may have very different absolute rates of change if they're on different scales.
• Lagging vs. leading indicators. If the chart shows two metrics, one may lead the other by several periods. The relationship between them — not their individual trends — is the insight.
• Smoothed vs. raw data. Some exhibits use moving averages that can hide volatility or recent reversals.

Scatter Plots

What they show: Relationships between two variables. Often used to show correlation — or the absence of it.

Common traps:

• Correlation vs. causation. A scatter plot can only show correlation. Never say the x-variable causes the y-variable based solely on a scatter plot.
• Outliers that shift the story. One or two outlier points can anchor the visual impression differently than the main cluster suggests. Identify outliers explicitly: "Excluding the two outliers in the upper right, the relationship between X and Y is actually negative."
• Logarithmic axes. Scatter plots frequently use log scales on one or both axes. The visual impression of the relationship changes dramatically depending on the scale.

How to extract the insight: Describe the general trend (positive, negative, or no correlation), call out any outlier quadrants, and connect to what this means for the client. A scatter plot showing no correlation between marketing spend and revenue by region tells you marketing allocation is inefficient — some regions are overspending with no return.

Waterfall Charts (Bridge Charts)

What they show: How a starting value becomes an ending value through a series of additions and subtractions. Extremely common in profitability analysis — a typical "EBITDA bridge" shows how EBITDA changed from one year to the next.

Common traps:

• Ignoring the magnitude hierarchy. In a waterfall showing profit bridge, the largest negative contributor is usually the key finding. Candidates who discuss the items left-to-right rather than by magnitude miss the main story.
• "Other" or "unspecified" categories. A large "other" bucket is often where interesting costs are buried. If "other costs" is the second-largest negative contributor, ask what's in it.
• Confusing absolute and relative changes. A waterfall can show absolute dollar changes or percentage point changes. Misreading which one changes your interpretation entirely.

How to read a waterfall: Start with the starting value and ending value. Calculate the total change. Then identify the 1-2 largest contributors to that change (positive or negative). These are your headline findings.

Pie Charts and Donut Charts

What they show: Proportional breakdown of a whole into its parts.

Common traps:

• Poor at showing changes over time. If you're comparing two pie charts (e.g., revenue mix in 2022 vs. 2025), compare specific segments rather than trying to compare the visual shapes. The most useful insight is usually which segment grew or shrank the most.
• Misleading with many small segments. Pie charts with 8+ segments become hard to read. Focus on the top 2-3 segments and the "other" bucket.

Dual-Axis Charts

What they show: Two metrics with different scales plotted on the same chart.

Common traps: This is the single most dangerous chart type in case interviews. Candidates regularly read both lines as if they're on the same scale and state relationships that don't exist. Always identify which variable uses the left axis and which uses the right axis before making any comparison.

How to avoid the trap: When you see a dual-axis chart, explicitly say which axis you're referencing: "Revenue, on the left axis, has declined 15%, while customer count on the right axis has remained stable. This tells me revenue per customer is falling." Stating the axes out loud prevents the most common dual-axis error and demonstrates precision to the interviewer.

Bubble Charts

What they show: Three variables simultaneously — x-axis, y-axis, and bubble size. Often used for portfolio analysis or market attractiveness comparisons.

Common traps:

• Ignoring the bubble size dimension. Candidates focus on the x-y position and forget that bubble size represents a third variable (often revenue, market size, or number of customers).
• Misreading the legend. Bubble size scales are not always linear. Check whether the legend shows the size for specific values.

How to read a bubble chart: Identify what each axis and the bubble size represent. Then look for clusters and outliers. A bubble in the upper-right quadrant (high on both axes) with a large size is usually the most attractive opportunity. A large bubble in the lower-left quadrant is the problem to address.

Worked Example: Reading a BCG-Style Exhibit

The exhibit: A grouped bar chart titled "Operating Margin by Product Line (2022-2025)." Four product lines (A, B, C, D) are shown across four years. Product lines A, B, and C show steady operating margins of 22-24%. Product line D shows margins declining from 18% in 2022 to 9% in 2025.

Step 1: Title, axes, and units. This shows operating margin (a percentage, not absolute dollars) by product line. Four years of data. The y-axis starts at 0%, which is correct — no baseline distortion.

Step 2: Main pattern. Three product lines are stable and healthy at 22-24%. One product line (D) is declining sharply.

Step 3: Exception. Product D is the clear outlier — not just lower margin (which was also true in 2022) but a margin that has been compressing rapidly. From 18% to 9% is a 9-point compression in three years, nearly cutting margin in half. The acceleration in 2024-2025 (a 4-point drop in the final year vs. 2-3 points in earlier years) is particularly notable.

Step 4: Implication. "The key insight is that this is not a company-wide margin problem — it's almost entirely isolated to Product D, whose margins have compressed from 18% to 9% over three years and are accelerating downward. The other three product lines are stable at 22-24%. This tells me I should focus my analysis specifically on Product D: what's happening on pricing or costs in that line that isn't affecting the others? The acceleration in the final year suggests something changed recently — a new competitor, a cost input issue, or a pricing decision."

