Whether an ad is in view or not has become a hot topic over the last 18 months, and rightfully so. There is no reason why advertisers should pay for ads that are not seen. However, there are two parts to this debate: one is viewability, and the second is fraud. While we should worry about both, advertisers and agencies have begun to realize that viewability is actually far less worrying, and more manageable, than fraudulent activity. Since I've been advised not to use the “F” word by my legal team, I’m going to refer to it as "suspicious traffic" from here on, so the lawyers don't shut me up.
Ad exchanges, which have the largest, most liquid pools of ad inventory, are where it’s easiest for bad guys to lurk and hide – though they can be found everywhere. Recently, we've seen some very scary statistics – whether that's because we're getting better or the problem is getting worse, I couldn't tell you, though I suspect the latter. One thing I can tell you is that not all exchanges are created equal. For example, up to a third of all ad impressions on some exchanges are not generated by people -- whereas on others, "suspicious traffic" is below 5%. One-third to less than 5% is a big delta. Thus, which exchange you use makes a big difference. I suggest looking into the tools and services out there that can identify this traffic and put them to use.
At the IAB Leadership Conference (shout out to Randall Rothenberg for a great event), I was engaged in a session about viewability. At one point, someone had the temerity to suggest that knowing whether an ad is shown or not, ALWAYS matters in EVERY campaign -- not only branding, but also in performance and direct response/lower funnel campaigns. The response from well-respected players was quite interesting, as they claimed that this isn’t true. Who cares if the ad is seen as long as you get results? That's right; buying an ad that is never viewed by a person can make you #1 on some plan,s because the technology is designed to mimic the results you are looking for. Even when you are tying it to outright purchases, "invisible" impressions can give you view-through conversions.
Here are some common tactics that you have fallen victim to if you’ve bought impressions on RTB:
1. Purposely serving ads (sometimes many, many ads) into a 1x1 pixel.
2. Viruses that run on computers and open websites on hidden windows so that a real person appears to be visiting web pages and seeing ads.
3. Bots that run on websites to mimic people – some bots also click on ads and show up on client sites. Examples can be seen by comparing large network volumes available by some players in RTB and cross-tabulating them with comScore metrics, whose panel’s population is 100% human.
While it’s important to understand how we have become victims, figuring out what we can do to improve the situation is even better.
Focus on attribution: If you run sophisticated attribution models, which ads are working vs. which ads are not will become apparent. In the end, bots don't buy widgets. With the help of Google, Adobe, C3 Metrics and Adometry, we've been running a series called Attribution Revolution to help educate marketers.
Use viewability products Most viewability products really work.
Find credible exchanges and reward the good guys. If you are a buyer on RTB, figure out which exchanges contain more “suspicious traffic” and move your buys to those that have less. Explain the issue to the suspicious exchanges; if the problem is fixed, you can always move traffic back.
Don't believe the unbelievable. If the biggest players in the industry cannot come close to matching the results of some of the smaller players on your campaign, then those results are probably too good to be true.
Be suspicious, ask questions and begin to understand these issues, as they aren’t going away anytime soon.