Quick Answer: Which Average Should You Use in Business?
- Mean when values are reasonably balanced, every observation should count equally, and the total amount matters (e.g., tracking total monthly output across a team).
- Median when data is skewed or contains influential extreme values — income, transaction values, delivery times, house prices, or customer spending where a few large observations distort the picture.
- Mode when the most frequently occurring category or value is the business decision — the most popular product size, the most common payment method, or the most frequent complaint type.
- Weighted mean when observations differ in importance, volume, revenue, or time — for example, calculating one average price across products that sold in very different quantities.
- More than one measure when data is skewed or the stakes are high. Report the mean alongside the median to show both the total picture and the typical experience.
| Measure | Best For | Sensitive to Outliers? | Categorical Data? | Business Example | Main Limitation |
|---|---|---|---|---|---|
| Mean | Symmetric numerical data where totals matter | Yes — strongly | No | Average daily sales revenue | Distorted by extreme values |
| Median | Skewed distributions; income, prices, times | No — robust | Only ordinal | Median customer transaction value | Ignores actual distances between values |
| Mode | Most common category or discrete value | No | Yes — ideal | Most popular shoe size ordered | May not exist or may be unstable |
| Weighted Mean | When observations differ in importance or volume | Depends on weights | No | Average selling price weighted by units sold | Results depend on chosen weights |
What Are Mean, Median, and Mode?
Mean, median, and mode are the three standard measures of central tendency in descriptive statistics. Each answers the same basic question — "what is a typical value in this dataset?" — but they answer it differently, and that difference matters when the data is uneven or contains extreme observations.
The word "average" is often used loosely to mean the arithmetic mean, but it can refer to any of these three measures. Before accepting any reported "average" at face value, it's worth asking: average of what, measured how?
Mean (Arithmetic Average)
Add all values, then divide by the count. Represents the balance point of the data. Used when every observation contributes equally and the distribution is not heavily skewed.
Median (Middle Value)
Sort values from lowest to highest; the median is the middle observation. Used when data is skewed or contains extreme values that would distort the mean.
Mode (Most Frequent)
The value or category that appears most often. The only measure that works with categorical data. Used when the most common option is the decision that matters.
Weighted Mean
Each value is multiplied by its weight before summing. Used when some observations represent more units, revenue, or importance than others.
Mean in Business
Where x̄ is the sample mean, Σx is the sum of all observations, and n is the count of observations. When you're working with the entire population rather than a sample, the formula is the same but the result is called μ (mu) rather than x̄.
Businesses use the arithmetic mean for average daily sales revenue, average order value, average delivery time, average customer spending across a period, average employee performance scores, average production output per shift, and average revenue per region. The mean is the right choice when you want each observation to count equally and when you need the result to connect to a total — because x̄ × n = Σx, the mean multiplied by the count gives back the sum.
Different Denominators, Different Averages
Before calculating a business mean, define the denominator. The same revenue figure produces very different results depending on what you divide by:
| Business Metric | Formula | What It Measures |
|---|---|---|
| Average sales per transaction | Total revenue ÷ Number of transactions | Typical basket size |
| Average sales per customer | Total revenue ÷ Number of customers | Revenue value of each customer relationship |
| Average sales per day | Total revenue ÷ Number of trading days | Day-to-day revenue level for forecasting |
| Average sales per employee | Total revenue ÷ Number of employees | Team productivity |
| Average sales per store | Total revenue ÷ Number of stores | Typical store performance |
A retailer with $120,000 monthly revenue, 800 transactions, and 350 customers produces three valid averages: $150/transaction, $342.86/customer, and $0/unit (meaningless if you divided by SKUs). Each answers a different question. Specify the unit before you calculate.
Worked Example 1 — Average Daily Sales
Business question: A café wants to know its average daily revenue over a trading week.
Revenue by day (Monday–Friday): $1,840 · $2,110 · $1,760 · $2,240 · $3,050
Sum all values: $1,840 + $2,110 + $1,760 + $2,240 + $3,050 = $11,000
Count observations: n = 5 days
Apply the formula: x̄ = $11,000 / 5 = $2,200 per day
✅ Result: Average daily sales = $2,200. The café can use this to forecast that a four-week month (20 trading days) should generate approximately $44,000.
