What is Descriptive Analytics?

Descriptive analytics looks at what happened in the past.

It analyzes current and historical data to understand past events and trends, providing a clear picture without making predictions or suggesting actions—this usually comes later.

Descriptive analytics is the first port of call on most data science voyages, helping businesses make sense of their data, identify patterns, and draw conclusions. They can use this information to apply more advanced analytics approaches, like predictive and prescriptive, to forecast future outcomes and get actionable recommendations.

In descriptive analytics, we ask “What happened?” (or sometimes “what is happening?”), typically focusing on specific occasions or timeframes.

You might present these findings in the form of financial reports, product sales figures, or marketing campaign outcomes—essentially, any area where a business wants to gather and share the results of certain events.

What is the relationship between descriptive and predictive analytics?

You can’t look to the future without considering the past—predictions are meaningless without a robust understanding of data and its context.

This is why descriptive analytics is integral to predictive analytics. It helps lay the foundation for deeper analysis by giving essential insight into the data, such as highlighting important patterns, trends, and relationships.

This background helps inform predictive modeling techniques by:

• Identifying important and influential variables
• Highlighting the data’s characteristics
• Providing comparisons for the predictive models.

Predictive models, supported by historical analysis, help businesses determine what might happen in the future. This enables them to make more informed decisions based on potential outcomes.

Descriptive analytics' main task is analyzing historical data. Its techniques and approaches might change depending on the data type, where it comes from, and what the business is trying to discover.

You’ll likely see this process broken down into these several phases:

• Data collection: As with other analytical models, the first step is to collect the data. Data can come from various sources, like databases, spreadsheets, warehouses, or other data repositories. You’ll usually collect information relating to business goals. For example, if you want to find out why a product didn’t sell as expected, you could collect purchase data, website usage, and customer information.
• Data cleaning and preparation: Once you’ve collected the data, it must be cleaned and preprocessed to remove any inconsistencies, errors, or missing values. This ensures the data is accurate and reliable for analysis.
• Data exploration: Analysts visually explore data using several tools and techniques during this phase. They’ll identify patterns, trends, outliers, and other important data characteristics. This forms the basis for answering questions like “what happened?”
• Summary statistics: Descriptive analytics also involves calculating and presenting summary statistics. These include things like measures of central tendency (mean, median, mode), and measures of dispersion (standard deviation, range) to give a concise overview of the data’s distribution and characteristics.
• Data visualization and communication: You can use bar charts, line graphs, pie charts, histograms, scatter plots and other visual representations to present data patterns and relationships clearly. Reports and interactive dashboards are also useful for communication, especially if you want to show real-time or periodic updates. Presenting the information in an understandable format makes it easier for stakeholders and other interested parties to grasp key insights from the findings.

Once you’ve gathered, explored, and visualized the data, you can interpret the findings and generate insights.

Businesses might choose to collect data over time, in which case they’ll apply time series analysis to help them spot trends or other information. They could also segment the data based on different characteristics (e.g., customer demographics, product categories, etc.) to find differences and similarities between groups.

This process offers a solid way for organizations to understand their current performance, identify strengths and weaknesses, and detect potential opportunities or issues—vital for improving business growth and efficiency.

Descriptive analytics uses several measures to summarize and understand data.

They’re the mathematical techniques behind descriptive analytics and what makes the whole process work.

These measures help analysts uncover the data’s characteristics and distribution, so businesses can discern what it actually indicates.

Measures of frequency

Measures of frequency tell us how often something has happened or is likely to happen, such as when looking at a specific response or event.

Typical measures include:

• Count: The number of times each value appears
• Percentage: The proportion of each value’s occurrence relative to the total number of data points.

Example: A customer support team receives 100 customer feature requests over the month and wants to know the number of times each feature or enhancement request appears. Counting the amount helps show how popular each request is.

Measures of dispersion

Also called variability measures, these show the spread of data points and how much it deviates from the central tendency.

Common measures are:

• Range: The difference between maximum and minimum values.
• Variance: Average of the squared differences between each data point and the mean.
• Standard deviation: Square root of the variance, which gives a more interpretable measure of dispersion.

Example: A cloud storage solution wants to know the dispersion of file size among its users. They use range to see the spread of file size from smallest to largest and standard deviation to understand how spread out the file sizes are from the average file size. This information helps the business optimize its infrastructure and meet customers’ storage needs.

Measures of position

Measures of position help find a data point's position in relation to others. Percentiles are the most used measures of position and work by dividing the data into 100 equal parts.

