Structuring Information Before You Build Slides

Jeff Bezos was right to ban PowerPoint presentations.

To be precise, not across the whole company: in 2004 he changed the format of senior leadership meetings. Six pages of plain prose replaced the slides, and the meeting began with everyone reading that text in silence. But his diagnosis was spot on.

The real objection is not to the slide but to what it hides

A presentation that fills a template without knowing what it wants to say, and hides behind the slide, does not improve the quality of understanding or thinking. Employees prepare for days for a presentation; what comes out is a clumsy, aimless meeting with no outcome. Their own time and the company's energy are both wasted.

That was Bezos's reasoning too: writing a good memo is harder than "writing" a long presentation, because narrative structure forces the writer to clarify what matters more than what, and how the ideas connect. A bulleted deck lets you skip over all of that.

The same diagnosis, from the data side

A year earlier, in 2003, Edward Tufte reached the same point through another door. He had two objections: the information density per unit of slide area is far too low, and the nested bullet hierarchy lines ideas up without showing the logical relationship between them.

Tufte made it concrete with the Columbia shuttle disaster: a critical engineering finding had been buried under nested bullets and lost its significance.

Two names, two separate disciplines, one conclusion: the problem is not the tool itself, but that it allows poor thinking to hide.

Two questions that must be answered

What matters?

How do the pieces of information I am presenting relate to each other?

Both must be answered before you start building the presentation. Only by clarifying them can we focus stakeholders on the facts and the solution. Slide design comes after that, not before.

Example: the same content, two different structures

Say you are presenting the quarterly review of a supply-chain project. Your bullet points:

• 3 new suppliers added in Q1
• Average integration time 6 weeks
• Delays at 2 suppliers
• Costs 8% over target
• 4 more suppliers planned for Q2

The list is accurate. But it says nothing. All five items carry the same visual weight; it is unclear which one requires a decision and which is merely information.

Now let's ask the two questions.

What matters? The cost overrun, because it is the only item that requires a decision.

What is the relationship? The other four items are not independent facts but a chain: integration takes long, so delays occur, so costs overrun. The Q2 plan will produce the same result if that chain is not fixed.

Once the structure emerges, the slide builds itself: a process flow with the bottleneck marked, ending in the cost impact. The same five facts, but now an argument.

What changed was not the design. What changed was the presenter's understanding of the subject.

So are slides unnecessary?

No. Bezos doesn't say that either; Amazon still uses plenty of slides today.

The distinction is this: in a decision meeting, a well-written memo is often the better tool. Slides win when you are presenting to a room, when the message is comparative, spatial or quantitative, and when the document will circulate without you.

What's wrong is not using slides; it's using slides without thinking.

Conclusion

The answer is not filling templates and decorating slides, but structuring and visualising information.

That is how a real presentation is made.