⚠️ Caution: Friday's $3,050 is notably higher than the other days. If Fridays are consistently busy, a simple weekly mean underestimates Fridays and overestimates other days. Consider reporting daily averages by day of week alongside the overall mean.
Median in Business
The median is a positional measure. It does not use the exact magnitude of every value the way the mean does — it only cares about rank order. This makes it less sensitive to extreme values on either end of the distribution.
Businesses use the median for customer transaction values, employee compensation, delivery times, property prices, invoice amounts, and any measure where a small number of unusually large or small observations would otherwise dominate the summary. Median salary figures are more informative than average salary figures in organizations where a few executives earn multiples of the typical employee's pay.
Worked Example 2 — Median Customer Spending with an Outlier
Business question: A software company tracks spending by nine customers in a quarter. Which average is more representative?
Customer spending ($): 820, 940, 1,050, 870, 990, 1,130, 880, 810, 18,500
Sort ascending: 810, 820, 870, 880, 940, 990, 1,050, 1,130, 18,500
Find the median (odd n=9): Position = (9+1)/2 = 5th value = $940
Calculate the mean: Sum = 810+820+870+880+940+990+1,050+1,130+18,500 = $25,990. Mean = $25,990/9 = $2,888
✅ Result: The mean ($2,888) is 3× higher than the median ($940) because the $18,500 enterprise deal pulls the mean sharply upward. Eight of nine customers spent under $1,200, so the median ($940) is far more representative of a typical customer's behavior. The mean better captures total revenue contribution.
⚠️ Caution: Neither measure is wrong. For pricing decisions and typical customer analysis, use the median. For revenue forecasting and capacity planning, the mean connects to the total and is more useful. Report both when the difference is large.
Mode in Business
The mode is most useful when the business decision is about what is most common, most popular, or most in demand. It answers "what should we stock more of?" or "what does the typical customer choose?" rather than "what is the arithmetic center?"
Practical business applications for the mode include the most frequently purchased product size or SKU, the most common payment method at checkout, the most popular delivery option, the most frequent order quantity from a supplier, the most common customer complaint category, and the most common reason for product returns.
Worked Example 3 — Mode for Product Size Selection
Business question: An online apparel retailer records the sizes ordered in one day. Which size should it prioritize for restock?
Sizes ordered: S, M, L, XL, M, L, M, S, M, L, M, XL, M, L, S, M, M, L, M, XL
Count by category: S = 3, M = 9, L = 5, XL = 3
Identify most frequent: M appears 9 times — the highest count
✅ Result: The mode is M (Medium). Medium accounts for 45% of orders and should be the first priority for restock. Note: mean and median are not meaningful for categorical size labels.
⚠️ Caution: A single day's data may not be stable. If M and L had been close (say, 8 vs 7), the mode would be unreliable as a restocking guide. Check across a longer period and track frequency over time before making large inventory commitments.
Weighted Mean and Weighted Average in Business
A simple arithmetic mean treats every observation as equally important. When observations represent different amounts — different sales volumes, different customer counts, different time periods, or different portfolio allocations — the simple mean gives misleading results. The weighted mean corrects for this.
x = each value (price, score, rate)
w = its weight (units, revenue, count)
Σ(w·x) = sum of (weight × value) products
Σw = total of all weights
Worked Example 4 — Weighted Average Selling Price
Business question: A retailer sells three versions of a product at different prices. What is the true average selling price, accounting for how many units each version sold?
| Version | Unit Price (x) | Units Sold (w) | w × x |
|---|---|---|---|
| Basic | $49 | 320 | $15,680 |
| Standard | $89 | 180 | $16,020 |
| Premium | $149 | 50 | $7,450 |
| Total | 550 | $39,150 |
Simple (unweighted) mean: ($49 + $89 + $149) / 3 = $95.67
Weighted mean: Σ(wx) / Σw = $39,150 / 550 = $71.18
✅ Result: The weighted average selling price is $71.18, not $95.67. The simple mean overstates the actual price because it treats all three versions as equally common, when in reality the lower-priced Basic version outsold Premium 6 to 1. Revenue planning, margin analysis, and pricing decisions all require the weighted figure.
