Presenting the story: Data visualisation missteps to avoid

Goitsemang Moagi, Data Engineer at PBT Group

Data visualisation keeps climbing the skills agenda, as this is how people in business see the story behind their work. But a chart is not a story on its own. If a dashboard overwhelms, hides the real issue, or uses the wrong language, users fall back to exporting into Excel to rebuild what they need. That is a failure of design and alignment.

Why visualisation matters now

Two shifts are raising the bar. First, teams expect actionable narratives, not just pictures of data. A useful view gives context, highlights the problem, and points to what can be done next. For example, if customer acquisition drops by 30%, the dashboard must make that clear and, where possible, highlight options for investigation or response. Second, AI inside modern tools accelerates delivery and lowers the barrier for business users to explore further, but it only works if the input data and framing are sound.

Common data visualisation mistakes

#1: Designing without knowing your audience

The fastest way to miss the mark is to build for a request, not for the people who will use it. Executives need a clear high-level view that flags what changed and why. The teams below them need to drill into the details. If the view is not adaptive, you will either overload one group or starve the other. Start with who they are, what decisions they must make, and how they will use the output after the meeting.

#2: No story, no context, no action

A neat layout without a narrative still leaves users guessing. Strong reporting sets the scene, surfaces the issue that jumps out, and suggests next steps. Increasingly, teams pair this with predictive or rules-based suggestions, so viewers do not just see a problem; they also see ideas worth testing.

3#: Treating definitions as an afterthought

When two teams use the same term differently, trust erodes. I am seeing more organisations include a compact glossary and a short “logic” page that explains how key measures are calculated and whether figures include things like VAT. This closes the integrity gap, improves credibility, and reduces rework.

#4: Using the wrong chart for the job

Good choices improve readability. For comparisons that are not time-based, a bar chart is usually the most straightforward way to see which product or region leads. For trends over time, line-based charts help readers follow movement and drill down to months or weeks. Pie charts have a place for showing makeup or contribution to a whole, provided the categories are few and the differences meaningful. Knowing these basics keeps the argument visible.

#5: Jargon that loses the room

Titles, labels, and annotations should use the language your audience uses every day. If the wording feels foreign, people disengage or misinterpret. Keep copy familiar and straightforward so attention stays on the signal, not on translation.

#6: Building at a distance from the people who need it

The best results come when report designers work closely with the people who consume the view, including data stewards and analysts in each department. They know the recurring questions and edge cases. That closeness replaces assumptions with quick feedback and helps the design evolve as new questions arise.

How AI helps, and where to be careful

AI speeds up the work. With clean inputs, assistants can suggest suitable chart types for time series or comparisons and let business users extend a view with their own follow-ups. That makes exploration faster. It however does not remove the need for good judgement about audience, context, and definitions.

A practical way to avoid the missteps

Begin every build with two questions: what is it for, and how will it be used. Design adaptive views that let people move from the headline to the detail. Include a brief glossary and logic notes. Choose chart types that match the question. Keep language plain. Work shoulder-to-shoulder with the teams who will rely on the output. Where it makes sense, add AI-assisted suggestions so that viewers see possible next steps alongside the problem.

When you get these basics right, visualisation stops being a picture of the past and becomes a guide to action. And that is the point.