In 2017, Piers Howe published a finding in Attention, Perception, & Psychophysics with a title that is also its argument: natural scenes can be identified as rapidly as individual features. He asked how briefly an image can be shown and still be understood, masked the displays properly so the answer would mean something, and then checked that answer against the simplest visual judgment he could construct: the orientation of a single line. Identifying an entire scene took no longer. The floor for both came in around thirty-five milliseconds. Not a hundred, which was the older estimate for a scene.
Put that next to the window your work actually gets. A single image in a feed is looked at for about nine tenths of a second. Thirty-five milliseconds is a small fraction of that, and it is shorter than the conscious mind takes to register that anything happened at all. And the brain has already gone to work: it has a rough read on the scene and is building a response before you know you looked.
The care in the method is the point, and so is what the method did not find. Howe timed the judgments separately and they came out within a few milliseconds of one another, close enough that the differences did not hold up. That is the result: not that one thing beats another, but that a whole scene is no harder to identify than one property of one line. Chapter 2 has the caveats that number depends on. One of them is worth repeating here, because the rest of Part II leans on it: these are minimum exposure times, not processing times. The image has to be present that long. The brain keeps working on it afterward, which is why the gist of a scene lands around one hundred fifty milliseconds even when the scene was gone in ten.
That sits underneath everything in Part II. If a scene has to be in front of the eye for something like thirty-five milliseconds before it can be understood, then even the median look at a piece of content, nine tenths of a second, is not the crisis the industry keeps calling it. It is about twenty-five times the minimum. The window is not too small. It is being spent badly.
What a Micro-Moment Actually Is
A micro-moment is the window in which a piece of content is actually looked at. Not the window in which it is on screen. Those are two different numbers, and the distance between them is the whole problem.
The two get confused constantly, and the confusion has a name: taking one layer of a nested set of durations and calling it “attention span.” The layers are all measured and they are all different. A work sphere runs about 6.1 minutes and a single work task about 3.4 (Talypova and colleagues, CHIWORK, 2025). An application window holds for forty-seven seconds on average. A visit to Facebook lasts about eighteen seconds, twenty-one times a day. A phone session has a median of ten seconds, and people start 228 of them a day (Brinberg and colleagues, Human Screenome Project, 2021). And a single image inside a feed holds the screen for a median of 0.90 seconds. None of those numbers contradicts another. They are nested, like rooms inside a building, and almost every argument about attention spans is two people standing in different rooms.
The other confusion is between being on screen and being looked at, and it is the one that costs designers money. How long something is on screen depends on the format, the device and the instrument that measured it. Chouaki and colleagues logged a median of 4.0 seconds for an ad unit (The Web Conference, 2024), but that was a Chrome extension counting pixels in a desktop browser, not a thumb on a phone. Parry and Masur’s phone figure for a single feed image was 0.83 seconds; the 0.90 quoted above is their median across all three devices, and the two are not interchangeable. Simonov and colleagues had an ad visible for nineteen seconds inside an article. A fourth instrument gives the largest number of all, and it belongs here because it is the one that cuts against this book. Zannettou and colleagues took donated TikTok histories from 347 real users, 9.2 million video views, and found a median attention of 82 percent of each video’s duration, at 89.9 videos and about 27 minutes a day (CHI ’24). That is not a glance. Short vertical video is a format where people watch most of what they are served, and any claim that attention has collapsed everywhere has to survive it. Read it with its own limits: platform log data rather than gaze, viewing durations inferred rather than timed, and more than half the participants in Africa. The line the evidence actually draws is between products, not between eras. A still image in a feed and an autoplaying video are different things, and the nine-tenths-of-a-second figure describes the first. Looking, where it has been measured as looking, varies far less: Simonov put gaze on that nineteen-second ad at 2.76 seconds (Journal of Marketing Research, 2025), and Lumen puts the average across formats between one and two. Whether pixel counters can stand in for gaze is a published question now rather than an assumption: Bruns and colleagues validated viewport logging against mobile eye tracking (Journal of Advertising 54(5), 2025). The micro-moment is the unit of design in the scrolling economy the way the pixel is the unit of resolution on a screen. Everything is built from it, and it is built per format.
The first three chapters set the context: the scroll is the medium, attention has collapsed, and the feed runs on economic logic. The micro-moment is where all of that turns into one constraint you can actually work with. It answers the question every designer in feed environments ends up asking: if I have less than two seconds, what can I do?
More than you think. But only if you design for the right things in the right order.
The Anatomy of 1.5 Seconds
Throughout this book, a micro-moment means one and a half seconds: the time a piece of content is looked at, not the time it occupies the viewport. It is a working unit rather than a measured average, and it is a generous one. It sits between the 0.90-second median and the 2.79-second mean of the best per-image measurement we have, which means designing for a micro-moment means designing for more time than the typical piece of content is actually given. I would rather the ruler err that way. Everything that follows uses it as a ruler, not as a finding. The brain spends it in stages.