Mean vs Median vs Mode vs Weighted Mean
Decision Table by Business Question
| Business Question | Recommended Measure | Why | Also Report |
|---|---|---|---|
| What was our average daily revenue last quarter? | Mean | Each day contributes equally; total revenue is the goal | Standard deviation; weekly breakdown |
| What does a typical customer spend? | Median | High spenders skew the mean; median shows the midpoint | Mean (for total revenue context); count |
| What product should we stock most of? | Mode | Most popular option drives stock decisions | Frequency count; % of total |
| What is our average selling price across products? | Weighted Mean | Products sell in different quantities; price must reflect volume | Volume breakdown by product tier |
| What is the typical salary in our company? | Median | Executive compensation pulls the mean high | Mean; salary bands by role |
| What is our average order value (AOV)? | Mean (standard AOV formula) | Total revenue / total orders; consistent industry metric | Median; distribution of order sizes |
| What delivery time can we promise customers? | Median or 90th percentile | Occasional delays skew mean; median is more dependable | Max; % on-time rate |
| What is the most common reason for returns? | Mode | Categories; most frequent reason is the priority to address | Frequency count; Pareto analysis |
How to Calculate Each Average
Mean Calculation — Step by Step
Add every value. Divide by the count. Order does not matter for the mean.
Median Calculation — Odd Number of Observations
Dataset: 14, 22, 9, 31, 18 → Sort: 9, 14, 18, 22, 31 → n = 5 → Position = (5+1)/2 = 3rd value → Median = 18
Median Calculation — Even Number of Observations
Dataset: 14, 22, 9, 31, 18, 25 → Sort: 9, 14, 18, 22, 25, 31 → n = 6 → Middle two = positions 3 and 4 = 18 and 22 → Median = (18 + 22) / 2 = 20
Mode Calculation — Numerical Values
Order quantities: 5, 10, 5, 20, 10, 5, 15, 5, 10 → Frequency: 5 appears 4 times, 10 appears 3 times, 15 and 20 once each → Mode = 5 units
Mode Calculation — Categories
Payment methods: Card, Cash, Card, Card, PayPal, Cash, Card, PayPal, Card → Card appears 5 times, Cash 2, PayPal 2 → Mode = Card
Worked Example 5 — How an Outlier Changes the Mean
Business question: A logistics team records six delivery times (in hours). How does one delayed shipment change the summary statistics?
Delivery times without outlier: 2.1, 2.4, 2.7, 2.5, 2.3, 2.6 (hours)
Without outlier: Mean = (2.1+2.4+2.7+2.5+2.3+2.6)/6 = 14.6/6 = 2.43 hrs. Sorted: 2.1, 2.3, 2.4, 2.5, 2.6, 2.7 → Median = (2.4+2.5)/2 = 2.45 hrs
Add one outlier (18.0 hrs — a customs delay): New mean = (14.6+18.0)/7 = 32.6/7 = 4.66 hrs. Sorted: 2.1, 2.3, 2.4, 2.5, 2.6, 2.7, 18.0 → Median (n=7, position 4) = 2.5 hrs
| Statistic | Without Outlier | With Outlier (18 hrs) | Change |
|---|---|---|---|
| Mean | 2.43 hrs | 4.66 hrs | +92% |
| Median | 2.45 hrs | 2.50 hrs | +2% |
✅ Result: The mean nearly doubled with one delayed shipment. The median barely moved. For customer-facing delivery promises, the median is a more honest indicator of typical performance. The outlier should be investigated separately — was it an error, a genuine event, or something that needs a process fix?