This interpretation takes about 20 seconds and immediately focuses the case on the right problem. For more context on how to connect this kind of insight to a final recommendation, see Case Interview Synthesis.

Worked Example: Reading a Waterfall Chart

The exhibit: A waterfall chart titled "EBITDA Bridge: FY2023 to FY2025 ($M)." Starting EBITDA is $120M. The bridge shows: Volume effect +$15M, Price effect -$28M, Raw materials -$12M, Labor costs -$8M, SG&A -$5M, Other +$3M. Ending EBITDA is $85M.

Step 1: Title, axes, units. This is an absolute dollar waterfall showing EBITDA change over two years. Total decline: $35M.

Step 2: Main pattern. EBITDA declined $35M. There's one positive contributor (volume growth) that is more than offset by several negative contributors.

Step 3: Exception. The price effect at -$28M is by far the largest single contributor to the decline — it accounts for 80% of the total EBITDA erosion. Volume actually grew, which means the company is selling more units at lower prices. Raw materials and labor are secondary factors.

Step 4: Implication. "The dominant driver of the $35M EBITDA decline is pricing erosion — a $28M negative price effect that overwhelms the $15M positive volume contribution. This is not a demand problem or a cost explosion. The company is growing volume but giving up margin on each unit. I'd want to understand what's driving the price compression — competitive pressure, a deliberate volume-for-price trade, or a product mix shift toward lower-priced SKUs."

Footnotes and Fine Print: The Hidden Insight Source

One of the most frequently neglected parts of any exhibit is the footnotes and data notes. In case interviews, these are often where the most important methodological clarifications live:

• "Revenue excludes inter-company transfers"
• "Market data based on company estimates"
• "2025 data is preliminary / annualized from H1"
• "Includes acquisition of [company] as of Q3 2024"
• "Constant currency basis"

That acquisition footnote changes everything about how you interpret a revenue increase — it might be entirely acquisition-driven, not organic growth. A "constant currency" note means foreign exchange effects have been removed. A "preliminary" note means the most recent data point may shift. Always read the footnotes before interpreting the data.

Footnotes are also a frequent source of exam-style traps in the McKinsey Solve assessment, where the Redrock Study module tests data interpretation with similar chart types and deliberate footnote-level complexity.

Common Mistakes and How to Avoid Them

Based on how MBB interviewers describe the most frequent candidate errors:

Mistake 1: Describing instead of interpreting. "Revenue went from $400M to $350M" is a description. "Revenue declined 12.5%, with the decline accelerating in the final year, suggesting a structural shift rather than a temporary dip" is an interpretation. Always ask yourself: am I telling the interviewer what they can see, or what they should think?

Mistake 2: Anchoring on the obvious trend. The first thing you see in a chart is usually the overall trend. But the overall trend is what every candidate sees. The differentiator is the secondary insight — the exception, the inflection point, the segment that behaves differently.

Mistake 3: Reading too many numbers. You don't need to cite every data point. The exhibit may have 30 numbers in it. Your job is to identify the 2-3 that matter most and explain why they matter.

Mistake 4: Forgetting to connect back to the case. An exhibit doesn't exist in isolation. After interpreting it, connect your finding to the case question. "This pricing erosion aligns with the competitive entry we discussed earlier and suggests the client may be in a price war they're losing."

Mistake 5: Not asking what's missing. Sometimes the most important insight from an exhibit is what data it doesn't show. If a profitability exhibit shows revenue but not costs, and the question is about margin decline, ask for the cost side. This demonstrates you're thinking about the full picture, not just what's in front of you.

Practice: Building Exhibit Reading Speed

The best way to improve exhibit reading is deliberate timed practice. Start with 60-second drills: read an exhibit, state your key insight, and connect it to a hypothetical case question. Our AI-powered cases on the dashboard include exhibit-reading exercises across 6 chart types with the same time pressure and question format you'll face in live case interviews. The AI debrief evaluates whether you led with implication, caught key exceptions, and connected the data to the case.

For broader practice on the quantitative skills that support exhibit reading — percentage calculations, growth rates, margins — see Case Interview Math Practice and Mental Math for Case Interviews.

Key Takeaways

• Over 70% of consulting cases include data exhibits — this skill is not optional.
• Use the 4-step method: read title/axes/units, identify the main pattern, find the exception, then state an implication.
• Lead with the "so what," not a description of the chart. Start with "The key insight is..."
• The exception or outlier in a chart is usually more valuable than the trend.
• Know all seven common chart types: bar, line, scatter, waterfall, pie, dual-axis, and bubble. Each has specific traps.
• Dual-axis charts are the most dangerous — always identify which axis corresponds to which variable before commenting.
• Read footnotes before interpreting any exhibit. Acquisition notes, currency adjustments, and data exclusions can change the entire interpretation.
• Practice timed exhibit reading to build the habit of fast, accurate, insight-led interpretation.