Network checks were blocked in this environment, so I’m citing only sources whose URLs I’m confident exist and dropping the one I couldn’t vouch for.
A marketing presentation works when it does three things in order: states the one decision you want the room to make, proves it with evidence the audience cannot get anywhere else, and ends with a step someone can take this week. Build a 10-to-15-slide core with one idea per slide and under 25 words of on-screen text, and move everything else to an appendix. Let visuals carry the argument instead of decorating it. Rehearse aloud at least three times and time the talk to about 60 percent of the slot, so questions fit. Then build a second, readable version, because most decks are forwarded, not presented. That is the whole method. What follows is how to execute each part and where it usually breaks.
Why do most marketing decks fail?
They ask the audience to read and listen at the same time. Working memory is small, and text on screen competes with the speaker’s voice for the same channel. Sweller’s cognitive load research and Mayer and Moreno’s multimedia principles both show that redundant on-screen text lowers comprehension rather than reinforcing it. A slide of six bullets read aloud is the textbook failure case.
The second failure is structure. Decks built as a tour of the company, the market and the product leave the decision to the final slide, by which point attention is gone. Amazon went so far as to replace slide presentations with six-page written narratives, a choice Jeff Bezos defended in the 2017 shareholder letter. You do not need to ban slides, but you should be able to write your argument as prose first. If it does not hold up as two pages of text, no template will save it.
How should a marketing presentation be structured?
Nancy Duarte’s analysis of persuasive talks, published in Harvard Business Review, found they alternate between what is and what could be, then close on a concrete call to action. That pattern maps cleanly onto a marketing deck.
| Slide | Job | Common mistake |
|---|---|---|
| 1 | The decision you are asking for, stated plainly | Title slide with logo only |
| 2–3 | The customer’s current situation and its cost | Generic market-size chart |
| 4–6 | What changes with your offer, shown not described | Feature list |
| 7–8 | Proof: results, named customers, before-and-after numbers | Unsourced “industry statistics” |
| 9–10 | Competitive position and why alternatives fall short | Checkbox grid where you win every row |
| 11–12 | Price, timeline, next step, who does what | “Questions?” |
Keep every supporting chart, methodology note and bio in an appendix. The appendix can be 30 slides long. The core cannot.
How much belongs on a slide?
One idea, one visual, and a headline that states the takeaway as a sentence. “Churn fell 31% after onboarding video” is a headline. “Results” is not. Body text, if any, stays under 25 words and above 24-point type so it survives projection and screen sharing.
Contrast matters more than brand colours. Text needs at least a 4.5:1 contrast ratio against its background to meet WCAG 2.2 Success Criterion 1.4.3. Light grey on white, the default in many templates, fails that test. The Section 508 presentation guidance adds the practical checklist: reading order, alt text on images, and no information carried by colour alone. Those rules also make the deck legible on a phone, which is where forwarded decks get read.
Where does video belong in a presentation?
We cut video for client decks regularly, and the clip that works is 30 to 60 seconds, placed after the problem is established and before the proof. It should show the product or the customer, not an animated logo. A single clip of that length cut from existing footage typically runs from the low thousands of dollars. Add a shoot day and the budget climbs into five figures, depending on locations, crew size and talent.
What goes wrong is nearly always technical, not creative. Google Slides cannot embed a local video file, so the clip has to live on YouTube or Drive, and a conference-room network decides whether it plays. PowerPoint embeds the file, which pushes a deck past 200 MB and makes it fail to send by email. Google’s published limits cap a Slides file at 100 MB when importing. Over Zoom or Teams, screen-shared video stutters unless the presenter enables the share-sound and optimize-for-video options, and even then the frame rate drops.
Our standing rules: export the clip as a 1080p H.264 MP4 at 8 to 12 Mbps, keep it as a separate file next to the deck, test it on the actual room system or call platform the day before, and have a hosted link ready to send so the audience can watch it full quality on their own screens if playback fails.
Should the deck you present be the deck you send?
No. The presented deck has sparse slides because you supply the words. The sent deck has to stand alone. Build the sparse version first, then duplicate it and add the speaker notes as short captions under each visual. Export the second version to PDF so fonts, layout and the embedded chart data stay fixed. Sending a 12-slide presenter deck without you attached is the most common reason a strong meeting produces no follow-up.
What has changed in how decks get built?
Three things since the early 2020s. First, 16:9 is the only aspect ratio worth designing for, because every laptop, television and call platform assumes it. Second, remote delivery is the default rather than the exception, which means your slides are competing with the viewer’s inbox on the same screen, and the 25-word limit matters more. Third, AI layout tools in PowerPoint, Google Slides, Keynote and Canva now handle alignment and formatting well. They do not handle argument. A generated deck looks finished and still fails the test of being written as two pages of prose.
Start by writing the decision you want in one sentence and the two pages of narrative that justify it. Build the 12-slide core from that narrative, strip every slide to one idea and one visual, and place a short product or customer video where the proof begins. Rehearse to time, test the video on the real system, and prepare the standalone PDF before the meeting, not after.
