As advertisers, we often think of the major ad platforms like Google as being black boxes with secretive, complex objectives, and it makes it hard to envision the future of advertising.
We can, however, find some clues by looking at the research they publish. Google has an extensive list of research papers, and they continually add to the vault. Reading these papers, we can start to get a peek inside the mind of Google’s decision makers, and use this insight to try to understand their priorities.
With that in mind, we took a look at two recent research papers to see what they can teach us about what ad platforms might be considering for the journey ahead.
Measuring Emotional Response
Digital advertising provides comprehensive metrics on actions people take after seeing an ad. While this data is highly usable for low-funnel channels like search, it’s less useful on upper-funnel channels like display and video. This is because these upper-funnel channels aren’t designed to elicit immediate action; they’re designed to build emotional connections.
People don’t buy a new car simply because they saw a car brand’s display ad. The objective of advertising on these channels is to build an emotional connection with users over time. So eventually, when they need a new car, they consider that brand.
But this presents a challenge: It’s not possible to measure emotions in digital advertising. No ad platform can tell you how many people smiled while watching your ad, nor how many people felt bored by the end of it.
A joint 2019 paper from Google and Microsoft, AttentiveVideo: A Multimodal Approach to Quantify Emotional Responses to Mobile Advertisements looks into how to solve this problem.
In the paper, the authors describe AttentiveVideo, software that uses a user’s mobile device to measure the user’s emotions while watching an ad. The user is instructed to hold his finger over the rear-facing camera lens to play the ad, at which point:
- Changes in the user’s blood flow are measured by turning on the camera’s flash, and recording how much is reflected back at the camera as blood flows through the user’s finger.
- Facial expressions are recorded by an algorithm using data from the front-facing camera.
What Does This Mean for Advertisers?
The researchers conducted a study with real users to gauge emotional response to ads. The collected data showed they were able to determine a user’s emotional response to an ad with 83% accuracy.
Given how valuable emotional data could be to brand advertisers, it’s not unthinkable that we might see ad networks crop up in future that are built around this idea.
Of course, there is likely to be opposition from some users, who may not feel comfortable granting advertisers access to such personal data. Still, as we’ve seen with other channels of advertising, the financial incentives for publishers may be so great that they find ways to overcome these barriers.
Understanding True Incrementality
Search advertisers have come to recognize that there is natural tension between paid and organic search. Both channels are competing for the same users, so there’s always a risk that paid search ads become pure cost, driving up prices for clicks that you could have otherwise acquired for free, organically.
The second Google paper we reviewed discusses this challenge in the context of organic search position. The paper, Impact Of Ranking Of Organic Search Results On The Incrementality Of Search Ads, explores how the position of a brand’s organic search results can affect the incrementality of their paid traffic on that results page.
The paper is in the form of a meta-analysis, through which the authors review 390 different Search Ads Pause studies — experiments where advertisers turn search ads off and measure the impact. The most interesting finding is that the proportion of a brand’s search ad clicks that are incremental — meaning they wouldn’t have occurred if there were no search ads — depends heavily on the position of the brand’s organic result.
If a brand is showing an ad on a particular term, and it has an organic result in position #1 for that term (immediately below the ad positions), then the authors estimate only 50% of the ad’s clicks will be incremental.
This means that the advertiser’s cost per incremental click — that is the cost of generating one more click than they would’ve otherwise gotten — is going to be twice their reported cost per conversion. The percentage of incremental clicks rises to 82% when the organic result is in positions #2 through #5, and to 96% when the organic result is below position #5.
In other words, the research suggests that when advertisers rank organically for a particular keyword, the higher the organic position, the more expensive it is to also use a paid ad for that keyword.
What Does This Mean for Advertisers?
These findings have huge significance for brands that rank well organically. While they don’t suggest that these kinds of brands should completely turn off paid search for terms where they appear at or close to position #1, it does demonstrate that advertisers should think carefully about how they bid.
One consideration is to incorporate these results into attribution. By looking at rank data for all organic search terms, you can then determine how many clicks, and therefore conversions, were incremental for that keyword. Knowing this, you can then apply different budgets accordingly.
Predicting Advertising’s Path
These are just two small glimpses into the questions platforms are actively answering and the challenges they’re trying to solve. In an environment where predicting platform changes is very difficult, this kind of insight can be very valuable to brands. These research papers offer us all a deeper understanding of the priorities these powerful platforms have, and allows us to consider what their mindset might be for the future.
If you want to discuss more about emotional responses, incrementality, or performance branding, WITHIN is here to help. Reach out any time.