Where pitch decks lose the room

Fibvo editorial · 10 July 2026 · 4 min read

Fifteen minutes into the meeting, someone asks what you are actually asking for, and the answer is on slide nineteen. Someone else is squinting at a chart with eleven lines and a legend in eight-point type. A third person, never in the room, gets the file forwarded that evening and cannot follow it without you narrating. The argument was sound. The deck was built in the wrong order, for the wrong reader, and lost a decision it should have won.

The ask belongs near the front

Most decks are built like a case in court: context, evidence, argument, then the request. That order fails in a meeting, because attention is highest in the first few minutes and the agenda can move without warning. An ask that arrives last often does not arrive.

Put the ask on the second or third slide, in the terms the decision needs, then spend the rest of the deck earning it. A reader who knows what you want tests the evidence against the ask, and raises objections while you are still in the room to answer them.

• The decision you want, in one sentence a reader could repeat to somebody else.
• The commitment attached to it: budget, headcount, access or scope.
• The date it needs to be taken by, and what slips if it is not.
• The person or committee who has to say yes.

One slide, one job

The second fault is a slide doing four jobs at once: context, data, a conclusion and a question, with nowhere for the eye to land. Everything on it is legitimate. It belongs on different slides.

The test is mechanical. Write the single sentence the slide exists to prove. If that sentence needs an "and", it is two slides. Splitting costs nothing: nobody leaves a meeting complaining about slide count.

Then make that sentence the headline. Not "Q3 pipeline" but what you would say out loud about the pipeline: which segments moved, which did not, and what follows. A reader who skims only the headlines should still get the argument, which is what lets a deck survive being forwarded.

Charts nobody has to decode

An unreadable chart moves work to the audience at the moment you need them thinking about the argument, not the axis labels. Most of the repair is subtraction.

• One comparison per chart. If it supports two conclusions, draw two charts.
• Label each series on the line or bar itself, so the eye does not travel to a legend and back.
• Sort categories by value, unless time or an agreed order says otherwise.
• State the units and the period on the chart, not out loud.
• Colour the one bar or line that carries the point and grey the rest.
• Check it at the size it will be seen: from the back of a room, and on a phone.

Two devices are worth dropping. A second vertical axis invites a comparison the data may not support, and the story it tells changes with the scale you choose. Anything three-dimensional distorts the lengths and areas a reader is trying to compare.

Designed for the presenter, not the reader

Three words and a photograph make a fine prompt for whoever is speaking. As a document, that slide says nothing: the argument stayed in the presenter's head. And the deck travels without you. It is opened before the meeting by the person deciding whether to attend, forwarded to a finance lead who was never invited, and read again weeks later when the decision comes back.

A deck that only works with you narrating it will be judged on the day it is read without you.

The remedy is not paragraphs on every slide. It is a conclusion in every headline, the evidence beside it, and detail only some readers want moved to an appendix. Speaker notes are for what you will say, not for reasoning a reader needs. In our deck work the readable version is drafted first: a slide that survives being read cold will also survive being presented.

A deck to present and a deck to send

These are two documents carrying one argument. One file doing both jobs ends up too dense to project and too thin to read alone. Build the argument once, then produce both versions.

• To present: one idea per slide, type large enough to read from the back, charts cut to one comparison, nothing essential in the notes.
• To send: the same slides, each with a written conclusion under the chart, the appendix attached, and a closing slide naming next steps, an owner and a date.
• Send a PDF with the fonts embedded, so the layout holds on someone else's machine.
• Name the file so it can be found in a mailbox six weeks later: organisation, subject, version, date.
• If one file must do both jobs, build the version you will present and add a written summary page at the front.

Ten minutes of checks before it leaves your desk

None of this needs design skill. It needs you to read your own deck as a stranger will.

• Read only the headlines, in order. Do they form the argument on their own?
• Ask someone outside the project to state your ask after the first three slides.
• Print two slides to a page in greyscale. Anything that vanishes was relying on colour alone.
• Hand it to a colleague and ask them to summarise it back, with no narration.
• Check the final slide says what happens next, with a name and a date.

A deck is neither a performance nor a report. It is an instrument for getting a decision taken by people who are busy, distracted, and usually reading it when you are somewhere else. Order the argument for them, give every slide one job, and write down what you want.

---

Written by Fibvo — Bengaluru and Erode, India — working with clients in other time zones. Questions, or think we have it wrong? Tell us.