BUSINESS STATISTICS • DESCRIPTIVE ANALYTICS

Percentiles, Quartiles & Boxplots — Percentiles, Quartiles, and Boxplots for Business Data

Transform raw business data into actionable distributional insights using positional measures and visual summaries.

Historical Context & Motivation

Summary statistics such as the mean and standard deviation have long served as the backbone of data analysis, yet they tell an incomplete story when a dataset is skewed or riddled with outliers—precisely the conditions managers encounter daily when examining sales figures, employee compensation, or customer wait times. The need to describe where individual observations fall relative to the rest of the distribution motivated the development of positional measures. By converting raw values into ranks and then into standardized positions, analysts can answer practical questions such as 'How does this quarter's revenue compare to our historical performance?' or 'Which branch is in the bottom 10% for customer satisfaction?'

The statistical tools we explore in this lesson—percentiles, quartiles, and boxplots—evolved over more than a century of statistical thought, each building on the last to give decision-makers an increasingly powerful vocabulary for describing distributional position and spread.

1885
Galton's Percentile Concept
Sir Francis Galton introduced the term percentile while studying anthropometric data, formalizing the idea of ranking observations on a 0–100 scale to compare individuals within a population.
1900
Quartile Formalization
Karl Pearson and his school refined the concept of dividing ordered data into four equal parts—quartiles—as part of a broader effort to build descriptive statistics for social-science research.
1970
Tukey's Box-and-Whisker Plot
John Tukey introduced the boxplot in his landmark work on Exploratory Data Analysis (EDA), giving analysts a compact visual summary of five-number statistics and a systematic way to flag outliers.
1990s
Business Intelligence Adoption
With the rise of spreadsheet software and BI dashboards, boxplots and percentile-based KPIs (such as the 90th-percentile response time) became standard tools in corporate performance management and Six Sigma quality programs.

The central question these tools address is deceptively simple: Given a dataset of business observations, how can we efficiently communicate the shape, center, spread, and potential anomalies of the distribution without listing every data point? As we will see, percentiles quantify position, quartiles partition the data into meaningful segments, and boxplots translate all of this into a single, powerful graphic.

Core Principles & Definitions

Before diving into calculations, it is essential to establish precise definitions. Percentiles, quartiles, and boxplots are all rooted in the idea of order statistics—the values obtained by sorting a dataset from smallest to largest. Once data are ordered, we can assign each observation a position that indicates what fraction of the data falls at or below that value. The following foundational concepts underpin the entire lesson.

1

Percentile (Pₖ)

The k-th percentile is the value below which k% of the ordered observations fall. For example, if your salary is at the 85th percentile, you earn more than 85% of the comparison group.
2

Quartile (Q₁, Q₂, Q₃)

Quartiles are special percentiles that divide sorted data into four equal parts: Q₁ = P₂₅, Q₂ = P₅₀ (the median), and Q₃ = P₇₅. They provide a robust, outlier-resistant summary of distribution center and spread.
3

Interquartile Range (IQR)

The IQR equals Q₃ − Q₁ and captures the spread of the middle 50% of the data. It is the foundation for outlier detection and the length of the 'box' in a boxplot.
4

Five-Number Summary

The five-number summary consists of Minimum, Q₁, Median (Q₂), Q₃, and Maximum. It provides a concise snapshot of center, spread, and range without assuming a normal distribution.
5

Boxplot (Box-and-Whisker Plot)

A boxplot is a graphical representation of the five-number summary, with whiskers extending to non-outlier extremes and individual points marking observations beyond 1.5 × IQR from the box.
KEY TAKEAWAY
Think of percentiles like the mile markers on a highway. If you're at mile marker 75 on a 100-mile stretch, you know you've passed 75% of the journey. Similarly, the 75th percentile tells you that 75% of data points sit at or below that value. Quartiles are simply the 25-, 50-, and 75-mile markers—they split the highway into four equal segments so you can quickly gauge whether a particular observation is in the first leg, the middle stretch, or the home stretch of the distribution.