Average Sales in Business
"Average sales" is one of the most commonly reported and most frequently miscalculated business metrics. Before anyone reports or reads an average sales figure, the denominator needs to be defined and documented. Two managers can report entirely different numbers from the same data — and both can be correct — if they're dividing by different units.
When Average Sales Can Mislead
Averaging averages — computing the mean of a set of averages calculated from groups of different sizes — produces incorrect results unless weighted by the original group sizes. A store with 20 transactions averaging $80 and a store with 200 transactions averaging $60 does not have a combined average of ($80+$60)/2 = $70. The correct combined average is ($1,600 + $12,000) / 220 = $62.55.
Worked Example 6 — Average Sales by Salesperson
Business question: Three salespeople generated the following revenue last month. Which metric should the manager use to set targets?
| Salesperson | Total Revenue | Deals Closed | Avg per Deal | Customers | Avg per Customer |
|---|---|---|---|---|---|
| Alex | $48,000 | 16 | $3,000 | 12 | $4,000 |
| Jamie | $52,000 | 40 | $1,300 | 35 | $1,486 |
| Sam | $44,000 | 22 | $2,000 | 20 | $2,200 |
| Team | $144,000 | 78 | $1,846 | 67 | $2,149 |
Team average per deal: $144,000 / 78 = $1,846 (weighted by deal count, not a simple mean of $3,000+$1,300+$2,000)
Note the contrast: Alex averages $3,000/deal but closes only 16. Jamie averages $1,300/deal but closes 40. Revenue is similar, but the strategies are opposite.
✅ Result: The team average per deal is $1,846. For target-setting, the manager should distinguish deal size from deal volume and avoid using any single average as the only KPI. A salesperson can hit the average deal-size target while underperforming on volume, and vice versa.
Median Income and Other Skewed Business Data
Income distributions are right-skewed: most people earn in a moderate range, but a small number earn vastly more. When the top earners are included in the calculation, the mean rises well above what most people actually earn. The median — the income of the person in the exact middle of the ordered distribution — resists that pull and gives a more honest picture of what is typical.
Worked Example 7 — Median vs Mean Salary Distribution
Business question: A company's 11 employees earn the following annual salaries. What is the typical employee's salary?
Salaries ($): 32,000 · 36,000 · 38,000 · 40,000 · 42,000 · 44,000 · 46,000 · 48,000 · 52,000 · 58,000 · 380,000 (CEO)
Mean: Sum = $816,000. Mean = $816,000 / 11 = $74,182
Median (n=11, already sorted): Position = (11+1)/2 = 6th value = $44,000
✅ Result: Nine of eleven employees earn less than the mean of $74,182. The median of $44,000 describes what a typical employee actually earns. The mean is accurate for total payroll calculation ($816,000) but misrepresents the employee experience. HR communications, job postings, and benchmarking should use the median.
The same pattern appears in customer spending, invoice values, project costs, claim amounts, and any business metric where a few large observations occur alongside many small ones. The rule of thumb: when the mean and median differ substantially, report both and explain the distribution.
Outliers and Averages
An outlier is an observation that is unusually distant from the rest of the data. There is no single universal rule for what counts as an outlier — a value might be flagged using the interquartile range method (more than 1.5 × IQR below Q1 or above Q3), using z-scores, or through domain knowledge. The appropriate method depends on the data and the business context. See the outliers in statistics guide for formal detection methods.
Outliers occur in business data for many reasons: data entry errors, fraud, legitimate one-time events (a single enormous bulk purchase), product launches, seasonal spikes, or system glitches. An outlier is not automatically an error, and removing it without investigation misrepresents the data.
How to Handle Outliers Responsibly
Investigate Before Acting
Check whether the value is a data entry error, a system issue, or a legitimate unusual event. A $50,000 transaction at a store with typical $40 tickets could be fraud, a data error, or a corporate bulk order — each requiring a different response.
Correct Errors, Document Exclusions
Correct confirmed data errors. If you exclude a legitimate observation under a pre-declared rule (e.g., "transactions over $10,000 are classified as B2B and analyzed separately"), document it clearly so readers know the scope of the reported figure.
