Presentation Statistics (2026): 16 Verified Numbers (and 9 Myths)

This page exists because researching it was genuinely annoying. The presentation-statistics genre is unusually polluted: numbers get quoted with confidence, sourced to a blog that sources it to another blog, and the trail dies. Two of the most repeated "facts" in the category are inventions. So the rule here is simple — if we couldn't reach a primary source, the number isn't in the first list.

How to read this page

• Primary source named for every figure, with the publication date and the fieldwork method where it's disclosed.
• Vendor-published is flagged. A statistic from a company that sells the thing being measured isn't worthless, but you should know.
• Stale is flagged. Anything over about three years old is marked.
• "Environment" stats are kept separate from "audience" stats. This is the single most common error in the genre — see the note before section 5.

1. The one presentation-specific behavioural stat

PowerPoint edits spike 122% in the last ten minutes before a meeting. Everyone finishes their deck at the last possible second, and there is telemetry to prove it.

Source: Microsoft WorkLab, "Breaking down the infinite workday", 17 June 2025. Method: aggregated, anonymised Microsoft 365 telemetry through February 2025. Vendor-published — Microsoft owns PowerPoint.

This is the only presentation-specific, behaviour-based statistic we could find from a named primary publisher. It's measured from what people actually did, not what they said in a survey, which makes it unusually trustworthy for this category.

2. The environment your deck lands in

All four from the same Microsoft telemetry report (17 June 2025):

• Workers are interrupted every 2 minutes during core hours — roughly 275 interruptions a day.
• 57% of meetings are ad hoc, with no calendar invite.
• About a third of meetings span multiple time zones, up 35% since 2021.
• The average worker gets 117 emails and 153 Teams messages per weekday.

Taken together these say something more useful than any attention-span claim: you are not presenting into a quiet room. You're presenting into a stream of interruptions, often to people in three time zones who joined a call that wasn't on their calendar.

3. AI adoption at work

21% of US workers say AI does at least some of their work — up from 16% a year earlier. Just 2% say it does "all or most," a figure that didn't move.

Source: Pew Research Center, 6 October 2025. Method: American Trends Panel, 5,010 employed US adults, fielded 2–8 September 2025, probability-based. Not vendor-published — the most methodologically solid source on this page.

85% of knowledge workers use AI at work, but only 29% have embedded it into their workflows. The gap between "uses AI" and "has changed how work happens" is the whole story of 2026.

Source: Atlassian Teamwork Lab, State of Teams 2026, 27 April 2026. Method: double-blind survey of 12,035 knowledge workers and 173 Fortune 1000 executives, Jan–Feb 2026. Vendor-published, but the methodology is unusually well disclosed.

66% of AI users say they spend more time on high-value work, and 58% say they produced something they couldn't have a year ago.

Source: Microsoft 2026 Work Trend Index, 5 May 2026. Method: Edelman Data x Intelligence, 20,000 knowledge workers across 10 markets, fielded 18 Feb – 7 Apr 2026. Vendor-published — Microsoft sells Copilot.

84% of developers use or plan to use AI tools — but only 3.1% "highly trust" its accuracy, and 45.7% actively distrust it.

Source: Stack Overflow 2025 Developer Survey. Method: ~49,000 responses, 177 countries — a self-selected convenience sample, so treat it as directional. Developer-specific; don't generalise it to all knowledge workers.

That last pairing is the most honest thing in this section: adoption is near-universal, trust isn't. Which is roughly the correct posture toward an AI first draft of anything, including a deck.

4. The scale of the tools

• Google Workspace: 3 billion+ users, 13 million+ customers — Google, 23 April 2026. Self-reported. Important: that's all of Workspace, Gmail included — not Google Slides. Google has never published a Slides-specific number, and "3 billion" has been the stated figure since 2021, so treat it as a floor.
• Canva: 265M+ monthly active users, 31M+ paying, $4B ARR — company figures disclosed February 2026, reported by TechCrunch. Unaudited.
• Gamma: $100M+ ARR, 70M users, $2.1B valuation — announced November 2025 alongside a funding round, independently reported by TechCrunch. Unaudited company figures. Useful as evidence that AI-native presentation tools reached real commercial scale.
• Microsoft 365 commercial seats grew 6%, with commercial cloud revenue up 19% — Microsoft FY26 Q3. Note that Microsoft does not publish a PowerPoint user count and hasn't for years — see the myths below.

5. Hybrid work — and a warning about attention stats

Among remote-capable US employees: 51% hybrid, ~26% fully remote, ~21% fully on-site, with hybrid workers averaging about 2.3 days a week in the office.

Source: Gallup, 2 September 2025. Method: Gallup Panel, probability-based; Q2 2025 wave surveyed 17,660 adults, 7–16 May 2025, ±1.1pp margin of error. Not vendor-published.

