GainHQ
Go back

What Is Descriptive Analytics? Meaning, Examples and a Worked Calculation

Donut chart and histogram with mean, median, mode and range labels

by GainHQ Editorial Team | Oct 20, 2026 | Software Development Insights | Editorial policy


Table of Contents
  1. What is the meaning of descriptive analysis?
  2. The four types of analytics
  3. Descriptive analytics vs descriptive statistics
  4. The measures behind descriptive analytics
  5. Worked example: 12 months of sign-ups
  6. How to do descriptive analytics: step by step
  7. Descriptive analytics examples by department
  8. Common pitfalls
  9. Tools for descriptive analytics
  10. How descriptive analytics feeds what comes next
  11. How GainHQ can help

Descriptive analytics is the analysis of historical data to summarise what has already happened: totals, averages, trends, distributions and comparisons over a period. It answers "what happened?" and "how much?", not "why?" or "what next?". INFORMS describes it as the branch of analytics that gives insight into past events using historical data, alongside predictive and prescriptive analytics.

Most business reporting is descriptive: a monthly revenue dashboard, a churn rate, average ticket resolution time, a sales-by-region chart. It is the base layer that every other kind of analytics depends on, because you cannot explain or forecast a number you have not measured correctly.

What is the meaning of descriptive analysis?

Descriptive analysis means condensing a large set of raw records into a few numbers and charts that describe it accurately. Instead of reading 50,000 orders, you report that the median order was $64, that 80% of orders came from three regions, and that volume rose 12% on the previous quarter. The goal is accuracy and clarity, not prediction.

The four types of analytics

Descriptive analytics is usually shown as the first of four types. INFORMS itself lists three (descriptive, predictive, prescriptive); the fourth, diagnostic, is a common industry addition.

Type Question it answers Typical output Example
Descriptive What happened? Totals, averages, trends, dashboards Sign-ups fell 8% in October
Diagnostic Why did it happen? Drill-downs, segment comparisons The drop came from one paid channel
Predictive What is likely to happen? Forecasts, risk scores November sign-ups forecast at 540
Prescriptive What should we do? Recommended actions, optimisation Shift budget to the best-converting channel

Each step builds on the one before. A forecast trained on badly summarised history will be wrong in the same way. Our guide to predictive analytics software covers the next step.

Descriptive analytics vs descriptive statistics

The two terms overlap but are not the same. Descriptive statistics are the mathematical measures (mean, median, standard deviation, percentiles). Descriptive analytics is the business practice of collecting data, choosing the right measures, comparing periods and segments, and presenting the result so people can act on it. Statistics are the tools; analytics is the job.

The measures behind descriptive analytics

The NIST/SEMATECH e-Handbook of Statistical Methods groups the core summary measures into location and scale. In practice you need five families:

Family Measures What it tells you
Frequency Count, percentage, share How often something happens
Central tendency (location) Mean, median, mode The "typical" value
Dispersion (scale) Range, variance, standard deviation, interquartile range How spread out values are
Position Percentiles, quartiles Where a value sits in the distribution
Change Growth rate, period-over-period change How values move over time

Two definitions matter most. The median is the middle value once data is sorted, so it is not pulled around by extreme values. The interquartile range (IQR) is the 75th percentile minus the 25th percentile (NIST), which describes the spread of the middle half of the data.

Worked example: 12 months of sign-ups

The numbers below are an illustrative dataset (not GainHQ data) for a SaaS product's monthly sign-ups. They show how each measure changes the story.

Month Sign-ups Change vs previous month
Jan 410 —
Feb 385 −6.1%
Mar 450 +16.9%
Apr 470 +4.4%
May 430 −8.5%
Jun 520 +20.9%
Jul 495 −4.8%
Aug 480 −3.0%
Sep 610 +27.1%
Oct 560 −8.2%
Nov 590 +5.4%
Dec 1,240 +110.2%

Step 1 — Totals. 6,640 sign-ups for the year. The second half (3,975) was 49.2% higher than the first half (2,665).

Step 2 — Central tendency. The mean is 6,640 ÷ 12 = 553.3. The median (average of the 6th and 7th sorted values, 480 and 495) is 487.5. There is no mode: no value repeats. The mean sits 66 above the median because one month is far above the rest.

Step 3 — Spread. The range is 1,240 − 385 = 855. The sample standard deviation is 227.2. Using the same interpolation as Excel's QUARTILE.INC, Q1 = 445 and Q3 = 567.5, so the IQR is 122.5.

Step 4 — Outliers. A common rule flags values above Q3 + 1.5 × IQR = 751.25. December (1,240) is far above it and alone accounts for 18.7% of the year. Remove December and the mean drops to 490.9, the median to 480 and the standard deviation to 73.0.

Step 5 — Report. A good descriptive summary would say: "A typical month brought about 490 sign-ups. Volume grew through the year, and December was an outlier at 1,240 (we should confirm the cause before treating it as a trend)." That last clause is where descriptive work hands over to diagnostic work.

The same steps work in a spreadsheet: =AVERAGE(), =MEDIAN(), =STDEV.S(), =QUARTILE.INC(range,1) and =QUARTILE.INC(range,3).