Report the Median Alongside the Mean
When outliers are present or the distribution is skewed, report both. If they're close, the data is likely symmetric and the mean is reliable. If they differ substantially, note the distribution and explain what each figure represents.
Never Remove to Get a Preferred Result
Removing observations solely because they make the number look bad is data manipulation. If the outlier is real, it belongs in the analysis — or in a separate category with transparent labeling.
Business Data Summary: What Should You Report?
A single average rarely tells the full story. A complete business data summary lets readers understand not just the center of the data but its spread, its shape, and how many observations it's based on. The following is a sample summary for a hypothetical ecommerce store's order values over one month.
| Statistic | Value | What It Tells You |
|---|---|---|
| Count (n) | 1,240 orders | How many observations the summary covers |
| Total revenue | $98,300 | Actual business impact |
| Mean (average order value) | $79.27 | Revenue per order; connects to the total |
| Median order value | $62.00 | What a typical order actually looks like |
| Mode (most common order value) | $49.99 | Most popular price point; pricing anchor |
| Minimum | $9.99 | Floor of the distribution |
| Maximum | $1,840.00 | Highest individual order; context for the mean |
| Q1 (25th percentile) | $38.50 | 25% of orders were below this |
| Q3 (75th percentile) | $94.00 | 75% of orders were below this |
| IQR (Q3 − Q1) | $55.50 | Spread of the middle 50% of orders |
| Standard deviation | $88.40 | Average distance from the mean; high SD suggests spread |
| Period | August 2026 | Scope — essential for comparisons |
The mean ($79.27) is well above the median ($62.00) because a few large orders — visible in the $1,840 maximum and the high standard deviation — pull the mean upward. Knowing both figures tells a stakeholder: "Most customers spend around $62, but our high-value orders lift the revenue average. Strategies to increase order frequency should target the typical $62 customer. Strategies to grow AOV should focus on the conditions that produce the larger orders."
Choosing the Right Measure for Different Business Functions
| Business Function | Typical Question | Recommended Measure | Caveat |
|---|---|---|---|
| Sales & Revenue | What is our average transaction value? | Mean (AOV = revenue ÷ orders) | Report median alongside when outliers are present |
| Marketing | What is the typical customer acquisition cost? | Median (skewed by expensive campaigns) | Mean useful for budgeting total spend |
| Ecommerce | What is the most popular price tier? | Mode (most common order value or price bracket) | Use across a meaningful sample size |
| Finance & Costs | What is average cost of goods across products? | Weighted mean (by units produced or sold) | Weights must reflect actual production volume |
| HR & Compensation | What is the typical employee salary? | Median (executive pay skews mean) | Report by band or level for fair comparison |
| Operations | What is our usual delivery time? | Median or 90th percentile | Mean inflated by occasional long delays |
| Inventory | Which SKU should we reorder first? | Mode (most frequently ordered item) | Combine with sales velocity and margin data |
| Customer Support | What is the typical resolution time? | Median | A few complex cases extend the mean; median is more representative |
| Manufacturing | What is average output per shift? | Mean | Check standard deviation — high spread means inconsistent quality |
| KPI Dashboards | What single number summarizes performance? | Mean (for totals-linked KPIs) or Median (for experience KPIs) | Always include count and time period |
Mean, Median, and Mode Calculator
Calculate Mean, Median & Mode — Enter Your Numbers
Weighted Mean Calculator
Calculate a Weighted Average — Enter Value and Weight Pairs
Mean, Median, and Mode in Business Decision-Making
The right average for a business decision is not determined by convention or by which measure produces the most favorable number. It's determined by the data structure, the business question, and what the audience needs to understand. Here is a repeatable ten-step framework.
Define the Business Question
Write the question in one sentence before touching the data. "What is the typical order value?" and "What is our total revenue per order?" are different questions that may call for different measures.
Identify the Unit of Analysis
Specify what each observation represents: a transaction, a customer, a day, a product, or an employee. The denominator in your formula must match this unit.
Confirm Population, Period, and Source
State exactly which records are included, the time period covered, and where the data comes from. An average without these boundaries cannot be compared to another.
