Storytelling for Data Analysts


    "Storytelling with data is critical. But the emphasis is on data, not story."
    -- Richard Hren, marketing strategist

Storytelling is a popular concept in data analysis circles these days. Thatís good! Sometimes, though, the story overshadows the data, and thatís bad.

What happens when data analysts present results in the same form that we get from data analysis tools? Not much. Decision makers donít care about that stuff. If we want action, itís up to us to speak managementís language, and managers appreciate a good story.

Stories get managementís attention. Stories help make your message understood, and encourage thoughtful discussion. But you must tell the right story, in the right way.

The heart of a good data story is the data itself, and the information that the data reveals. Some people tell stories that are engaging, believable, and persuasive, yet untrue. They hammer data into the story they want to tell, instead of crafting a story to fit the dataís message. Some of these stories bend facts, others are simply make-believe. Those are not good data stories.

A good data story must be true.

Whatís true? The sequence of events in your story must be realistic, in light of the available data and your analysis of it. The data comes first, the story second.

Good data stories also need the same things that any other story needs.

A good data story needs the right protagonist.

Letís be clear. Your data story must not be about you. Management doesn't care about you. Clients don't care about you. A good story is about a person that the listener can identify with. So a good protagonist might be your manager, a client, or some other person (real or hypothetical, such as a customer persona), that is important to your audience.

A good data story needs a challenge.

Interesting stories introduce conflict early. Whatís your hero or heroineís problem? High costs? Low revenue? Legal risk? What will happen if the challenge is not overcome?

A good data story needs a happy ending.

The ending may not be as happy as a fairy tale. Perhaps not all will live happily ever after. But the ending will suggest a course of action that makes the best of the situation. Sometimes that may be very happy indeed, as when a test program worked so well that you can recommend expansion and everyone will gain. Sometimes it will only be the best of a bad situation, as when the test goes so poorly that you must recommend pulling the plug to minimize loss.

Next time that you present, donít do a data dump. Instead, tell a true story. It will be about a hero very much like your manager or client, and tell the tale of facing a challenge and taking it on for the best possible result.

Do you tell stories with data? Share your experience here.

Meta S. Brown, Business Analytics Consultant

Meta S. Brown is a consultant, speaker, and writer who promotes the use of business analytics. A hands-on analyst who has tackled projects with up to $900 million at stake, she is a recognized expert in cutting-edge business analytics. She has conducted more than 4,000 hours of presentations about business analytics, and written guides on neural networks, quality improvement, statistical process control, and many other statistical methods. Meta's seminars have attracted thousands of attendees from across the US and Canada, from novices to professors.

Tell Me a Story: Why Data Analysts Must Be Storytellers, Too

Data alone won't make an analyst's work memorable or actionable in the eyes of a business executive. A story puts it into perspective.

It's the Data, Stupid

When it comes to acquiring the data that will feed your analytics initiative, "free" isn't always the best approach.


Re: Tapestry
  • 11/6/2013 12:50:19 PM
NO RATINGS

Looks suspiciously devoid of women.

Re: Tapestry
  • 11/6/2013 12:49:25 PM
NO RATINGS

You bet!

Re: Tapestry
  • 11/6/2013 12:46:49 PM
NO RATINGS

This is the first I've heard of it. Will investigate, thank you!

Tapestry
  • 11/6/2013 12:42:45 PM
NO RATINGS

Hi Meta, I just came across a mention of the Tapestry conference for storytelling with data, and thought of you. Have you ever attended? Looks like a lot of fun. (It came to my attention from fellow A2 blogger Jon Schwabish, a data visualization creator for the US government, and one of the event's planned storytellers.)

Re: Checks & balances?
  • 11/5/2013 3:20:19 PM
NO RATINGS

Narrative Science takes a set of structured facts - like the scores of a ball game, and presents them as narrative text.  So, you are on the mark, that a news provider whose intention is to provide original research, information that is new, would not get value from this or any other automated content generator. Even Narrative Science depends on some other source to provide the facts it uses.