The first thing that happens is the glance, and almost everything visual happens inside it. Before the conscious mind has decided to pay attention, the viewer already has the scene and most of what is in it: the layout, the density, whether there is a face, whether this is a product on an empty field or a crowd in a street. Thorpe, Fize and Marlot showed that a categorization that sophisticated can finish in under 150 milliseconds, from photographs held on screen for twenty (Nature, 1996). None of it is chosen. You don’t decide to take in a picture. It arrives.
And the glance is coarse. Larson and Loschky asked which part of the visual field does the work of getting the gist, and found peripheral vision more useful than central vision (Journal of Vision, 2009). That tells you what kind of information survives the trip. Not fine detail. Large shapes, broad tonal masses, the overall arrangement. Anything in your creative that only exists at full resolution is not in the glance at all. It is waiting with the words.
Then comes the read, and it is a different machine running on different rules. Text has to be foveated: the eye must land on it, hold for about a fifth of a second, and move on seven to nine letters at a time. Nothing about that is involuntary. The viewer may not experience it as a decision, but a decision is what it is. The glance is free. The read has to be earned, and most creative work is built as though the reverse were true.
Recognition happens inside the glance too, and it is less impressive than people assume, including people who sell brand equity for a living. Fabre-Thorpe and colleagues trained subjects on a set of images for three weeks and then tested them (Fabre-Thorpe, Delorme, Marlot and Thorpe, Journal of Cognitive Neuroscience, 2001): completely novel scenes were categorized just as fast as highly familiar ones. Familiarity bought no speed at all. So what visual equity buys is not a faster glance. It is a match. When the picture arrives, a brand with a distinctive system has something already stored for it to land on.
Then, if the content is still in front of the eye, and that is a large if, the conscious mind engages. Text becomes legible. The viewer reads the headline, builds meaning out of it, and decides: stop or scroll.
All of this happens inside the micro-moment, and there are two states in it, not four. This book’s first draft had a ladder: colour, then shape, then recognition, then meaning, each with a number attached. The numbers had no source and the rungs did not survive checking. What survives is simpler and harder to argue with. The picture arrives. The words wait. If your composition says nothing in the glance, the read may never happen, because the thumb has already moved on.
Source: Howe, Attention, Perception, & Psychophysics
Why Most Creative Fails the Micro-Moment
Once you know how the brain handles a scroll, the failure of most creative work stops being mysterious.
Most creative is designed message-first. The brief starts with what we want to say. The headline comes first in the brainstorm. The copy is written before the visual concept is finalized. The message is the star; the design is the vehicle.
In a micro-moment, though, the message doesn’t arrive first. The signal does. The composition registers as a whole before anything inside it is identified, and a headline cannot register until the eye lands on it and reads it. By the time your message is legible, the viewer’s brain has already formed judgments out of signals your team may never have consciously designed.
This is the gap, and it is a gap in sequence rather than in talent or effort. The creative was designed in the wrong order for the environment it lives in.
Source: Howe, Attention, Perception, & Psychophysics (2017) 10.3758/s13414‑017‑1349‑y · Larson & Loschky, Journal of Vision (2009) 10.1167/9.10.6 · Rayner, Journal of Eye Movement Research (2009) 10.16910/jemr.2.5.2 — The Scrolling Economy · Carlos Murguía, 2026
Consider how most brand campaigns are built. A team receives a brief. The brief identifies key messages, target audiences, and deliverables. The creative director develops a concept, usually anchored around a headline or a central idea. The design team builds the visual around that concept. The work is reviewed in a presentation room, where a group of stakeholders sits in focused silence and evaluates it on a large screen for several minutes.
Every step of this process is optimized for comprehension, for the slow, deliberate reading that only happens once someone has agreed to it. Every step ignores the glance: the part that arrives for free, and the part that decides whether the reading ever gets its chance.
None of this argues against messages or headlines or ideas. It argues about sequence. In the scrolling economy, signal comes before message. You have to engage the visual system before you ask the cognitive system to work. And the creative process has to start where the viewer’s experience starts: with the first thing the brain sees, not the first thing the brand wants to say.
The One-Thing Test
One practical tool comes straight out of the micro-moment framework. I call it the one-thing test, and it is the most useful question you can ask about any piece of creative headed for a feed:
If this creative can only accomplish one thing in 1.5 seconds, what is it?
Not two things. Not a message and a brand signal and a call to action. One thing.
The one-thing test forces a brutal clarity that most creative processes avoid. When you’re working inside a ninety-minute brainstorm or a multi-round review process, it’s easy to add layers. Another element, another supporting visual, another tagline, another logo placement. Each addition seems small in the context of the full design. But in the context of a micro-moment, each addition is competing for a share of the same finite cognitive resource, and when everything competes, nothing wins.
