Sunday, 15 December 2013

Your data story

I recently participated in a trade mission with UKTI - think David Cameron in China, but instead it was Baroness Bohnan Carter in Mexcio. And as part of this GREAT Week, I gave a presentation on how to use data to tell stories. The presentation was to a room full of journalists, social scientists and programmers, who had been brought together, given a bunch of Government data and challenged to visualise it in 24hrs.

It was super exciting for many of the Mexicans in the room (and me!) because this was the first time certain Government departments were making their data freely available and it was good stuff! Data from the Mexico City bike scheme (their equivalent to Boris Bikes); stats from the Government helpline call centre; data from hospitals and air pollution levels.


Before they split into teams and got stuck-in to the data, my job was to provide some context for the day. I started by explaining that making the data available is only the first part. People need help to:

  • Understand why they should trust the data, what is the motivation behind making it public? Who owns the data? Where did it come from? How was it gathered?

  • Interpret what it means to them. How can they use it in their every day life? So we need to make it resonate, by making it relevant.



People don't connect to data, people connect to stories. You need to find the story in your data. Kantar, the company I work for, is the data investment management division of WPP and we have a staggering amount of data. When I first joined, one of my challenges was to get my head round the data we have and I struggled. And the more I learnt about the company, the more I realised it was a struggle that all the spreadsheets in the world weren't going to solve! What helped me was a simple story about how one of our clients used our data to develop their product and ultimately grow their business. And it's these stories that have been helping me ever since.

It comes back to journalism (doesn't it always with me!) 

  • What? (data)

  • Where? (context)

  • And why (why this is happening, so you can learn from it).



When I was a News Editor one of the important lessons I would have to drill into all new journalists was the problem of burying the lead; this can happen surprisingly often when you become embroiled in a story. 

Authors Chip and Dan Heath provide a framework for creating a story in their book 'made to stick' and I find it to be a useful checklist when working with data:



Simple - Data is complicated, but it is your job to make it simple and bring it to life.

Unexpected - get people to pay attention
Concrete - help people understand and remember
Credible - so they can believe and agree
Emotional - so they care
Story - because it spells success?! and because a good story will encourage people to act


So after stealing shamelessly from the Heath brothers, I felt I needed to present the Mexican audience with at least one idea of my own! So I talked them through my five-step process for telling a story from data:
  1. I spend time looking at and playing with data, creating hypothesis and seeing whether they are supported by the data. Is it different for men v women? Do older people respond differently to younger people? Is there a problem with X or Y? I've become a dab-hand at pivot tables!

  2. Pull out the key points, accept that at this stage there may be a dozen, or more, different little nuggets of information. Collate them all and worry about whittling them down to a more succinct story later.

  3. Bring the data to life by interviewing experts and asking their opinion on the most interesting data points you've pulled out. This is your opportunity to sense check the data.

  4. Use the data and these interviews to create a narrative, following the SUCCESs framework.

  5. Then take that story and tell it in a dozen different ways to get the maximum impact and traction: data visualisation, charts, snackable content, short articles, longer pieces, white papers, social media posts, video et al.

I talked again about the 'scrappy or epic' way of working that I learnt at Google. This is something I talked about in my blog post on measurement and it is just as applicable when talking about data stories. You can use data in a scrappy way to tell quick stories, or you can use epic data to tell vast game-changing stories, but if you find your data tells a rather mundane dull story (which you will find on occasion) then my advice would be to stop and do something else instead.

Recommended reading:


'Made to Stick: Why some ideas take hold and others come unstuck'