A (1:51)

So from a pre testing standpoint, yes, I would go with my first two. One is immediately after and then we've done this ourselves where you would then do intercepts a couple days later.

B (2:02)

Okay, gotcha. Sorry, I just wanted to clarify that. So this study, it starts by explaining that most people look at recall within a few minutes, right? Like you watch the ad and then you ask some follow up questions. And they think that's a problem because this paper, they argue that short term recall is a poor proxy for how advertising actually works. Because people don't usually buy something within five minutes of seeing an ad, especially something like a television ad. They move through stages of awareness, consideration and evaluation over time. So the researchers set out to understand what actually makes ads stick over time and whether we can predict or even engineer that memorability. To answer those questions, they created the first large scale long term ad memorability data set. And they called it Lambda, thanks to a great name which. This data set included more than 1700 participants, 2200 real ads, and 276 brands across 113 industries. This wasn't just like a memorability game or some kind of lab setup. People were watching ads normally and then days later they were asked what they remembered. So that's why this study stands out. But they found a lot of things didn't matter for memorability. That included ad length, whether there was music or speech or the time of the day. But a few factors stood out in a big way. So, Rob, based on what we've talked about so far and what you know about ad memorability, what factor do you think mattered the most for long term ad memorability?

A (3:24)

I'm gonna go with distinctive assets for $200 there, Alex.

B (3:30)

Well, okay, that's not really a good guess because distinctive assets aren't even like used by a lot of researchers. Is that your final guess?

A (3:40)

What factors do you think matters most for long term ad memorability?

B (3:44)

Yeah, that's a good answer.

A (3:45)

Well, that's a good answer for that question.

B (3:49)

Yeah. So that's not. That wasn't this one, but that is a good answer. I'm sorry, I shouldn't give you a hard time about that. That's a good answer.

A (3:55)

I feel shamed.

B (3:55)

I'm sorry, do you want to try to guess another thing?

A (4:00)

Not anymore.

B (4:01)

Okay, sorry. Okay. The answer was emotion. That's what they found was that mattered the most. So I'm sure distinctive as it's also very important. But for that memorability, it was emotion, or more specifically, emotional intensity. So ads that triggered a strong emotional response were far more likely to be remembered days later. Now, I think distinctive brand assets, maybe if you're talking about like months or years down the line, that one would probably pop. More thinking.

A (4:26)

Yeah, but there was a twist.

B (4:29)

Yeah, negative emotion stuck out more than positive emotion. That doesn't mean your ads should be depressing. It just means that the intensity mattered more than positivity and pace. Also mattered. Ads with more movement were slightly more memorable. That's kind of fun. It's not a magic bullet, but visual energy did help memories form. And then there's an important one. Another factor was brand relevance. So one of the strongest correlations in the entire study was how relevant the brand already felt to the viewer. If someone already found a brand familiar or meaningful, they were more likely to remember its ads later. This showed up more clearly in categories like food, entertainment, and tech. Now things got interesting. You're going to like this part of the study, Rob. They built an AI model. They named it Henry, and it was designed to predict long term ad memorability. So it combined visual features, emotional cues, brand knowledge, all those things that they found mattered, and world knowledge from large language models. Henry outperformed existing models across every major memorability data set. But they didn't stop there. They asked, what if we then use this model to generate more memorable ads? So using a Data set of 5 million ads, they trained AI systems to create ads optimized for memorability based on what they knew and what they built this model for. Rob, how well do you think AI did at optimizing ads for memorability?

A (5:47)

Gosh, you know, AI follows frameworks really well. So I suppose if they've identified a framework that can repeatedly tap into people's emotions, then that would support their thesis and it would work.

B (6:02)

You're correct. The AI generated ads scored an average of 44% higher in predicted memorability than the original. So they were able to outperform the original ads. So what should marketers actually do with this research? First, stop treating recall like a pop quiz. Most ad testing still asks, do you remember this ad right now? But that's not how buying works. This study shows long term memorability is a much better proxy for real world impact. If your creative testing only measures immediate recall, you could be optimizing for the wrong outcome. It could be helpful to think about more lasting memories. Second, emotion beats information. That's maybe a headline from this study. It's that emotion is the strongest driver of long term memory. Ads that made people feel something, especially when it's intense or unexpected, were the ones that tended to stick more. And interestingly, slightly uncomfortable negative emotions often lingered longer than cheerful ones. That doesn't mean that all your ads should be sad. But it means maybe you shouldn't assume that rational messaging can do all your heavy lifting. Third, brand building makes your ads memorable. So people remember ads from brands they already find relevant. Which means your memorability is going to compound over time. The stronger your brand, the easier it is for each new ad to stick. Which is why brand building really isn't optional. And then AI can help spot patterns so models can predict memorability and even generate ads that score higher on memory metrics. So hey, if you're not using AI for some of your pre testing or creative iteration, it might be a good thing to test out. All right, time for a Rob GPT. Ads are like seeds. Most land on hard ground and never take hold. The ones that grow are planted where there's already some familiarity and emotion in the soil. This research shows that memorability follows those same conditions. AI can help spot which seeds are more likely to sprout, but the brand still decides what kind of garden it wants to grow.

A (7:45)

Super interesting. It reminds me of a study that Roy Williams, the wizard of ads, quoted. I can't remember if it was in one of his books or a lecture that I saw of him where he talked about adrenaline and the ability to increase adrenaline in a consumer will increase the ability for it to stick, like the memory to stick. He quoted a study that referenced people who were assaulted, which I know is a horrible thing to have to think about, but that's why people who are assaulted have such rich detail in what was going on in that moment, because the level of adrenaline in their system was so high. So to correlate this with what you were talking about with their study, that even negative emotions, actually negative emotions, maybe even score higher than positive emotions, is because you have that much more kind of adrenaline flowing through the system.

B (8:44)

Yeah, that's super interesting. And I, I think that you're right. That's probably why. And this study, I really like it. I think it's good for marketers to think, all right, how am I if I'm pre testing ads for memorability or I'm trying to measure memorability, how am I doing that? And then also, we've talked a lot about the power of emotion in advertising, but we've also talked about the importance of rational and emotional. And this doesn't mean go create a super scary ad, but still interesting takeaways. I think that's it for this episode of the Marketing Architects. We'd like to thank Taylor Delos Reyes for producing the show. You can connect with us on LinkedIn. And if you like the podcast, please leave us a review. Now go forth and build great marketing Marketing architects.