Anatomy of a Boxplot

A boxplot condenses an entire distribution into a single compact graphic. The diagram below illustrates the anatomy of a standard boxplot using a hypothetical dataset of quarterly revenue figures (in thousands of dollars). Each component maps directly to a positional statistic, making the plot both information-dense and visually intuitive.

The box spans from Q₁ to Q₃, capturing the interquartile range. The pink median line inside the box shows the center. Whiskers extend to the most extreme non-outlier values, while the red dot marks an outlier beyond 1.5 × IQR from Q₃.

In the diagram above, the box represents the middle 50% of quarterly revenues, immediately revealing whether the distribution is symmetric (median centered within the box) or skewed (median shifted toward one edge). Here, the median at $258K is closer to Q₁ ($198K) than to Q₃ ($318K), suggesting a mild right skew—a few quarters generated unusually high revenue, pulling the upper tail outward. The outlier at $510K exceeds the upper whisker fence of $498K (computed as Q₃ + 1.5 × IQR) and would warrant investigation: perhaps it reflects a seasonal spike or a one-time contract win.

Mathematical Framework

Several methods exist for computing percentiles, but the percentile locator formula is the most widely taught approach in business statistics courses. It identifies the position (index) in the sorted dataset that corresponds to a given percentile, and then interpolates if necessary.

PERCENTILE LOCATION
L = (k / 100) × n
where L is the locator (index position), k is the desired percentile (1–99), and n is the total number of sorted observations. If L is not a whole number, round up to the next integer to find the position. If L is a whole number, the percentile is the average of the values at positions L and L + 1.
QUARTILE DEFINITIONS
Q₁ = P₂₅ Q₂ = P₅₀ (Median) Q₃ = P₇₅
Quartiles are simply the 25th, 50th, and 75th percentiles. Apply the percentile locator formula with k = 25, 50, and 75 respectively to compute each.
INTERQUARTILE RANGE
IQR = Q₃ − Q₁
The IQR measures the spread of the central half of the data. Unlike the range (Max − Min), the IQR is resistant to extreme values and thus provides a more reliable measure of variability for skewed business data.
OUTLIER FENCES
Lower Fence = Q₁ − 1.5 × IQR Upper Fence = Q₃ + 1.5 × IQR
Any observation falling below the lower fence or above the upper fence is classified as an outlier. These fences define the maximum extent of the boxplot's whiskers; the whiskers actually terminate at the most extreme non-outlier data point within the fence.
📌 Convention Note
Multiple conventions exist for percentile computation (e.g., exclusive vs. inclusive methods, Excel's PERCENTILE.EXC vs. PERCENTILE.INC). The locator formula presented here aligns with the method used in most introductory business statistics textbooks (Anderson, Sweeney & Williams; Levine, Stephan & Szabat). Always verify which convention your course or software uses.

Interpreting Percentiles & Boxplots in Business Contexts

Positional statistics become most powerful when applied to real business scenarios. The table below maps common percentile and boxplot metrics to the business decisions they inform, spanning functions from human resources to supply-chain management.

Common business applications of percentile-based metrics
MetricBusiness ApplicationWhat It Reveals
P₉₀Customer service — response time SLAs90% of tickets are resolved within this threshold; useful for setting service-level agreements.
Q₁ (P₂₅)Compensation benchmarkingStarting salary floor for new hires positioned at the 25th percentile of the market.
Q₂ (Median)Real-estate pricingThe median home price is preferred over the mean because extreme luxury listings inflate the average.
IQRManufacturing quality controlA narrow IQR indicates consistent product dimensions; a wide IQR signals process variability.
Outliers (Boxplot)Fraud detection in financeTransactions flagged as outliers on a boxplot of daily spending merit investigation for potential fraud.
Comparing three regional sales distributions side by side. The West region has the highest median and widest IQR, indicating both stronger and more variable sales. The East region shows a single high outlier that merits investigation. The South region displays the most compact distribution with no outliers.

Side-by-side boxplots are especially valuable in business settings because they allow rapid visual comparison of distributions across categories—regions, product lines, time periods, or customer segments. Managers can instantly spot which group has higher central tendency, greater variability, or potential anomalies, all without examining tables of numbers. When the median line inside one box is visibly higher than the entire box of another group, the performance difference is striking and communicable even to non-technical stakeholders.