Check Data Quality and Missing Values
Gaps, duplicates, and data entry errors affect all averages. A mean calculated on data with undetected errors is wrong, regardless of how carefully the arithmetic is done.
Sort and Visualize the Data
A quick histogram or box plot often reveals skewness or outliers before you calculate anything. See the data visualization guide for chart selection guidance.
Examine the Distribution
Is it symmetric? Skewed right or left? Are there outliers? Roughly symmetric data with no extreme values works well with the mean. Skewed data or outliers call for the median. Categorical data calls for the mode.
Select Mean, Median, Mode, Weighted Mean, or a Combination
Apply the decision table in Section 7. When the mean and median diverge noticeably, report both. Neither cancels the other out — they answer complementary questions.
Report with Units, Count, and Context
"$62" means nothing without context. "$62 median order value, n = 1,240 orders, August 2026" is a complete, comparable statistic.
Test Stability Across Segments and Time
Does the average change significantly if you split the data by region, product line, or month? Averages that vary substantially across segments may need to be reported separately rather than combined.
Connect the Statistic to a Decision Without Overstating It
A descriptive average describes what happened. It does not prove why it happened, predict what will happen next, or guarantee results. State what the statistic shows, then add appropriate qualifications.
Common Mistakes When Using Averages in Business
| Mistake | What Goes Wrong | How to Avoid It |
|---|---|---|
| Treating every "average" as the mean | Median and mode are both valid averages; each means something different | Always specify which measure you're reporting |
| Using the mean for skewed data | A few extreme values distort the mean away from the typical value | Check mean vs median; if they differ greatly, use the median or report both |
| Ignoring outliers | Unexamined outliers inflate or deflate the mean without anyone noticing | Visualize the data; investigate every unusual observation before summarizing |
| Averaging averages incorrectly | The average of group averages is only valid when all groups are the same size | Always go back to the raw data and compute the weighted average |
| Wrong denominator | Dividing by the wrong unit produces a meaningless number | State the denominator in writing before you divide |
| Comparing averages from different periods or groups | Changes in the composition of the group, not actual performance changes, can shift an average | Match the population and time window precisely before comparing |
| Confusing median with mode | The median is a position (middle value); the mode is the most frequent value — they are not interchangeable | Use clear labels; mode is for "most common," median is for "middle of the distribution" |
| Reporting without sample size or time period | An average without context cannot be verified or compared | Always include n, the time period, and the data source |
| Using mode for continuous data with near-unique values | If every value is different, there is no mode; the metric is meaningless | Use mode only for discrete values or categories with genuine repetition |
| Making causal claims from descriptive averages | "Higher average spend caused satisfaction" — a descriptive statistic describes; it does not prove causation | Use language like "associated with" or "observed alongside," not "caused" |
Worked Example 8 — Comparing Mean and Median Before Making a Pricing Decision
Business question: A SaaS company wants to set a "standard plan" price that reflects what its current users pay. Should it use the mean or median monthly spend?
Monthly spend across 10 users ($): 29, 49, 49, 79, 49, 99, 49, 199, 29, 4,999 (one enterprise account on a legacy deal)
Sort: 29, 29, 49, 49, 49, 49, 79, 99, 199, 4,999
Mean: Sum = $5,630. Mean = $5,630 / 10 = $563
Median (even n=10): Positions 5 and 6 = $49 and $49. Median = $49
Mode: $49 appears 4 times = $49
✅ Result: Setting a "standard plan" at the mean price of $563 would price out 90% of current users. The median and mode both point to $49 as the most representative and most common price point. The enterprise account at $4,999 belongs in a separate enterprise tier analysis. For pricing a new plan to appeal to typical users, the median and mode are the correct inputs.
⚠️ Limitation: This is a tiny sample (10 users). With more data, the mode and median might shift. Pricing decisions should use the largest available dataset and should account for willingness-to-pay research, not just historical spend.
Visualizing Business Averages and Outliers
A number without visual context is harder to interpret than a chart that shows the distribution around it. Each chart type serves a different purpose.
