Data-driven storytelling

"The goal is to turn data into information, and information into insight." – Carly Fiorina

Data is the start of the story not the story itself

Data when used responsibly can be wielded to help make compelling stories. This is a skill I am actively working on and hope you gain something from me exploring it.

Having spent long enough in the finance industry I soon realised that numbers weren’t impartial at all. They're not just data points; they're weaved into stories, carefully selected and presented. Like many things, our trust in the numbers becomes a function of who presented them or whether we like the message.

Equally bad is a page of numbers with no explanation. The lack of direction makes us lose attention. The best piece of advice I have received on my presentations, is that a reader should be able to quickly figure out the answer to "So what?".

What are the issues with data's reputation?

• We don't explain the insight and why it is important, and what we should do about it

• Skewed incentives can lead to torture of data so it admits the right story!

What is data?

Data itself is single metric from many stories. If, for example, you are looking at the median level of education in the United Kingdom, there are millions of people underlying it. Some people might have been exceptional at music but struggled with history, or others were encouraged too much or not at all. Data cuts through all that, which can be helpful.

This can often work in reverse. We might hear that 50,000 people were killed, and it sounds awful. But we feel even more invested in the story, if we hear about Mustafa, a school-age child who was bombed whilst on the way to school. Our brain is not objective about these things. The personal story of a single Mustafa, may be more resonant than the loss of an abstract 50,000 people.

Why We Need Data (and What to Do When It's Missing)

We instinctively trust data because it feels reliable, concrete, and objective. Numbers guide our narratives, providing credibility and confidence. But what happens when data is incomplete or unavailable?

Our mind goes into overdrive in this situation. In these spaces for interpretation our feelings and bias can fully express itself. This is not unreasonable. After all, what else can we draw upon?

Trying to take action without data is harder and storytelling becomes even more critical. Experiences, anecdotes, and visions of the future are needed to help bridge the gap.

In trying to answer complex questions like 'should we go with a new company for this complex product', our minds tend to simplify with more accessible questions like 'would I look stupid if this went wrong?'. The value of data, is that it can help decision-makers justify unorthodox or bold action, to overcome the tendency in complexity to play it safe.

When storytelling is intentionally biased

I think storytelling is always and everywhere biased. The world is complex and we communicate in simple stories. There will always be details missed out.

It becomes more of a problem, where personal, professional or institutional incentives strongly favour a particular course of action. Knowingly or through not being able to consider the other side, data can be abused.

Data is so frequently misrepresented:

1. Cherry-Picking Timeframes – Selecting specific start and end dates to favour a particular narrative.

2. Selective Outcome Reporting – Highlighting only positive results while ignoring neutral or negative ones.

3. Averaging Away Extremes – Using mean values that hide volatility or outliers.

4. Misleading Visuals – Using inconsistent axis scales, truncated charts, or overly complex graphics to obscure reality.

5. Generalisation – Drawing broad conclusions from narrow or context-specific datasets.

6. Ignoring Sample Size – Presenting small-sample insights as robust findings.

7. Using Relative Percentages Without Base Rates – Quoting 200% increases without clarifying that it’s 2 out of 1.

8. Attributing Causation to Correlation – Implying cause-and-effect relationships where only correlation exists.

The list can go on, but a vast majority of the analysis contains one or more of these misrepresentations.

Data analytics is a sport with an ethical code

These practices do more harm than good. Investors become rightly skeptical. After all, if a deal looks too good, the question inevitably arises: "If this is so brilliant, why are you offering it to me?"

We see similar problems beyond finance, notably in medical journalism, where sensationalized headlines distort benign studies into alarming news, exploiting the same biases. How many times have people flip flopped between margarine and butter, or whether wine is good for you?

A practical approach to data and storytelling

1. Great compelling stories take you on a journey. A journey feels more real and compelling if every key step has a strong quantitative justification.

2. If possible, a killer chart for each argument made is great. Show not tell.

3. Data can be dull, and couched in jargon. Make sure your data is explained and relayed in language that your audience will be able to easily understand.

4. Numbers are good, visualisations are better! One day (been saying this for years) I am going to spend a month learning advanced data visualisation.

5. Even if you are selling, use data that shows risks. There is nothing less reassuring than being offered a "risk-free" investment. Showing risks builds trust.

6. Too much data is as dangerous as too little. Whilst there is temptation to show all your work, do you really need to show 5 charts making the same point? Better to footnote things you have left out, with "x, y and z charts tell a similar story".

7. How confident should this make us? Often historical data is presented as enough. What might differ in the future?

8. Where don't you have data? It is important to note gaps. It shows you have thought of how your argument could be even more compelling.

Who Does It Well? Examples of Ethical Data Storytellers

Not everyone succumbs to these temptations. Some storytellers consistently handle data responsibly, setting standards we can aspire to:

1. The gold standard is BBC's 'more or less' which debunks often-quoted stats BBC Radio 4 - More or Less

2. Andrew Huberman (Huberman Lab) and Peter Attia (Home - Peter Attia) in medicine.

3. Our World in Data: By presenting data transparently and thoroughly, it minimizes anchoring and availability biases, enabling audiences to evaluate information objectively.(Our World in Data)

So What?

1. Data alone doesn’t tell a story - and every dataset contains many stories. Statistics represent individual lives. We resonate more deeply with personal anecdotes than with abstract numbers.

2. Data feels solid – but in the absence of data, our minds simplify, bias creeps in, and storytelling becomes essential to bridge gaps.

3. A story without insight is forgettable – Readers need to know why they’re being shown data. Without purpose, attention wanes.

4. Incentives corrupt clean storytelling – Whether in marketing or media, data is often tortured to tell the ‘right’ story. Misrepresentation is rife, but there are some people who have excellent data ethics.

5. Clear structure helps ethical storytelling – Purpose, context, visuals, integrity, and curiosity form the backbone of data communication done right.

Next week I will be discussing "Stories as reminders". Until then, please sign up to receive the blog directly to your email at Blog | Deciders.