Telling the Story of Your Data and Its Science

Image for post
Image for post

As CEO of an AI company I always have to tell stories based on our findings to our clients, and take strategic decisions for our company and its teams. I am not perfect, but here are a five things I have learnt the hard way:

1. Start your project with the story, not vice versa

This is the only thing that matters — do not start on your data science project unless you know what business questions you are supposed to answer. This is where most of us fail — most sophisticated data analysis is useful only if it was commissioned with a clear purpose or to answer a clear set of questions.

2. Say as little as possible, say it in English

If you followed #1, by the time you complete your analysis, you have answers to all the initial business questions or you know that the data cannot find these answers. Just put these answers (or I-don’t-knows) into a simple narrative with no word that is not English (or your language). There are many resources on the internet that can help you compose this. Do not include data science terms in the narrative; and do not say anything unless you have to.

3. Add insights to the original story

You can take liberty and add to the story some insights that — a) explain some point in the original story; b) challenge one of your/ your business colleagues’ assumptions so far; or c) are super-cool fun-facts to break the ice or end your meeting at a high note. That’s it. Nobody is interested in anything else. Even for these, say as little as possible and say it in English or your business language.

4. Handicap everything with your confidence in the insight

It is very important to be transparent about how confident you are in the insights that you have just reported based on your analysis. Early on, find a metric that you will use to judge this for yourself, and report this metric. You don’t have to say things like R = 0.83, because then you will have to talk about what is R, why is it important etc. However, do compute such metrics and scale them into High, Medium and Low confidence at the minimum.

5. Decide on next steps with business colleagues, not data teams

Once you are done narrating a story, it’s time to move on to the next one. Again, start with the story. Think of the next story you would like to tell, and then agree with your business colleagues on that — it may turn out that the story you want to tell is not important to them at all. As soon as you and your business colleagues agree on the next story, you are back at step 1.

Hope this helps.

Written by

Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store