Learning Objectives
By the end of this lesson, learners should be able to:
- Define business analytics and distinguish major analytical approaches.
- Explain the purpose of descriptive analytics.
- Explain the purpose of diagnostic analytics.
- Distinguish correlation from causation.
- Identify appropriate analytical questions for descriptive and diagnostic analysis.
- Explain how descriptive and diagnostic analytics support business decision-making.
1. Meaning of Business Analytics
Business analytics is the systematic use of data, statistical methods, analytical techniques and technology to generate insights that support organizational decision-making.
Business analytics can help organizations:
- Understand performance.
- Identify patterns.
- Investigate problems.
- Forecast future outcomes.
- Evaluate alternatives.
- Improve resource allocation.
The value of analytics does not come simply from calculating statistics. It comes from connecting evidence to meaningful business questions and decisions.
2. Major Types of Analytics
Business analytics is commonly divided into four broad categories:
Descriptive Analytics
Answers:
What happened?
Diagnostic Analytics
Answers:
Why did it happen?
Predictive Analytics
Answers:
What is likely to happen?
Prescriptive Analytics
Answers:
What should be done?
These categories are related but serve different purposes.
3. Descriptive Analytics
Descriptive analytics summarizes historical or current data to explain observed performance.
Examples include:
- Total revenue for a quarter.
- Average customer order value.
- Monthly employee turnover.
- Number of products sold.
- Percentage change in operating costs.
Descriptive analytics does not necessarily explain why an outcome occurred.
4. Diagnostic Analytics
Diagnostic analytics investigates relationships, patterns and contributing factors to understand why an observed outcome occurred.
Questions may include:
- Why did revenue decline?
- Why did operating costs increase?
- Why did customer cancellations rise?
- Why did productivity vary between periods?
Diagnostic analysis may involve:
- Drill-down analysis.
- Comparisons.
- Correlation analysis.
- Segmentation.
- Variance analysis.
- Root-cause investigation.
5. Descriptive Versus Diagnostic Analytics
Consider an organization that reports:
Revenue decreased by 12% during the quarter.
This is descriptive analytics.
Further analysis discovers that:
- One major product category declined substantially.
- The decline was concentrated in two distribution channels.
- The affected channels experienced increased customer cancellations.
This moves toward diagnostic analytics.
The distinction is therefore not simply about the complexity of the analysis. It concerns the question being answered.
6. Business Questions and Analytical Questions
A strong analyst begins with the business problem rather than the available technology.
For example:
Business question:
Why has customer retention deteriorated?
Possible analytical questions include:
- When did retention begin to decline?
- Which customer segments experienced the largest change?
- Are changes concentrated in particular products?
- Is the decline associated with service issues?
- Did pricing or competitor activity coincide with the decline?
A clear analytical question helps determine which data and methods are appropriate.
7. Data Aggregation
Aggregation combines individual observations into meaningful summaries.
For example:
Individual transactions can be aggregated into:
- Daily sales.
- Weekly sales.
- Monthly sales.
- Product-level sales.
- Regional sales.
Aggregation simplifies complex datasets but can also hide important variation.
8. Drill-Down Analysis
Drill-down analysis moves from a high-level summary to increasingly detailed information.
For example:
Total revenue → Business unit → Product category → Product → Individual transaction
This can help identify where a change originated.
However, analysts should avoid assuming that the lowest-level explanation is automatically the most important explanation.
9. Segmentation
Segmentation divides data into meaningful groups.
Examples include:
- Customer segments.
- Product categories.
- Geographic markets.
- Distribution channels.
- Time periods.
Segmentation can reveal patterns hidden in overall averages.
An organization may have stable overall performance while one important segment is deteriorating significantly.
10. Correlation and Causation
Correlation describes the degree to which variables move together.
However:
Correlation does not by itself establish causation.
Two variables may move together because:
- One causes the other.
- Both are influenced by another factor.
- The relationship is coincidental.
- The relationship changes under different conditions.
Executives should therefore avoid treating statistical association as proof of cause and effect.
11. Diagnostic Reasoning
Diagnostic analytics should follow a structured process.
Step 1: Identify the observed outcome
What changed?
Step 2: Establish the magnitude
How large is the change?
Step 3: Establish the timing
When did it occur?
Step 4: Segment the problem
Where is it concentrated?
Step 5: Identify potential contributing factors
What variables changed at approximately the same time?
Step 6: Test competing explanations
Which explanation is most consistent with the available evidence?
12. The Importance of Context
Numbers without context can produce misleading conclusions.
Suppose sales increase by 20%.
This may appear positive, but management should ask:
- Did prices increase?
- Did sales volume increase?
- Did costs increase faster?
- Was the increase temporary?
- Did competitors experience similar changes?
- Did profitability improve?
Analytics should therefore interpret numerical results within their business context.
13. Limitations of Descriptive and Diagnostic Analytics
These analytical approaches have limitations.
They may:
- Depend on historical data.
- Reveal associations without proving causality.
- Be affected by data-quality problems.
- Oversimplify complex organizational systems.
- Miss unobserved variables.
Good analytical practice requires recognizing these limitations.
14. Role in Executive Decision-Making
Descriptive and diagnostic analytics help executives:
- Monitor organizational performance.
- Detect deviations.
- Identify emerging problems.
- Investigate performance gaps.
- Challenge assumptions.
- Direct management attention.
They provide the foundation for more advanced predictive and prescriptive analysis.
Lesson Summary
Descriptive analytics answers what happened, while diagnostic analytics investigates why it happened.
Descriptive methods include:
- Aggregation.
- Summarization.
- Segmentation.
- Performance reporting.
Diagnostic methods may include:
- Drill-down analysis.
- Variance analysis.
- Correlation analysis.
- Comparative analysis.
- Root-cause investigation.
Executives should distinguish association from causation and interpret analytical findings within their broader organizational context.
References
- IBM — Business Analytics
IBM Analytics - Microsoft — Power BI and Business Intelligence
Microsoft Power BI - DAMA International — DAMA-DMBOK
DAMA International - NIST — Data and AI Resources
NIST
Review Questions
- What is business analytics?
- How does descriptive analytics differ from diagnostic analytics?
- What questions does predictive analytics address?
- What is data aggregation?
- Why can aggregation hide important patterns?
- What is drill-down analysis?
- Why is segmentation useful?
- What is the difference between correlation and causation?
- What steps can an analyst follow when investigating a performance decline?
- Why is context important when interpreting analytical results?
- What limitations should executives consider when using historical analytics?
- How do descriptive and diagnostic analytics support executive decision-making?