December 1, 2022
Previously we've shown that TV ads that are relevant to the viewer have an addressability uplift -- that is, they engage greater focus, are liked more and remembered better than ads that are not relevant to the viewer. Here we investigated why addressable TV is so effective. Is there something special that happens in a person’s brain when they see a relevant ad, something that makes it particularly memorable?
We conducted two experiments to investigate the impact of addressable ads.
The first used functional magnetic resonance imaging (fMRI) to directly measure brain activity while 24 volunteers watched TV programs with embedded ads that were related to their interests (addressable ads) or not. This first experiment studied how specific neurological systems were activated by addressable content.
The second experiment was conducted with 200 volunteers who participated in a behavioural experiment using the on-line experimental tool Gorilla to measured the behaviour impact of addressable ads. These participants were shown the same content as in the first experiment, then their memory and recall for addressable content was measured using an online behavioural survey. This second experiment was designed to measure what impact addressable TV ads have on participants’ memory and recall of relevant (addressable) ads vs. non-relevant ads. In addition, it explored whether the TV context (that is choosing your own show or having it chosen for you) influenced the addressability uplift.
• Produce higher brain activation in four key networks relating to attention, emotion, reward and memory, showing that multiple brain systems contribute to the addressability uplift effect
• Are robustly recalled better and faster than non-addressable ads. In other words, memory for relevant ads is higher than for less relevant ads.
• Are recalled better in any context - whether participants chose their TV show or not
• In the context of a TV show chosen for the participant, have an even greater effect on brain networks relating to attention, emotion, memory and reward
• Have a recall effect that changed according to age and context
The results helped Finecast to organically grow their client base, to inform their future product roadmap and to enhance their position as leading the industry narrative. Of course, they are also a company of really curious people who love to learn more about how TV addressability works for both consumers and advertisers.
“The outcomes of the study were impressive and have significant implications for the media industry. They shed light on why addressable TV campaigns continually achieve substantial results for clients, and they provide further evidence that these campaigns deliver a strategic advantage to brands who include this technology in their budgets.” Harry Harcus, UK managing director at Finecast
Ads (n=4) were chosen from four different categories of products (auto, travel, pets and video gaming), for a total of 16 ads. In addition, we used five half-hour TV programs (The Bold and the Beautiful, Carp Wars, Dog the Bounty Hunter, Married with Children and Modern Family).
1304 volunteers were recruited via Prolific to complete an on-line, pre-screening survey that asked about their leisure time activities. Embedded within the questions were four asking about whether participants owned a car, whether they went on cruise vacations, whether they owned a dog, and whether they owned a video game console. These four questions were used to assess the suitability of each person to participate in the main experiments. Responses were statistically analysed to account for response bias and to find participants with interest in two, and only two, of the ad categories featured in the experiment. This resulted in potential participants who fell into one of the six categories below:
In this way we ensured that half of the ads ads were relevant to a participant and half were not. In other words, all participants saw exactly the same 16 ads. The only thing that changed was whether the ad's topic was relevant to the participant or not.
The fMRI experiment
The first experiment looked at what impact addressable ads have on brain activity. We sought to establish what effect (if any) addressable ads have on neurological activity in brain regions specifically related to attention, emotion, reward and memory.
Participants (n=24) made two, one-hour visits to the Birkbeck-UCL Centre for Neuroimaging approximately seven days apart. Each visit consisted of two shorter (7-minute) fMRI screenings followed by a longer, 30-minute scan. The short scans were designed to localise brain activity associated with attention, emotion, reward, and memory, whereas the main experiment measured brain activity in each of these four networks while participants watched a 30-minute TV episode with embedded ads (addressable and non-addressable).
To enhance the reliability of the results, conditions at each of the two participant visits were varied. During the first visit, half of the participants chose the show they watched, while for the other half, the show was randomly selected by a computer; on the second visit, the conditions were reversed. This allowed us to compare how participants responded to addressable and non-addressable ads under both conditions (self-selected vs. other-selected).
The most striking finding was that across all four brain networks - attention, emotion, reward, and memory - we found evidence of an advantage for addressable ads. This suggests that multiple brain systems contribute to the addressability uplift effect.
The brain’s attentional network, shown below in blue, was identified by a visual search task. There was significant activation in two regions typically associated with sustained visual attention:the frontal eye fields (FEF) and the intra-parietal sulci (IPS). We then measured brain activity in this network and examined the difference between addressable and non-addressable ads. We found there was consistently greater attention-related activity for addressable relative to non-addressable ads, indicating greater attention for relevant ads:
To illustrate these results, imagine that Mel is an auto-enthusiast but is less keen on cruises. When she sees a car ad, her attention increases much more than when she sees an ad for a cruise. This increase in attention is slightly larger when the program she is watching was chosen for her.
An emotional image task identified emotion-related brain activity in the amygdala, shown in red. Within this region, there was less emotion-related brain activity for advertisements than for the TV shows, which shown by the downward direction of the bars. It is not surprising that the longer form of the TV program generated a greater emotional response than seen in the shorter ads.