How to do descriptive analytics: step by step

  1. Start with a question. "How did sign-ups change this year, and which months were unusual?" A question decides which data and measures you need.
  2. Collect the data. Pull from the source system (product database, CRM, billing, analytics tool) for a clearly defined period.
  3. Clean it. Remove duplicates and test accounts, fix time zones, and agree definitions (what counts as a "sign-up" or an "active user").
  4. Summarise. Calculate counts, central tendency, spread and change over time. Always report the median next to the mean when data can be skewed.
  5. Segment. Break the totals down by channel, region, plan or cohort. Many surprises only appear at this level.
  6. Visualise. Use line charts for trends, bar charts for comparisons and histograms or box plots for distributions.
  7. Publish and monitor. Put the recurring numbers in a dashboard with a stable definition so the same question gets the same answer next month.

If you are building internal reporting, our guide to custom dashboard development covers how to scope one.

Descriptive analytics examples by department

Team Descriptive question Typical measure
Sales How did revenue change by region last quarter? Revenue by region, quarter-over-quarter %
Marketing Which channels brought sign-ups this month? Sign-ups and share by channel
Product How many accounts used the new feature in week one? Active accounts, adoption %
Customer support How long do tickets take to resolve? Median and 90th percentile resolution time
Finance What did we spend by category? Spend by category, variance to budget
HR How has headcount and attrition changed? Headcount, attrition rate
Engineering How often do we deploy? Deployments per week (see our engineering KPIs guide)

Public statistics are descriptive analytics too. The US Census Bureau headlines median household income ($83,730 in 2024) rather than a mean, because income is skewed: a small number of very high incomes would pull an average up.

Common pitfalls

Averages hide the shape of the data. In 1973, Francis Anscombe published four small datasets with near-identical means, variances and correlations that look completely different when plotted (Anscombe, 1973). Always chart the data before you summarise it.

Totals can reverse when you segment them (Simpson's paradox). In fall 1973 graduate admissions at UC Berkeley, about 44% of 8,442 male applicants were admitted versus about 35% of 4,321 female applicants. Compared department by department, the data showed a small bias in favour of women; women had mostly applied to departments that were harder to get into (Bickel, Hammel and O'Connell, Science, 1975). An aggregate number can point the opposite way from every segment inside it.

Mean vs median on skewed data. As the sign-up example shows, one extreme value can move the mean a long way. Report both.

Changing definitions. If "active user" meant something different last year, a trend line is comparing two different things. Write definitions down.

Description is not explanation. Descriptive analytics can show that sign-ups jumped in December. It cannot tell you why. Treat any cause as a hypothesis until you test it.

Tools for descriptive analytics

Tool type Good for Limits
Spreadsheets (Excel, Google Sheets) Small datasets, one-off analysis, quick charts Manual refresh, error-prone at scale
BI tools (Power BI, Tableau, Looker Studio) Recurring dashboards, sharing, drill-down Need clean, modelled data underneath
SQL on a database or warehouse Large datasets, repeatable metric definitions Needs SQL skills; output still needs charting
Python or R (pandas, dplyr) Complex cleaning, statistics, automation Needs programming skills
Product analytics tools Event and funnel reporting Limited to the data they collect

Most teams use two or three together: SQL to define metrics once, a BI tool to show them, and a spreadsheet for quick questions. For a one-off question your standard reports don't answer, see our guide to ad hoc analysis and reporting.

How descriptive analytics feeds what comes next

Descriptive work produces three things later stages need: a clean baseline ("a normal month is about 490 sign-ups"), flagged anomalies to investigate (December), and well-defined historical series to train forecasts on. Teams that skip it tend to build predictive models on inconsistent data. For the difference between forecasting approaches, see generative AI vs predictive AI.

How GainHQ can help

GainHQ has spent 15+ years building software and runs its own products, including Payrun (HR and payroll) and Gain.io (CRM and helpdesk). If you need reporting built into your product or an internal dashboard on top of your data, see our custom software and SaaS development services.

Frequently asked questions

What is descriptive analytics in simple terms?

It is summarising past data to show what happened, using totals, averages, trends and charts. It does not explain causes or predict the future.

What is an example of descriptive analytics?

A monthly report showing that sign-ups rose from 410 in January to 590 in November, with a median of about 490 per month, is descriptive analytics. So is a sales-by-region dashboard or an average ticket resolution time.

What are the four types of analytics?

Descriptive (what happened), diagnostic (why it happened), predictive (what is likely to happen) and prescriptive (what to do about it). INFORMS formally lists three; diagnostic is a common industry addition.

What is the difference between descriptive and predictive analytics?

Descriptive analytics summarises historical data. Predictive analytics uses that history to estimate future outcomes, such as next month's demand or a customer's churn risk.

What is the difference between descriptive analytics and descriptive statistics?

Descriptive statistics are the measures (mean, median, standard deviation). Descriptive analytics is the wider process of collecting, cleaning, summarising and presenting data for decisions.

What measures are used in descriptive analytics?

Counts and percentages, mean, median and mode, range, standard deviation and interquartile range, percentiles, and period-over-period change.

What tools are used for descriptive analytics?

Spreadsheets, BI tools such as Power BI, Tableau or Looker Studio, SQL, and Python or R. Most teams combine a database, a BI tool and spreadsheets.

What are the limitations of descriptive analytics?

It only describes the past, it cannot show cause, and summary numbers can mislead when data is skewed, badly defined or aggregated across very different groups.

Related Blogs