Worked Example — Employee Salary Analysis

Consider the following dataset of annual salaries (in $K) for 12 employees in a mid-sized consulting firm: 42, 48, 53, 55, 60, 64, 68, 72, 78, 85, 92, 145. We will compute the quartiles, IQR, outlier fences, and identify any outliers suitable for boxplot construction.

Quartiles, IQR & Outlier Detection
1
Step 1 — Sort the DataThe data are already sorted in ascending order: 42, 48, 53, 55, 60, 64, 68, 72, 78, 85, 92, 145. We have n = 12 observations.
n = 12
2
Step 2 — Compute Q₁ (25th Percentile)Apply the locator formula: L = (25/100) × 12 = 3.0. Since L is a whole number, Q₁ is the average of the values at positions 3 and 4: Q₁ = (53 + 55) / 2.
Q₁ = 54.0 ($K)
3
Step 3 — Compute Q₂ (Median, 50th Percentile)L = (50/100) × 12 = 6.0. Since L is a whole number, Q₂ is the average of positions 6 and 7: Q₂ = (64 + 68) / 2.
Q₂ = 66.0 ($K)
4
Step 4 — Compute Q₃ (75th Percentile)L = (75/100) × 12 = 9.0. Since L is a whole number, Q₃ is the average of positions 9 and 10: Q₃ = (78 + 85) / 2.
Q₃ = 81.5 ($K)
5
Step 5 — Compute IQR and Outlier FencesIQR = Q₃ − Q₁ = 81.5 − 54.0 = 27.5. Lower Fence = Q₁ − 1.5 × IQR = 54.0 − 41.25 = 12.75. Upper Fence = Q₃ + 1.5 × IQR = 81.5 + 41.25 = 122.75. Any observation below 12.75 or above 122.75 is an outlier.
IQR = 27.5 | Upper Fence = 122.75
6
Step 6 — Identify Outliers and Five-Number SummaryThe salary of $145K exceeds the upper fence of $122.75, so it is flagged as an outlier—perhaps a senior partner whose compensation is structured differently. The five-number summary is: Min = 42, Q₁ = 54.0, Median = 66.0, Q₃ = 81.5, Max (non-outlier) = 92. The boxplot whisker would extend from 42 to 92, with the $145K observation plotted as an individual point.
Outlier detected: $145K

Strengths, Limitations & Comparisons

Like all analytical tools, percentiles and boxplots have contexts in which they excel and contexts in which other methods may be more appropriate. Understanding these trade-offs is essential for choosing the right visualization or summary statistic when presenting findings to business stakeholders.

Strengths and limitations of boxplots for business analytics
CriterionStrengthsLimitations
Robustness to OutliersThe median and IQR are resistant to extreme values, making boxplots ideal for skewed business data such as incomes, housing prices, or claim amounts.Extreme outliers are shown but their magnitude is compressed; a point at $1M looks identical to one at $10M on a boxplot.
Distributional ShapeAsymmetry in the box and whiskers reveals skewness at a glance—valuable for quick executive presentations.Boxplots hide multimodality. A bimodal distribution (e.g., two customer segments) appears as a single box, masking important structure.
Comparison PowerSide-by-side boxplots allow instant comparison across groups using a common scale—effective for regional, temporal, or categorical comparisons.With very small datasets (n < 10), the quartiles may be imprecise and the boxplot can be misleading.
Data DensityBoxplots compress thousands of observations into a compact glyph, perfect for dashboards with limited screen real estate.Individual data patterns (clusters, gaps) are invisible. Histograms or violin plots may be preferred when data texture matters.
WHEN TO CHOOSE A BOXPLOT
Use a boxplot when your primary goal is to compare distributional position and spread across categories or time periods, especially when the data may be skewed. If you need to see the full shape of the distribution—including potential bimodality or density clusters—pair the boxplot with a histogram, overlay a strip plot of raw points, or upgrade to a violin plot. In practice, many BI tools (Tableau, Power BI) render boxplots natively, making them a go-to default for initial exploratory analysis.