The answer to the one-thing test should fall into one of three categories:
1. Brand recognition. The viewer’s brain connects this content to a brand they already know. It needs no message and no comprehension. The signal drops a recognition cue that accumulates over repeated exposures. This is the strongest outcome available in the scrolling economy, because the whole thing happens inside the glance, before the viewer has agreed to anything.
2. Emotional resonance. The creative triggers a feeling, curiosity or warmth or surprise, that doesn’t ask the viewer to read or understand anything specific. It earns its place on this list, but not as the shortcut past attention the industry usually sells it as. Every brain region that responds differentially to emotional faces, the amygdala included, does so only when attentional resources are available, and the processing appears to be under top-down control (Pessoa, McKenna, Gutierrez and Ungerleider, PNAS, 2002; Pessoa and Adolphs, Nature Reviews Neuroscience, 2010). The one exception on record is narrow enough to state in a sentence: an amygdala response at 74 milliseconds, to fearful faces only, carried by low spatial frequencies only, and explicitly not to arousing photographs of scenes, which is what advertising actually uses (Méndez-Bértolo and colleagues, Nature Neuroscience, 2016). Emotion works because a feeling can be taken from a composition without reading it, not because it skips the queue.
3. A single idea. The creative communicates one clear, immediate concept. Not a story or an argument, but one thing expressed visually that the viewer can grasp in whatever is left of the second and a half once the glance has bought it. A number. A contrast. A before-and-after. A visual metaphor so clear it requires no explanation.
If your creative is trying to do something other than these three things in a micro-moment, it is almost certainly trying to do too much. And if it’s trying to do two or three of these simultaneously, it is definitely trying to do too much.
The one-thing test doesn’t mean your creative can only contain one element. It means your creative has to be built so that one element dominates, clearly enough to survive a glance. The glance carries the coarse whole; the read carries everything else, and the read has to be earned. The one thing you’re optimizing for decides which of those two you are designing for.
Framework: The Scrolling Economy · Carlos Murguía, 2026
Micro-Moments Are Not Small. They Are Precise.
There’s a natural resistance to the micro-moment framework, and it goes something like this: if I’m designing for 1.5 seconds, doesn’t that mean I’m making shallow work? Isn’t this just an argument for oversimplification? For dumbing things down? For reducing design to a thumb-bait exercise?
Designing for the micro-moment does not mean making simpler creative. It means making more precise creative, and the difference is large. Simplicity removes elements. Precision arranges them. Simplicity says: do less. Precision says: do the right thing first.
A haiku is not a simple poem. It’s a precise poem. Every syllable carries weight because the form demands it. The constraint doesn’t reduce the art; it concentrates it.
The micro-moment operates on the same principle. When you know the glance carries the coarse whole and little else, you don’t strip color out of your design. You deploy it more deliberately, because a palette is a property of the whole and is therefore already in the glance. When you know the picture arrives and the words wait, you don’t remove your logo. You design a visual system so distinctive that the logo becomes optional. When you know the viewer’s conscious attention might never arrive, you don’t abandon the message. You earn the right to deliver it by putting something worth recognizing in the part that arrives for free.
The creative teams that thrive in the scrolling economy focus their work rather than simplify it. They know what a glance can carry, and they design for it.
There is real craft in making something that communicates in a fraction of a second. It takes more design intelligence than a billboard does, and more strategic clarity. The micro-moment is the hardest creative problem in modern communication, and it should be treated that way.
Framework: Howe, Attention, Perception, & Psychophysics (2017) 10.3758/s13414‑017‑1349‑y · Larson & Loschky, Journal of Vision (2009) 10.1167/9.10.6 · Rayner, Journal of Eye Movement Research (2009) 10.16910/jemr.2.5.2 — The Scrolling Economy · Carlos Murguía, 2026
From Unit to System
The micro-moment is the unit. But a unit is not the whole.
If your creative succeeds in the micro-moment, if the glance carries a signal strong enough to be recognized and the second and a half behind it delivers something worth the stop, it has done something real. But it has done it once. In one scroll, on one screen, for one viewer.
The scrolling economy, as Chapter 3 established, rewards cumulative signal over transactional impact. One successful micro-moment is valuable. Hundreds of them, all built on the same hierarchy and all depositing the same recognition cues, are an asset.
This is the bridge to the rest of Part II. The micro-moment is the unit; the chapters that follow are about what to put in it, in what order, at what scale, and by what act of translation.
Each of these ideas builds on the micro-moment. None of them work without it. Because in the scrolling economy, you don’t design for campaigns, deliverables, or approval rooms. You design for the 1.5-second window where your work either earns attention or ceases to exist.
Framework: The Scrolling Economy · Carlos Murguía, 2026
That window is not a limitation. It’s a lens. And everything you create should be evaluated through it.