Typical percentiles include the median (50th percentile), quartiles (25th and 7th percentiles), and deciles (10th, 20th, 30th etc., percentiles).

Example: An online marketplace for freelancers wants to analyze hourly rates. It uses the median to see the typical rate freelancers charge and quartiles to see the spread and variability of rates.

Measures of central tendency

These measures show the central or average value around where the data clusters.

Common measures are:

• Mean: Average of all the data points, calculated as the sum of the values divided by the number of data points
• Median: The middle value of the dataset, separating the higher and lower half
• Mode: The value that occurs most often.

Example: An e-learning platform wants to examine its students' average course completion time. Mean provides the typical time, median shows the central completion time, and mode highlights students' most common completion time.

Understanding the current state of affairs is a must for every business, which is why most industries apply descriptive analytics in some way.

Even at this foundational level, businesses can use descriptive analytics to look at their historical data and make decisions to optimize processes.

Let’s examine the areas and scenarios organizations can apply these measures.

Financial analysis

Descriptive analytics in financial analysis involves analyzing historical financial data to get insight into a company’s revenue, expenses, profit, and cash flow.

Central tendency and dispersion measures help assess financial indicators’ variability and stability. Businesses can also use data visualization techniques to represent trends and highlight key financial ratios.

Aggregated survey analysis

In aggregated survey analysis, businesses use descriptive analysis to examine data from multiple survey respondents, summarizing the answers to identify prevailing opinions and trends.

Businesses could count the frequency of responses for each question to find the most common or use measures of central tendency to show numerical survey responses.

Demand trends

Descriptive analytics in demand trending focuses on historical sales data to understand patterns and fluctuations in demand for products and services.

Organizations can apply time series analysis to identify seasonality, cyclic patterns, and overall growth or decline in demand and use data visualization to show these changes over different periods.

Goal progression analysis

Descriptive analytics helps track and visualize goal or KPI progress to monitor their attainment.

Businesses can compare actual results against targets or benchmarks to help find areas of success or improvement. They can also use data visualization, such as progress charts or performance dashboards to show progress and highlight areas needing attention.

Descriptive analytics is usually the first step for businesses when looking into their historical data. It’s relatively simple to grasp, carry out, and communicate, making it an appealing and obvious (and often necessary) starting point.

However, its straightforward approach has limitations, especially regarding predicting future trends or providing recommendations for the identified problems.

It’s all very well finding issues or highlighting opportunities—but what’s the best way to address and achieve them? This is when predictive or prescriptive approaches are more useful, offering comprehensive data analysis when they are combined with descriptive analytics.

Before you take this route, let’s expand on the advantages and disadvantages of descriptive analytics in more detail.

Advantages

• Help businesses gain valuable insights into past events and trends, enabling them to understand historical performance.
• Provides a concise data summary, making complex information more accessible and understandable using charts, graphs, and other visualizations.
• Assists data exploration, enabling analysts to spot patterns, anomalies, and relationships, leading to deeper investigation.
• By understanding historical data patterns, decision-makers can make informed choices and strategies based on evidence rather than intuition.
• Provides metrics and KPIs to measure the success of initiatives and track progress toward goals.
• Fosters a data-driven culture, where businesses make decisions based on evidence and data rather than solely on personal experiences or gut feeling.

Disadvantages

• It only provides information about the past and doesn’t offer insights into future trends or events.
• Although it highlights patterns, it doesn’t give any actionable recommendations or solutions to address issues or capitalize on opportunities.
• The accuracy and reliability of descriptive analytics rely heavily on the quality and completeness of the data. Inaccurate or missing data can lead to misleading insights.
• Relying only on descriptive analytics could prevent organizations from adopting more advanced analytical techniques, which offer more proactive conclusions.
• It might not account for particular event timings or the changing nature of patterns, especially in dynamic environments.
• With lots of data being collected, there can be a risk of information overload and challenges in selecting the most relevant insights.

As we’ve seen, there’s great value to be had in looking at the past.

Amplitude Analytics gives you a clear, unobstructed view of your businesses’ past performance, and is an all-in-one data analytics platform that applies predictive, prescriptive, descriptive, and diagnostic analytics to help you garner insights across several areas of your business.

See what’s been happening without fuss or pageantry—just straightforward, valuable information that is the jumping-off point for deeper explorations.

Enhance your team's decision-making approach and level up their data game. Chat with Amplitude today to see how we can help.

Other Analytics Guides

• What is Predictive Analytics?
• What is Diagnostic Analytics?
• What is Prescriptive Analytics?