How to Communicate The Story in your Data

Once you have worked out what the data means, the next step is to package it up in a way that the audience can quickly grasp its meaning and implications. This is a challenging thing to do. There are three steps:

1. Build recommendation pyramids
2. Order the narrative
3. Use techniques to propel the story forward.

Why data storytelling is challenging

Traditional storytelling is about entertainment.  Making people laugh, cry, and perhaps scream. Commercial storytelling primarily involves sales pitches or communicating a big idea. Data stories are different and very challenging.

McKinsey & Company has an expression for a bad data story: the anxious parade of knowledge. Slide after slide of data is presented. The client is perhaps interested, but more often than not bored. The presenter hopes that the client will find something interesting.

A number of things conspire to make data storytelling a challenge. The first is that you can't just create a story. You have to find it in the data. Even once it is found, it is often confusing, ambiguous, and weak. And, the whole point is that the data is going to be used as an input to decision making and planning.  This leads to an important implication: in a data story, there is a need to present the data in a way that the audience can stress test it. They need to be able to verify that the data support any conclusions and recommendations. 

Another complication is that data stories often need to work as standalone documents so that somebody who was not at the presentation can still understand them and stress test the facts. 

The last part of the challenge is that there is often a need for the data presentation to be re-purposed, cut up, and added to other documents (e.g., strategic plans).

Step 1: Build recommendation pyramids

The first step in working out how to communicate the story in your data is to create a recommendation pyramid, which indicates the relationships between all the data and how they lead to conclusions and recommendations. An example is below.

For more information, see Recommendation Pyramids for Data Storytelling.

Step 2: Order the narrative

Once we have created the recommendation pyramid, we need to pull it apart and create a linear narrative (also known as a story arc). 

The basic way to do this is to start from the top of the story pyramid, put the recommendation at the beginning, and gradually reveal the supporting detail. It is very important to first lead with the recommendation rather than to save it to the end, as the audience can only meaningly stress test data and conclusions if they understand what recommendations they are being used to support. The way this is done is that the presentation starts with the recommendations and logic that supports those recommendations, and then reviews all the underlying data, giving the reader/viewer the opportunity to stress test facts with full knowledge of how they are being used. The structure for the margarine example is shown below.

Step 3: Propel the story forward

A challenge with the recommendations-first structure is that it kills the suspense. The audience no longer has to wait until the end to see how the story ends. While this is ideal from the perspective of ensuring that the data story is useful, it does mean that we have to find other ways of propelling the story forward.

Specific techniques for propelling the story forward are:

• Have interesting facts with a good structure.
• Create pages with a pyramid structure.
• Visual design.
• Metaphors and analogies.
• Emotion.
• Find specific, concrete, relevant, detail and zoom in on it.
• Show and emphasize the story gap.
• Force the audience to participate.

For more information, see Techniques for Propelling a Story Forward.

Next

Recommendation Pyramids for Data Storytelling.