The story came from an investigation on Habr, whose author set out to verify a claim familiar to anyone who has ever sat through a lecture on “proper presentations”: the brain processes images 60,000 times faster than text. The number is round, catchy, and supremely convenient—it gets used to justify infographics, video courses, and the eternal “less text on the slide.” The author combed through the literature and found not a single scientific study behind it.
The only trail led to a presentation booklet from 3M—the company that makes films, sticky notes, and protective coatings—where the number appeared with the attribution “according to behavioral research,” with no author, title, or year. I wanted to follow the trail to the end and see who, exactly, first printed those 60,000, rather than merely retelling someone else’s debunking. And, while I was at it—see what real measurements have to say.
The document most often cited is a two-page 3M booklet, “Polishing Your Presentation,” a promotional leaflet on how to deliver a striking projector presentation. There, in the second line, sits the famous phrase: “Did you know that visual aids increase learning by up to 400 percent? Are you aware that we process visuals 60,000 times faster than text?” Then comes the attribution that ought to explain everything: “These findings from behavioral research confirm our everyday experience.”
Not one citation—not to a person, not to a study. The booklet is an advertisement for slides, not an experimental report, and its entire “source” comes down to the phrase “behavioral research.”
The original was removed from the 3M website long ago and survives only in the web archive; in 2012, educator Alan Levine wrote to 3M asking for help finding the original source, and a company employee, replying a month later, sent him… the very same booklet.
There is a respectable branch to this story—the one people seize on more often than anything else. In 1986, the University of Minnesota and 3M conducted a real study: doctoral student Doug Vogel examined how visual aids affected the persuasiveness of a presentation. Its finding: presentations with visual aids proved 43 percent more persuasive than presentations without them.
This was an actual study, with an author and a method: graduate students were persuaded to sign up for a seminar using different presentations, and the researchers counted the responses. It did not measure image-processing speed at all.
When Levine found Vogel and asked him directly, he replied: “My work had nothing to do with visual processing speed.” Forty-three percent persuasiveness and sixty-thousand-fold speed are numbers from different universes, yet it is the former that has been dragged into footnotes beneath the latter for decades.
The earliest known point is not 3M, but the June issue of Business Week in 1982. Though that was not an article either: it was an advertising insert titled “Computer Services: Management’s Answer to Data Overload.” The president of a company called Computer Pictures Corporation, Philip Cooper, wrote in the ad that “people assimilate visual information approximately 60,000 times faster than printed text.” No citation, no method—just a claim purchased in ad space.
Levine tracked Cooper down at MIT Sloan, wrote him two letters, and called twice. There was no response; the voicemail box was full. Over the years of digging, he had time to offer first 60, then 120 dollars to anyone who could identify the real source—the reward went unclaimed.
This is where the sources begin to diverge slightly. An investigation by Instructional Design Tips claims that Cooper eventually admitted he no longer remembered where he got the number. In Levine’s records, which collect the entire correspondence, there is still no reply from Cooper—only auto-responses and a full voicemail box. It is unclear which version to believe, but the outcome is the same: the number with its six tens of thousands has been roaming the world for more than forty years, and its only known author cannot name the source.
To grasp the scale of the stretch, it helps to look at proper measurements. MIT neurophysiologists led by Mary Potter showed in 2014 that the brain recognizes the gist of an image in 13 milliseconds—faster than anyone had thought, and a genuinely astonishing result. The caveat almost always left out: this is the roughest possible recognition (“picnic,” “smiling couple”), not understanding.
The participants also trained from trial to trial and received feedback, meaning they rapidly learned to snatch the content at maximum speed.
On the other side—text. Photo: recognizing a word takes an experienced reader usually 200–500 milliseconds, and the fastest readers 160–180. Reading at a relaxed pace is far slower than people tend to think: Mark Brysbaert’s meta-analysis of 190 studies and 18,573 participants gives 238 words per minute for nonfiction and 260 for fiction (183 aloud).
That gives an honest estimate of the gap: if you compare recognizing an image with recognizing a word, you get somewhere between 6 and 15 times—not 60,000, not once. Just do the reverse calculation: with a word taking 200 milliseconds, an image would have to be processed in 0.003 milliseconds—that is faster than a signal travels along a neuron.
Here, too, the sources split according to taste. Some reviewers, having checked the myth, still conclude that the brain definitely handles images faster than text—just without attaching any number. Others, like the instructional design specialist at Instructional Design Tips, go further and discard the premise itself: there is nothing to compare, because low-level vision—edges, color, pattern—is indeed lightning-fast, while high-level understanding, where meaning must be connected and shaped into a conclusion, takes comparable time in both modalities.
In other words, the dispute is not over whether it is 6 times or 15. It is over whether we are measuring detection or understanding. Those are different questions with different answers.
Sixty thousand is not a loner. It has a large family, and they all came out of the same incubator: numbers printed without a method. The most famous relative is Edgar Dale’s “cone of experience” from 1946, to which someone, twenty years later, glued percentages of retention. Dale himself gave no numbers at all: he simply arranged ways of presenting material in an intuitive order.
The figures—10/20/30/50/70/90—first appeared in a 1967 article by a Mobil Oil employee named D. G. Treichler, also without a single citation, and apparently trace back to a handout from the Socony-Vacuum oil company in the forties. Learning-myth researcher Will Thalheimer called the resulting hybrid “a fraud.”
Two more stand nearby. Mehrabian’s 7–38–55 formula—the claim that supposedly 7 percent of meaning lies in words, 38 in intonation, and 55 in facial expression—was derived from experiments involving extremely ambiguous remarks, and Mehrabian himself disowned that interpretation. And “65 percent of people are visual learners,” a number without a source, was attributed to the VARK model, which has not held up in meta-analyses.
At the foundation of it all lies the harmless saying “a picture is worth a thousand words,” which, as Quote Investigator found, has its own evolution and competing authorial candidates—including a “ten thousand words” version attributed to newspaper man Arthur Brisbane in 1915. Sixty thousand is the same saying, only run through the machinery of advertising exaggeration.
There is one genuinely beautiful scientific fact in this story, obscured by the myth. In 1960, George Sperling conducted an experiment that became a classic: he showed people a grid of letters for a fraction of a second.
If asked to name everything they remembered, a person produced an average of about 4.5 letters—and that number barely changed no matter how many letters were in the grid. But if, after the display, they heard a sound cue and were asked to name just one row—any row, chosen at random—the participants gave 75–90 percent of its contents.
The logic is cruelly simple: if three quarters of any random row are available, then immediately after the flash, nearly the whole grid must be sitting in memory. The brain really does see everything at once and in parallel, far more than it can say aloud. The catch lies elsewhere—this stockpile melts away in fractions of a second. Delay the cue by less than a second and the advantage vanishes completely; the image drains away like water from a cupped hand.
A 2023 online replication with 64 participants confirmed Sperling again: iconic memory—as Ulric Neisser later called it—lasts roughly from 120 to 500 milliseconds. That is the whole “visual advantage”: not reading speed, but a brief window in which the brain manages to grasp a structure as a whole before it dissolves.