Lately we’ve heard a chorus of skepticism regarding the importance of viewability, and some say that there is no correlation between viewability and conversion rate.
In reality, there are only three reasons why one could legitimately argue that viewability doesn't matter:
Let's start with the first point: This would only be the case in an "efficient market" in which there was extensive information available on every impression. This scenario also assumes complete transparency, wherein buyers and sellers all had similar information. Unfortunately, this is not today’s reality, nor will it be for some time.
So, let's move on to the second and third reasons.
Imagine walking through a busy intersection in New York City and seeing a billboard that has been turned backward and is now facing a wall. Because it is facing the wall, you can't see the ad.
Now, if it doesn't matter whether this billboard is facing the wall or facing the crowded city street, then we are saying that the message on the billboard has no effect on any of the thousands of people who pass it daily. In other words, the ad has no impact on the viewer’s opinion or consideration and the advertiser is wasting money on advertising in general, since billboard advertising -- or this ad in particular -- simply doesn't work for its brand whether it’s seen or not. Or perhaps the method for measuring the effectiveness of this advertisement is incorrect.
As it turns out, the latter is a problem in the industry. Most of our attribution models assume all online advertisements are seen, regardless of where they appear on a Web page. As we are learning, this is a dangerous assumption. Worse yet, many people using these attribution models put a premium on quantity (e.g., number of "view throughs") rather than quality views (e.g., was my ad actually seen?).
Imagine two identical billboards side by side, one facing the wall -- one facing the intersection. Now imagine that we could track purchases for everyone who walked by the billboard over the next month. If we were using the online attribution model, we would treat all ad exposures the same way, regardless of whether the ad was facing the wall or facing the intersection. And in doing so, we would attribute the same amount of value to each billboard.
We know that this is crazy, because one of the billboards wasn't actually showing an ad. And worse yet, we underestimate the value of the billboard actually displaying the ad. Instead of giving equal credit to both billboards, we should give 100% of the credit to the billboard that was working and 0% to one that wasn't.
The majority of attribution, conversion calculation or ROI models online do not take viewability into account. And yet we know that on average, between 50-60% of ads are in view for less than a second. This means that only 40% of impressions even have the potential to be viewed (i.e., are ever on the user’s in-view screen, or “viewport”), despite the fact that they still may not be actually viewed based on other factors -- the consumer not looking at the browser at the time, or too many ads on the page.
The result of these outdated measurement models is that we are often giving credit to the wrong media suppliers. Fortunately, we are finally coming together as an industry to agree on how viewability should be measured. We are taking steps to ensure accurate data for these calculations. It’s time to add viewability to our success measurement methodologies and reward quality over quantity.
However, despite the industry’s fascination with this topic, it’s important to note that even if all ads are in view, it’s not at all the final word on ad effectiveness. In fact, viewability is merely a symptom of a broader issue surrounding media quality or the lack thereof. Although it is important, viewability is only one of many media quality considerations that are transforming traditional media buying for the programmatic world.