I write fiction for fun. Short stories, mostly. It is the opposite of my day job, which is helping enterprises make sense of their data. Except it is not the opposite at all.
The Throughline Problem
When I sit with a draft that is not landing, the first question is always: what is this actually about? A story can have twelve characters and six subplots, but if there is no throughline, the reader puts it down.
Data presentations break the same way. A CEO gets a chart on user acquisition, another on server latency, a third on Q3 revenue. Three slides, three metrics, no connection.
I worked with a logistics client where the operations team had built forty-plus visualisations. The COO looked at it once and never came back. When we rebuilt the entire view around three questions he actually needed answered on Mondays, he started checking it before his standup. The data was never the problem. The structure was.
First Drafts Are Not for Your Audience
The first draft is for you. The final draft is for the reader. Data analysts do this constantly. Exploratory analysis is the first draft. It is necessary and private. The moment you put that exploration in front of a stakeholder, you have handed them your working notes.
The shift is from exploratory to explanatory. Analysts are trained to show their work. In school, showing your work got you marks. In a boardroom, showing your work gets you ignored.
Editing Is the Job
A scene that is well-written but does not advance the story gets deleted. Dashboard design needs the same ruthlessness. We suffer from metric bloat: every KPI jammed onto one screen because the data exists.
One of our Qlik implementations for a manufacturing client started with 90+ metrics across eight tabs. The final version had 14 metrics across two views. The plant manager told us it was the first dashboard he actually understood without someone walking him through it.
Make Them Do Something
A story wants the reader to feel something. A data narrative wants the stakeholder to do something. If the audience leaves the room without knowing what to do next, the presentation did not work.
Every data story I build now ends the same way: here is what we found, here is what it means, here is what we recommend doing on Monday morning. Not "consider the implications." A specific action tied to a specific finding.