Then we examined the difference between addressable and non-addressable ads. We observed an uplift for addressable ads with greater emotion-related activity for addressable than for non-addressable ads. The addressable ads were less negative than the non-addressable ads. In other words, the level of emotional brain activity for addressable ads was closer to what was seen in the TV programs.
As an illustration, imagine that Lorenzo loves cruises but not video games. When he sees an ad about a cruise, he has a stronger emotional response to it than to a video game ad. If the two ads occurred while watching a show Lorenzo did not chose, then his increase in emotional engagement for the cruise ad is even larger.
A gambling task was used to identify reward-related activity in a region of the brain called the ventral striatum, shown in green. Again, we observed a clear advantage for addressable ads which showed greater reward-related activity than non-addressable ads.
As an illustration, imagine Ava loves video games but is less interested in dogs. When she sees a gaming ad, she finds it more rewarding than an ad about dog food. If the two ads occurred while watching a show Ava did not chose, then she finds the gaming ad even more rewarding than the dog food ad.
Finally, we identified brain activity associated with successfully encoding new information engaged the hippocampus, shown in purple. Here, we observed an overall reduction in memory-related brain activity for advertisements relative to TV shows – in other words, the TV programs were more memorable than the ads. But once again, we observed greater activation for addressable ads.
For example, Andre loves dogs but is not fond of cars. When he sees dog ad, he encodes the ad into long-term memory better than he would for an automobile ad. If the dog ad occurred while watching a show Andre did not chose, then he remembers the ad even better.
To summarise the fMRI experiment, we found that addressable ads produced higher brain activation in four key neural networks - relating to attention, emotion, reward and memory - demonstrating that multiple brain systems contribute to the addressability uplift effect. These were present regardless of whether the participant chose the TV show or not. This is encouraging because it suggests that addressable ads are effective whether you’re the person holding the remote or not.
The behavioural experiment
This second experiment sought to replicate our findings of a behavioural uplift effect using a new set of ads and an independent, larger sample of participants in order to ensure the validity and robustness of the original results. In addition, we investigated whether the TV context affected the addressability uplift.
Participants (n=200) completed an on-line experiment in which they watched a TV program with ad breaks. They were split evenly between male and female and came from three age bands: 18-34 (40%), 35-54 (45%) and 55-70 (15%) to reflect a nationally representative demographic distribution.
In this experiment, participants were randomly placed in one of two context conditions. Half of them were in the choice condition, and were asked to choose which TV show they wanted to watch during the study. The other half were in the random condition, and the show was selected for them at random. Participants then watched four ads, followed by roughly 6 minutes of the TV show, and then repeated this cycle four times. There were four ads from each of four categories, and they were presented in a fully randomised order.
After watching, we tested participants’ memory. They were shown an image and asked to identify, as quickly as possible, whether or not they had seen it previously in one of the adverts. Half the time it was a still from one of the ads or the product logo they had seen; the other half of the time it was a still or logo from a matched ad that they did not see. Finally, they were asked to rate how much they liked each ad.
Participants liked addressable ads more and remembered them better than ads that were not relevant to them. Participants’ accuracy in correctly recalling ads was roughly 10% higher for addressable content. In addition, they recalled addressable ads more quickly, providing corresponding evidence of better memory encoding for these ads.
We investigated participants’ memory further by examining the effects of context and age. We found a surprising but apparently robust result. In the plot below, we show participants’ age on the horizontal axis. On the vertical axis is their memory score. The separate lines for the addressable and non-addressable ads show the model’s estimate of how participants respond differently to this content at different ages. We have also split the participants by ad context. Those that chose the TV themselves are on the left; those that had the TV show chosen at random by the computer are on the right. The strength of the evidence is shown in two ways. The coloured percentages show the evidence we have that age had an effect on memory, for each type of ad in each context. For example, there is very strong evidence that when people chose their TV show, addressable content was remembered better as people got older. Regions are shaded grey when there was no evidence that participants at their age remembered addressable content better than non addressable.
This intriguing pattern of results shows that when people have chosen the TV show, they remember better the addressable content the older they get, whereas their memory for non addressable content doesn’t really change. Conversely, when the show is chosen for people at random, memory for addressable content is the same across age, whereas memory for non-addressable improves with age. Another way to put this is by asking - what context does addressability work the best? The answer here is that for older people, it is when they choose their own TV show, but for younger people it is when the show is chosen for them.
We found compelling evidence that addressable advertisements have robust effects on participants, influencing both their neurological and behavioural engagement. We demonstrated that the addressability uplift effect can be seen in four different brain mechanisms, related to attention, emotion, reward, and memory. In other words, these results suggest that relevant TV ads are attended to more, are more emotionally engaging and rewarding, and are encoded into memory more effectively than ads that are less relevant. In addition, participants like and remember ads that are relevant to them better, even when memory is measured implicitly (ie via reaction times). The effect is robust to TV context, with equivalent uplift seen when the participant chose their own TV show or when one was randomly selected for them. Interestingly, for older people when they choose their own TV show this effect was enhanced while for younger people, the effect was enhanced when the show is chosen for them. These findings provide novel insights into the efficacy of addressable advertising.
For additional information on consumer neuroscience, download our free Guide to Quality Neuromarketing.