Skip to content

Ten Kinds of Stories to Tell with Data

Time

1- Reporting: perform descriptive analytics. Tell what happened.

2- Explanatory survey: analyze what people or objects are up to. Ask people what they think about something. Conceive a statistical model; what factors drive others.

3- Prediction: perform predictive analytics. Use historical data, add a statistical model, probabilities, and assumptions, predict the future. Find out what customers are likely to buy. Assess how likely it is for an event to happen. Forecast economic conditions.

Focus

4- What: tell what happened with a focus on one issue (Reporting is not as focused).

5- Why: tell what underlying factors caused the outcome. Focus on the outcome.

6- How to address the issue: explore various ways to improve the situation. Focus on the situation.

Depth

7- CSI: run a small, ad hoc investigation. Find out why something is happening. Find out why some customer are dropping online transactions when they get to the postal code input form; what if some rural locations don’t have postal codes?

8- Eureka: invest in long, analytically-driven searches for a solution to a complex problem. Discover the right way to refer and price potential buyers to real estate agents on a website. Since the project is core to the company’s business model, it needs a corporate buy-in. The project involve several different analytical methods, false starts, dead-ends, discoveries.

Methods

9- Correlations: find why the relationships among variables rise and fall at the same time.

10- Causation: argue that one variable caused the other. Controlled experiment.

Afterword

  • These 10 approaches are not mutually exclusive.
  • Begin the report with the result and recommended outcome. Follow with the demonstration.
  • For the c-suite, keep technical terms to the minimum.