Data Storytelling: In absence of an action, give direction

Data analysts often find themselves in a situation where they don't know what the insights behind the data are. Here is a simple way to handle such a situation. The application of these strategies is not limited to data analysts but to anyone who is wants to be good at data storytelling.

Last week I conducted a workshops on Data Storytelling in UAE and Malaysia. During the workshop I got the participants to apply the learnings from the workshop on a data slide that they will use at work. Here are some things I observed during the exercise.

Imagine you are a data analyst and you are presenting the data given below, which shows the volume of tickets received versus volume of tickets processed in a year.

[caption id="attachment_7438" align="aligncenter" width="400"]

Image Source : Storytelling with Data By Cole Nussbaumer Knaflic[/caption]

Many participants when shown this graph said , " In the later half of the year we processed less tickets than we received." The moment you hear something like this, as an audience, what goes through your mind is, " Well, that I can see!" That is an observation not an insight. And sadly, this observation can be made by anyone who has access to the graph. It is an obvious observation!

So what exactly is the purpose of you standing up and speaking?

As a data storyteller you should always want your audience to know or do something. If you cannot articulate that, then you should revisit whether you need to communicate in the first place at all. When I brought this up and questioned my workshop participants as to why were they not sharing insights and suggesting actions, the answer I got was, "The audience know better and should be able to decide on how to act."

With that kind of a thinking, we have just announced that data analysts are far from understanding why things happen the way they do and even further away from an ability to suggest actions. Their roles are limited to being number crunchers with little insight in to actual business activity. Not a desired outcome is it? But the participants were right that they did not have the domain knowledge to be able to figure out what happened? This reminded me of AI experts with no domain knowledge.

So, with such a limitation what can you do? Step 1: Accept you are not the keeper of all the information but that doesn't mean you are just a number cruncher.

Firstly, accept that you are not the keeper of all the information and you may need to get more information from the business before you can come to a conclusion. Data analysts tend to put themselves under an unnecessary pressure that they must have all the information. And when they don't have the information, they tend to tell themselves a story that the business knows what to do. It is their domain.

Step 2 : Shift possible causes to probable causes

To add value, we have to analyse the findings and come up with some possible causes. Then reduce the list from possible to probable causes

Here are some ways to achieve that,

Ask the Questions. The New York Times published a great article, The Power of ‘Why?' and ‘What if ‘? which highlights the importance of asking questions. Feedback from senior executives reflects that they need employees and leadership to ask more questions. This is true especially when investigating facts with data.

In this specific data presentation if you are not aware what caused the difference between number of tickets received to number of tickets processed. Here are some questions you could ask

• Were there any resignations and were we facing a manpower crunch?
• Was our ticketing system having any troubles that led to slowing down of the process?
• Did we have new system that the team was still learning how to use?
• As for historical data. For example, did we experience something similar exactly at the same time last year?
• Ask for more data points required to find out what is happening. For example, you may ask data on background information on the staff. Did the staff started working remotely? To further investigate.

It is totally fine that you do not know the answer to why things happened the way they did. If you simply said what was obvious on the slide, you are really not needed. But if you asked the questions you can lead the group in to some productive conversations which will reveal the insights that can be used to drive business outcomes. And even if you are highlight the wrong thing, it prompts the right sort of conversation.

Your role sometimes is to trigger a conversation to find an insight not to provide a solution. And that you can do by being a Questionologist. Questinlogist is someone who asks good questions.

Step 3: Give the data a direction

Be comfortable with sharing a point of view based on the best information available to you and giving the data a direction. Be descriptive, share what you see and what you don't.

Your data storytelling is not compromised because you did not give an action. You can give a direction which will eventually lead to an action.

Today our roles are super specialised and many experts come together to get to the same goal. For a data analyst, the goal is find the insight for better business outcomes and the data analyst alone can't achieve that goal in a complex business environment

In absence of being able to explain a cause, , come up with possibilities, ask questions to shift possibilities to probabilities, ask for more information that might assist, describe what you find and what you don't. Give data a direction which will make you a trusted advisor not just a number cruncher.

Hope you found this insight on storytelling helpful and if you did, please share the article and join Storied Book Club by clicking here. If you are already a member of the book club, then Stay Storied!