Connection to Advanced Descriptive & Inferential Techniques

Percentiles and boxplots serve as a gateway to more sophisticated analyses. In advanced business analytics, positional measures extend into areas such as quantile regression, value-at-risk (VaR) in financial risk management, and percentile-based control charts in operations management. The table below contrasts the introductory tools you have learned with their more advanced counterparts.

From introductory positional measures to advanced business analytics
Introductory ToolAdvanced ExtensionBusiness Context
Percentile (Pₖ)Quantile RegressionModels how the entire distribution of a response variable (e.g., delivery time) changes with predictors, not just the mean—critical for logistics SLAs.
IQR Outlier FencesMahalanobis Distance / DBSCANMultivariate outlier detection for fraud analytics where anomalies span several variables simultaneously.
BoxplotViolin Plot / Ridgeline PlotDisplays full density shape alongside positional summaries—useful for large customer-behavior datasets.
Five-Number SummaryValue-at-Risk (VaR)The 1st or 5th percentile of a portfolio's return distribution estimates the maximum likely loss over a given time horizon.

As you advance through your statistics coursework and into electives in finance, operations, or marketing analytics, you will find that the conceptual foundation laid by percentiles and boxplots recurs constantly. Quantile thinking—asking not just 'What is the average?' but 'What happens at the extremes?'—is at the heart of risk management, supply-chain resilience planning, and customer-experience optimization. Mastering the basics here prepares you to engage confidently with these more nuanced frameworks.

Practice Problems

PROBLEM 1CONCEPTUAL
A manager states: 'Our average customer wait time is 4.2 minutes, so most customers wait about 4 minutes.' Explain why this claim could be misleading and describe how percentiles or a boxplot might reveal a more accurate picture of the customer experience.
PROBLEM 2BASIC CALCULATION
A dataset of 20 monthly profit figures (in $K) is sorted as follows: 12, 15, 18, 22, 25, 27, 30, 33, 35, 38, 40, 43, 45, 48, 52, 55, 60, 65, 70, 80. Compute Q₁, Q₂, and Q₃ using the percentile locator formula.
PROBLEM 3INTERMEDIATE
Using the dataset from Problem 2, determine the outlier fences and state whether any observations qualify as outliers. Then describe where the boxplot whiskers would terminate.
PROBLEM 4APPLIED
A logistics company tracks delivery times (in hours) for its 15 most recent shipments: 18, 20, 21, 22, 24, 24, 25, 26, 27, 28, 30, 32, 35, 48, 52. The company's SLA guarantees delivery within the 90th percentile time. Compute P₉₀ and explain whether the two longest delivery times should be classified as outliers using the 1.5 × IQR rule.
PROBLEM 5CRITICAL THINKING
A retail chain produces side-by-side boxplots comparing daily revenue across four store locations. Store A's entire box sits above Store B's box (no overlap between A's Q₁ and B's Q₃). Store C has a very narrow box but extremely long whiskers. Store D's median is almost touching Q₃. For each store comparison or observation, explain the business insight a manager should draw and recommend at least one follow-up analysis.

Lesson Summary

This lesson established a complete toolkit for describing distributional position and spread in business data. Percentiles rank any observation on a 0–100 scale using the locator formula L = (k/100) × n, enabling comparisons across datasets of different sizes. Quartiles (Q₁, Q₂, Q₃) are special cases at the 25th, 50th, and 75th percentiles, dividing the sorted data into four equal groups and forming the basis of the five-number summary (Min, Q₁, Median, Q₃, Max). The interquartile range (IQR) captures the spread of the middle 50%, and the 1.5 × IQR rule provides a systematic threshold for flagging outliers.

The boxplot translates the five-number summary into a compact visual—box for the IQR, line for the median, whiskers to the most extreme non-outlier values, and individual points for outliers. Side-by-side boxplots are especially powerful for comparing distributions across business categories such as regions, product lines, or time periods. Recognizing boxplot limitations—including the inability to reveal multimodality—guides analysts toward complementary visualizations like histograms and violin plots. These foundational skills prepare you for advanced techniques such as quantile regression and Value-at-Risk that you will encounter in finance and operations analytics courses.

Varsity Tutors • Business Statistics • Percentiles, Quartiles & Boxplots