Histogram
Shows the distribution of values across ranges. Use it to spot skewness and decide whether the mean or median is more appropriate. Available at the histogram maker.
Box Plot
Shows the median, quartiles, and potential outliers in one chart. The box spans Q1 to Q3; the line inside is the median; points beyond the whiskers are outliers. Try the box plot generator.
Bar Chart
Compares categories or shows the mode visually. Ideal for showing which product, payment method, or complaint type is most common.
Line Chart
Tracks average sales, median transaction value, or any summary statistic over time. Reveals trends, seasonality, and anomalies that a single average obscures.
Scatter Plot
Compares two numerical variables. Shows relationships between, say, deal size and close time, without reducing either to an average prematurely.
KPI Table
Presents multiple summary statistics side by side for management reporting. Always include n, time period, and at least one measure of spread alongside the average.
Common Visualization Mistakes to Avoid
Showing only a mean line on a chart without the surrounding distribution hides whether the data is symmetric, skewed, or has outliers. Truncating the y-axis makes differences appear larger than they are. Comparing bar heights without disclosing different sample sizes creates false impressions of relative performance. Omitting the sample size and time period from a chart makes it impossible to verify or reproduce.
Frequently Asked Questions
Key Takeaways
Summary
When to Use Each Measure
The right average depends on the data, not on convention or habit.
- The mean is the appropriate first choice for symmetric numerical data where every observation should count equally and the result needs to connect to a total. Check it against the median before reporting.
- The median is more representative than the mean whenever data is skewed or contains extreme values — income, prices, delivery times, order values, and salaries commonly fall into this category.
- The mode answers "what is most common?" and is the only measure that works with categorical data. It is most useful for inventory, product, and preference decisions.
- The weighted mean is required when observations differ in importance, volume, or scale. Using an unweighted mean in those situations produces a result that does not reflect the actual business situation.
- Always report averages with the count, time period, unit, and data source. Consider including at least one measure of spread — standard deviation, interquartile range, or the min-to-max range — so readers understand variation, not just the center.
- When the mean and median differ substantially, report both and explain why. The difference itself is informative — it signals a skewed distribution or the presence of outliers.
References
- NIST/SEMATECH. e-Handbook of Statistical Methods: Measures of Location. National Institute of Standards and Technology. itl.nist.gov
- U.S. Census Bureau. Income and Poverty in the United States. census.gov
- UK Office for National Statistics. Personal and Household Finances: Income and Wealth. ons.gov.uk
- Weisberg, H.F. (1992). Central Tendency and Variability. Sage University Paper. ISBN 0-8039-4007-6.
- Statistics Fundamentals. Descriptive Statistics Guide. statisticsfundamentals.com/descriptive-statistics/
- Statistics Fundamentals. Weighted Mean Explained. statisticsfundamentals.com/descriptive-statistics/weighted-mean/
Related Guides on Statistics Fundamentals
| Topic | Link | Why It's Relevant |
|---|---|---|
| Descriptive Statistics | descriptive-statistics/ | The parent framework that mean, median, and mode belong to |
| Standard Deviation | standard-deviation/ | Measure of spread to report alongside averages |
| Outliers in Statistics | outliers/ | How to detect and handle extreme values |
| Interquartile Range | interquartile-range/ | Robust spread measure for skewed data |
| Percentiles | percentiles/ | Context for where any value sits in the distribution |
| Weighted Mean | weighted-mean/ | Deep dive into weighted averages |
| Mean Examples | mean-examples/ | Additional worked examples for the arithmetic mean |
| Five-Number Summary | five-number-summary/ | Complete data summary beyond a single average |
| Data Visualization Guide | data-visualization/ | Choosing the right chart for business averages |
| Sales Forecasting with Regression | sales-forecasting-regression/ | Moving beyond averages to predict future sales |
| Business Decision-Making | business-decision-making/ | Connecting statistics to real business choices |
| Mean & Median Calculator | calculators/mean-median-mode/ | Online calculator for all three measures |