Median screen focus is about 40 seconds (mean ~47s), down from roughly 2.5 minutes in 2004.

Source: Gloria Mark et al., "Neurotics Can't Focus," CHI 2016 (ACM, peer-reviewed). Method: 40 information workers, computer-logged activity over two work weeks.

Read that one carefully, because it is the most misused statistic in this field. It measures how long someone stays on one screen before switching, during self-directed desk work. It is not a measure of how long an audience can listen to you. Any blog that presents a 40-second figure as "your audience's attention span" has misread the study. Use it to describe the environment, never the audience.

95% of business professionals admit to multitasking during meetings.

Source: Prezi State of Attention Report, fieldwork by Kelton Global, 2,036 US business professionals, 27 Feb – 7 Mar 2018. Vendor-published and eight years old — flagged on both counts. Worth knowing because it's the actual origin of a great many uncredited "attention" claims. Note the irony: Prezi's own survey found 59% said they could focus better than a year earlier, contradicting the shrinking-attention narrative it gets cited for.

The number nobody actually has

Here's the finding that surprised us most: there is no credible, recent, independent research measuring how long people spend building slide decks. Every figure in circulation traces back either to a 2020 vendor study by a company selling PowerPoint add-ins — i.e. with a direct commercial interest in that exact number being large — or to nothing at all. We couldn't verify it at source, so it isn't on this page.

Which means the honest summary of this entire topic is: almost every presentation-specific statistic is published by a company that sells presentation software (Microsoft, Canva, Gamma, Prezi). The only conflict-free sources here — Pew, Gallup, the CHI paper — aren't about presentations at all. That gap is worth knowing before you quote anything from this category, including us.

9 statistics to stop quoting

Each of these is widely repeated in presentation content. Each one dead-ends.

1. "Attention spans have fallen to 8 seconds — shorter than a goldfish"

Fabricated. Attributed to a 2015 Microsoft Canada report, which footnoted it to a site called Statistic Brain, which could not produce a source. The research bodies cited as backing had no record of it. Debunked by BBC More or Less, March 2017. (Goldfish don't have nine-second memories either — they retain spatial memory for months.)

2. "People remember 10% of what they read, 20% of what they hear…"

Fabricated, and falsely attributed to Edgar Dale. Dale's Cone of Experience (1946) contained no percentages at all; someone superimposed invented figures onto it around 1970. Debunked in a peer-reviewed four-article special issue of Educational Technology (Subramony, Molenda, Betrus & Thalheimer, 2014) — see "The Mythical Retention Chart".

3. "Communication is 7% words, 38% tone, 55% body language"

Real research, routinely misapplied. Mehrabian's 1967 studies involved 30 women judging emotional tone from single spoken words and 37 matching a word to a facial photograph. No conversations, no body language beyond facial expression, all-female samples. The ratio applies only to inferring feelings when verbal and nonverbal signals conflict — it says nothing about a presentation's informational content. Mehrabian himself objected to the misuse.

4. "30 million PowerPoint presentations are created every day"

Untraceable. Originates in a 2001 New Yorker piece phrased as "according to Microsoft estimates." Microsoft never published or substantiated it. Recycled for 25 years, sometimes inflated tenfold. Related: "PowerPoint has 500 million users" — no Microsoft source exists.

5. "63% remember stories, only 5% remember statistics"

Traceable, but not research. From Chip & Dan Heath's Made to Stick (2007), describing an informal classroom exercise with students giving one-minute speeches, recall tested ten minutes later. No sample size, no controls, never published as a study. Citable only if you describe it as what it is.

6. "Stories are 22 times more memorable than facts"

Unverified. Commonly attributed to a Stanford researcher; no published study supports the multiplier.

7. "Unproductive meetings cost US businesses $37 billion a year"

Unverified. Circulates via branded infographics from around 2011–2012 with no published methodology and undisclosed wage assumptions.

8. "Employees spend 31 hours a month in unproductive meetings"

Unverified against a primary source. Widely attributed to Atlassian research, but it doesn't appear in the 2026 State of Teams report, and we couldn't locate it on any primary page.

9. "Audiences lose attention after 10 minutes"

Unverified. Usually traced to a popular science book that doesn't itself cite a controlled study establishing the threshold.

What this means for your next deck

Strip away the folklore and the defensible evidence says something fairly plain: people are interrupted constantly, a third of your audience is in another time zone, and most decks get finished in the ten minutes before the meeting. None of that is an argument for shorter slides or more animation. It's an argument for making the point early, making it once, and making it legible to someone who just joined.

And if the 122% spike describes you — the deck that gets built at the last minute — that's the part worth fixing with tooling rather than willpower. See the free AI presentation maker guide, or how many slides you actually need for your slot.

Corrections welcome. If you can source any number in the myths list to a primary study, we'll move it up the page and say so.