Seen is not remembered

In 2018, Megan deBettencourt and her colleagues ran an experiment with an unusual twist. People sat at a computer doing a long, repetitive task, and the computer watched how quickly they responded - a well-used signal of whether their attention was on track or drifting.

Then the computer used that signal. When someone’s attention was in a good state, it showed them a new image to remember. When their attention had drifted, it showed them one too.

Later, people remembered the images shown during high attention better than the ones shown while their attention had slipped. Same images, same people, same screen. The only difference was the state of their mind at that moment.

Which means a media plan that counts seconds on screen might be counting the wrong thing. Being seen and being remembered are different things, and most of what we measure only captures the first.

Why this matters for marketers

Marketing buys attention by the second: viewable impressions, time in view, watch time, and a growing set of vendor “attention metrics”. Some of those are useful. But I couldn’t find any peer-reviewed study validating vendor attention metrics against real outcomes like sales or pipeline, and the research on memory suggests why that matters.

In the last essay I listed the brain facts that didn’t survive. This one is about a finding that did, and what it means for how we design and measure creative.

So I’ve broken it into three parts: why memory depends on the moment, what earns attention that sticks, and how to measure and design for memory rather than exposure.

1. Memory depends on the moment

The deBettencourt study isn’t a one-off. In 2020, a team led by Kevin Madore recorded brain activity and pupil size in 80 young adults as they tried to remember things. Signs of an attention lapse in the moments just before a memory attempt - a particular brain rhythm and changes in pupil size - predicted that they’d forget. It also found that people who multitasked more with media in daily life were more prone to those lapses, and that went with worse memory.

What we pay attention to also depends on what we’re trying to do. In a brain-imaging study, Mariam Aly and Nicholas Turk-Browne found that when the hippocampus - the region central to forming new memories - was tuned to the task a person was focused on, the details relevant to that task were remembered better.

My inference for B2B is that buyers store what fits the job they’re doing right now. Whether content drives action depends on what they’re thinking about at that moment - the same message might land or vanish depending on when it arrives. So frequency could be as important as the message itself, because it raises the chance that one exposure lands at the right time to resonate.

But how you add frequency matters. A burst of exposures can land entirely inside one bad moment, and the gain from repetition falls off quickly after the second time someone sees a message. Spread out over time, frequency gives a message more chances to meet the right moment - which is the subject of the next essay.

2. What earns attention that sticks

If memory depends on the moment, creative has to work harder to be stored, and the eye-tracking research has some practical answers.

Show the brand early. Rik Pieters and Michel Wedel tracked the eyes of more than 3,600 consumers looking at 1,363 print ads. The picture was best at capturing attention, but the brand was best at passing attention on to the rest of the ad. In their earlier modelling, looking at the brand fed later brand memory.

Cut the clutter, not the design. Dense detail - lots of small elements competing - hurt attention to the brand, while a well-designed, more elaborate layout helped. A lot of B2B creative fails on exactly this: logos, badges, screenshots and five claims in one frame.

Put surprise next to the brand. In studies of video ads, surprise and joy concentrated viewers’ attention and kept them watching. Entertainment that came after the brand appeared was linked to purchase intent, while entertainment before it wasn’t, and brief, repeated brand moments reduced skipping compared with one long logo at the end. The usual B2B pattern - a 30-second story, then a logo end-card - might be the weakest version.

Match the context. In an eye-tracking study of online text ads, most people looked at the ads at least once, but the ads that matched the page’s subject were remembered better. It suggests “banner blindness” is often closer to banner amnesia: the eyes landed, but nothing was stored.

3. Measuring and designing for memory

So what should we measure instead of seconds?

The most defensible answer I’ve found is shared attention. When a group’s brain activity rises and falls together while watching something, it predicts how the wider audience will rate it. In online education, when viewers’ eyes moved together on the key moments of a video, it predicted their test scores - and the finding held with ordinary webcams and more than 1,000 people watching at home. It’s a better question than “how long were they on the page”: were people looking at the same thing at the moment that mattered?

And for long content, you can design for memory directly. When students watching online lectures were given short tests between segments, their minds wandered less, they took more notes and they learned more - and they were less anxious about the final test. Mind-wandering is most costly on complex tasks, which is exactly what a B2B webinar or product demo usually is.

I think that’s one of the most practical findings in this whole series. A poll or a “which of these would you do first?” question every few minutes isn’t a gimmick. It’s a tested way to keep a complex message from slipping away.

What good looks like

A few checks I’d use on creative that needs to be remembered:

• Frequency, spaced out. More chances for a message to land at the right moment, but not all in one burst.
• Brand early, and more than once. Let the brand pass attention to the message, and pulse it through video rather than saving it for the end.
• One idea per frame. Cut the competing logos, badges and claims.
• Surprise next to the brand. Put the moment people remember where the brand is.
• Fit the buyer’s job. Match content to the task they’re trying to do, and the context they’re in.
• Build in retrieval. Polls and questions every few minutes in webinars and demos.
• Measure shared attention, not just time. Ask whether people were looking at the right thing at the right moment.

For me, yes and no

There’s a fair challenge here. Attention still matters - nobody remembers an ad they never saw, and time in view is at least a rough signal that the ad had a chance.

I agree that attention is necessary. Most of these studies also used consumers, students and lab tasks rather than B2B buyers, so the step to your media plan is my inference.

But the research is consistent on the point that matters: exposure is where memory starts, not where it’s decided. For me, the better question for any campaign is what buyers will be able to recall when they’re ready to buy, and that depends on far more than how many seconds they spent looking.

Where to start on Monday

1. Check where your brand appears. Look at your three main videos or ads and note how many seconds pass before the brand shows up.
2. Cut one frame down to one idea. Pick your busiest piece of creative and remove everything that competes with the main claim.
3. Add retrieval to your next webinar. Put a short poll or question in every five minutes or so, and compare the follow-up engagement.

Three tests to run

1. A brand shown early and often might be remembered better than a logo end-card. Run two cuts of the same video ad with an even split: the control saves the brand for the end, the variant shows it in the first few seconds and again through the video. Measure completion rate and click-through, and ad recall if your platform offers a brand-lift survey.
2. Ads that match the page’s subject could be remembered better than ads that don’t. Split a paid budget with the same creative: half on contextual placements about your buyer’s problem, half on audience targeting alone. Over a quarter, measure post-click conversion and pipeline created from the accounts reached.
3. Content that fits the job a buyer is doing might be acted on more than a general recap. Split event follow-up: the control sends the standard recap, the variant sends one piece tied to the problem each attendee named when they registered. Measure reply rate and meeting rate.