The 4 Problems Holding Attention Metrics Back from Wider Use
Jan 24, 2018
Traditional forms of video advertising on the web have been pretty straightforward; you pay a fixed amount based on how many people see your ad. But there’s one big problem in the ambiguity of this definition: what does“see” mean?
The Seeing Problem
Depending on your definition of this metric, you might end up paying very different rates for the same activity. For example, under the traditional idea of video advertising, a person “seeing” an ad can be described when an impression is served. In other words, whenever an ad is called from a server, an impression is said to have taken place and it counts as a user “seeing” that ad. But what happens if the ad isn’t properly displayed? What happens if the user clicks away before the video has a chance to finish loading? What if the user’s internet connection is so poor that the video doesn’t render the way it’s intended?
These scenarios occur regularly, and all of them prevent a user from getting the “full experience” of an advertisement. This is a problem – and the core problem that attention metrics hope to solve.
Attention Metrics (in Theory)
Attention metrics are a categorical type of metric designed to analyze whether or not a user actually paid attention to an ad, or whether it was simply called from the server. There are many different types of attention metrics, some of which are in broader use than others. For example, some video ad servers quantify an impression as taking place only if more than 50 percent of the pixels of a video ad are viewable for two or more contiguous seconds. Newer models are experimenting in many different directions, such as determining the amount of time an ad was in view and charging accordingly. Such models would allow for greater efficiency (and presumably, a greater ROI) for advertisers and could give ad servers a competitive advantage.
The problem is, there isn’t much in the way of a set standard for attention as a variable, and attention metrics in general are still in their infancy. Most video ad servers are still using old methods of counting views, which is leading to decreased efficiency on all sides.
Why Aren’t Attention Metrics in Wider Use?
Hypothetically, the simple addition of a few attention metrics could potentially increase advertising relevance on all sides and help the process become smoother and more cost efficient. So why are ad buyers and publishers so slow to adopt them?
- The subjectivity of attention. The first, and one of the biggest problems, is the subjectivity of user attention. There are many variables we can track, such as how a video displays, when it displays, and even certain actions that a user takes; but attention is still an internal, subjective experience. There’s no way to pop into someone’s head and figure out whether or not they’re actually paying attention to an ad. Video advertisers aren’t necessarily striving to achieve this (at least, not yet), but they are working in that general direction. Such progress demands an assumed correlation between certain measurable statistics (like percent of pixels viewed) and the unknowable factor of subjective experience.
- Too many variables. Another problem facing the industry is too many potential variables that could count as indicating a user’s attention. Obviously, the ad needs to be served, but various authorities have suggested different variables, such as number of pixels present, viewability, interactability, and new models, such as paying for the amount of time a video is displayed rather than on an impression basis. These are all good ideas, but how are they all supposed to come together? Will different advertisers opt for different variables? Should we stack all these variables on top of one another and make sure every served ad adheres to them? It’s confusing, and it’s why there’s been no standard model created.
- Flexibility in existing models. Another problem is the layers it takes to integrate such attention metrics into already functioning systems. For an advertising server to integrate new functionality like this, it would require a slight overhaul to the system as it exists today—likely one that counts on a per-impression basis only. Again, because there is no standard format, ad servers need to have a degree of flexibility when they integrate attention metrics—but this conflicts with the typical desire to make advertising as quantifiable and calculable as possible. If there were a more comprehensive or standardized way to do it, more publishers might be on board.
- Early development. Even though some of the latest advancements in attention metric tracking are impressive, we have to remember that it’s a technology still in the early stages of development. Video advertising is constantly evolving and changing as new types of software and faster integrations expand what we can measure. It’s frustrating to see such a potentially powerful technology going underdeveloped and under-utilized, but some of the best programmers of the industry are working on resolving these problems. In a few years, a more mature version of attention metric tracking may emerge to revolutionize the video ad industry.
Attention metrics are in their infancy, which is a weakness in most regards—but it’s also a strength. Attention metrics have a potentially bright future, and most advertisers and ad technology developers realize this. Because these types of metrics haven’t been fully developed or standardized, we have more flexibility with how we choose to develop them. There are hundreds of options, and our newer models of video advertising can be built around them. We just have to keep pressing forward for more advanced forms of attention tracking.Go back