There are other automatic content generators. These are used primarily to produce content that isn't intended for humans to read. That type of content is targeted toward automated readers - search engines. Although most of it is very poor quality, the best of the lot looks convincingly like it was written by a human being, but not necessarily an interesting human being.

So far, the biggest Narrative Science story is about the deal they didn't get, with the Washington Post.

Re: Checks & balances?
  • 11/5/2013 1:15:46 PM
NO RATINGS

@Meta, I can see Narrative Science being used to good effect in producing basic who-what-where-when-how types of news stories. But that's a far cry in my mind from the stories analysts would be using internally to deliver insight and intelligence to the business. Maybe it's just my perception as a writer, but I would think they'd want more control than what they could get by running their data through a story-writing algorithm!

Re: This Story Is About Me
  • 11/5/2013 9:40:35 AM
NO RATINGS

Oh, that reminds me of something.

I don't have much, if any, reason to tell data stories about myself. The usual thing is to tell a story that the listener percieves as being about himself or herself. But when the listener is a big shot, sometimes a story about a little person can be very compelling.

Marketers and product designers often use personas - fictitious profiles of intended users or buyers of a product. Personas are useful, but I haven't seen executives get very excited about them. However, I have seen presentations which were peppered with brief recordings of real customers or other "ordinary" people, talking about a concern or experience. I have seen some of these draw in executives and seriously engage them.

This Story Is About Me
  • 11/5/2013 9:01:28 AM
NO RATINGS

Story telling is how we get understanding. And usually the theme is some decision making that's not necessarily obvious at first glance. A perfect way to get movement from the decision makers is to tell a story. But, although the gneral guidance might be to not tell about oneself, there could be exceptions. Say the story teller is widely known or respected in the particular community receiving the presentation. Not only would the audience like to hear more about the storyteller in this case, his perceived high ranking would perhaps lend credence to the story. 

Re: Checks & balances?
  • 11/5/2013 7:58:14 AM
NO RATINGS

Ariella and Beth,

It's interesting that Kris Hammond made a prediction that 90% of news will be computer generated in 15 years, and that he did so publicly. The public statements I have seen from Narrative Science did not make any claims like that.

I have seen some samples of news generated by Narrative Science, and they were pretty good, so I do not doubt that we will see computer generated news in some places. But Mr. Hammond's forecast seems to be based on nothing but his own impressions.

The examples that I recall were sports reports. And they were written. Sports is not really my bag, so I find myself thinking about the sports fans I know and how they get their sports news. Personality is important to them. They watch the TV sportscaster they prefer, or listen to a favorite radio announcer.

Perhaps they would check any handy written source for quick information about a game, but if so, I wonder if they might not prefer just the facts - the scores as numbers, rather than a machine generated narrative. If they are in a real hurry, they might not even want a narrative written by a human.

If they have the time and motivation to read, I would still expect personality to matter. Even if professional sports writers went away, there will be amateur sports writers all over the web, and discussion groups populated by fans.

I'm sure that Narrative Science will do well. There is a market for cheaply generated content, especially online. There are others who provide machine generated content, but theirs seems to be of better quality.

As for Mr.Hammond's forecast that 90% of news will be machine generated in 15 years, that's just talk. Let's see some real evidence. That prediction was posted on Wednesday, December 7, 2011, almost two years ago, long enough to have some numbers to show how much traction Narrative Science and its competitors have made. I'd like to see those numbers.

Re: Don't anticipate your story before the analysis is over
  • 11/5/2013 7:24:17 AM
NO RATINGS

Musa,

You've made an important point.

Nate Silver wrote about this in his book, "The Signal and the Noise". He explains how the best forecasters gather all the information available, and do not discount sources because they suggest an outcome that the forecaster doesn't like. It's well worth reading.

 

Page 1 / 3   >   >>
INFORMATION RESOURCES
ANALYTICS IN ACTION
CARTERTOONS
VIEW ALL +
QUICK POLL
VIEW ALL +