How I do Data Storytelling

During my last 5 years of working as a Business Intelligence Analyst, I have received many positive feedbacks on my data storytelling skills from both internal and external audiences.

Therefore, with this very first blog post, I would like to share four tips that I found really helpful in conducting a good data storytelling. They are:

• Know your audiences.
• Develop your story together with your audiences.
• Know your data through and through.
• Create a visually affordable presentation.

Any feedback will be appreciated!

Happy reading.

Know your audiences.

The first and the foremost principle is to have a good profiling of your audiences. Ask yourself the following questions and make sure you get the answers before you start to work on your story:

• What roles do they have in the organisation? Are they developers, marketing experts or sales agents?
• How familiar are they about the topics today? Are they subject matter experts or new to the topics?
• What brings them to this presentation? Are they looking for some good-to-know fun facts or key information to support their decisions?

Knowing your audiences is the key because this will determine what you should present and how your should present it. In other words, this will help you to set up a stage for them and bring them into your story.

For example, I used to run an onboarding session for all new joiners every month. This session covers not only an intro to our team, but also some basic numbers and fun facts about the business. These new joiners come from different industries so they have various levels of data knowledge—some of them are database experts and some of them don’t even use Excel. They are all new to the business. They come here because this is a part of their onboarding journey and they are supposed to gain some basic knowledge about the business. The team intro is not difficult, but the data part can be a bit boring to some audiences.

Hence, I develop my storyline like this: “… now imagine this Christmas (or Easter or summer party depending on the timing 😆) you meet some friends & family and they know you now have a new job, they might ask questions like “what kind of business is it?”, “how big is the business?” or “what kind of customers are you serving?”. I would like to help you prepare for such questions, and here is an overview: our business has an annual revenue of XXX and a customer base of XXX, our customers have an average age of XXX, etc. When it comes to the top objectives of the business, instead of simply listing the KPIs, I would say something like “…as an employee I am concerned whether I will receive my pay check every month so I am interested in learning how much money comes in every month and that is why we look at revenue…”

By doing so I could feel that my audiences start to stay focused on my presentation and ask questions proactively. Today I still remember one feedback from a new joiner: “that is the best prepared onboarding session I’ve ever had!”

Develop your story together with your audiences.

I first learnt about this method from the great book “Pyramid Principle”. With this method, when I develop a story I constantly shift my perspectives between a speaker and an audience, ask myself questions that may come from and answer them on my own. I then repeat this process over and over until I feel like there is no more question I could ask.

Therefore, my storyline is connected not by the bullet points that I plan to talk about, but a series of Q&As I run with my (imaginary) audiences.

For example, instead of “First, I want to talk about …, then I want to talk about …, and then …”, my plot looks like the following:

• First, Alice wants to have an overview of our sales performance this year, so I will begin with presenting our monthly sales revenue and also show a YoY trend.
• Then Alice will notice that this September has a negative YoY growth and wonder why that is the case, so I need to zoom in here for a dissection by breaking down that total sales revenue into # of Orders * € Revenue per Order and compare these two components against how they look like last September.
• Then Alice will see that the number of orders is actually higher than last September yet the order value drops — that is an alarm! Now I need to look at the order value across categories/demographics to show where this drop stems from.
• …

This change of perspectives helps me to not only develop a story line that caters to the interests or questions of my audiences, but also make my verbal communication more natural. When I present some insights, I don’t see it as a one-way reporting in a way that “I am obliged to say something in front of a group of people because I find something I want to talk about”, but to replay a two-way conversation that I have had multiple times with my (imaginary) audience(s) — I know Alice has some questions for which she wants to know the answers and I am ready to walk them through together with her.

One fun fact: it happened to me very often that during my presentation, some colleague asks me a follow-up question and my answer to that is usually: “Well, that is what we will find out right on the next page!” 😃

Know your data through and through.

Know every single detail of the data behind your story and be prepared for questions.

This means you are expected to know not only the definition of metrics (e.g., the number of new customers) and dimensions (e.g., marketing source), but also how the data has been collected and transformed (e.g., what has been filtered out), so that you are well prepared for questions like “Does this chart also include XX channel?” or “Do we only count website visits here?”.

This also means that you should know not only the contents that are present in the slides or sheets, but also those that are not. In the last example with Alice, we may stop on the slides with finding out that some categories have a lower order value this September due to a promotional campaign. However, the real underlying story plot should always go at least one step further: What is the impact of this campaign on the annual revenue? Are there any implications other than revenue? Does the team plan to do this promotion again?

Why is it important to know your data well and be prepared for all questions that may come? It is about establishing your credibility especially when you present to a group of senior audiences who are looking for information to support their key decisions. Personally when I see a presentation, ask how something is calculated and hear uncertain answers like “I don’t know” or “Maybe something is off”, an internal alert is triggered and I start to question silently the validity of the numbers that I see, as I am not sure if the speaker knows what he/she is talking about. 🤔

It is perfectly fine to say “I don’t know” when you don’t know the answer to a question about your data! However, it would be a little bit sloppy if you just stop there. Why? Because now the audiences face not only the uncertainty about what you already presented, but also a new uncertainty about whether they will have the first uncertainty resolved. An expected answer in this scenario would be “I don’t know at the moment but I will do more research and get back to you by XXX (a specific date & time).” This answer is more welcome because it conveys to the audiences both your authenticity (i.e., you do not pretend to know something you actually don’t know) and your accountability (i.e., you know how to fix something and you commit to fixing it).

Create a visually affordable presentation.

Most of the time, your data storytelling is accompanied with some slides. The slides should be a great complement to your verbal communication rather than distracting your audiences from it. To put it in this way: imagine you are a tour guide and you presentation is a map. If the map is too complex to understand, your tourists will need to spend a lot of time on reading the map rather than paying attention to your fantastic introduction — or even worse, they may lose their interests in this trip at all.

Therefore, it is important to design your presentation in a manner to minimise the effort that it takes your audiences to extract information from it — this is why I call it “visually affordable”.

Here are some tips that I find pretty helpful (I will probably write another post about this with some concrete examples!) to achieve that goal:

• Always put conclusions first and let details follow — the most important lesson I learnt from the Pyramid Principle!
• Keep your visualisation as simple as possible. Avoid a fancy sunburst chart when a pie chart can do the job.
• Make layouts and colours you best allies and let them help you form a pattern for your audiences to follow.
• Use meaningful titles such as “Sales outlook is promising in the next 12 months” rather than simply “Sales outlook”.
• Let consistency rule every corner of your slides. If you use “July 2020” for a month label, then stick to it in your presentation instead of switching to “2020/07” or “2020–07–01”.
• Use number formats wisely. If you are showing a number series ranging from 2,000 to 200,000, use data labels “3K” instead of “3,543”.

That’s it for today.

Thanks for your time!