By Amit Joshi on June 6, 2017
Originally posted by Ross Benes on Digiday.
Two of the big trends in digital media aren’t compatible: The drive to enforce viewability standards and the shift to mobile, particularly apps.
Ad buyers say they receive misleading mobile viewability reports because measurement vendors struggle to integrate into publishers’ apps. In effect, they know the figures reported aren’t true. Until more pubs adopt third-party checks against their app inventory, buyers should be aware the mobile viewability scores that get reported to clients are unlikely to represent the campaign’s true metrics.
“It has generally been accepted that there is not a real simple way to measure in-app,” said Lee Garfield, director of programmatic sales at gaming platform Zynga, which drives about 80 percent of its usage through apps. “And rightfully or wrongfully, people have given the verification companies a pass because of that.”
Katie Farmer, associate media director at 22squared, began questioning the in-app viewability numbers vendors provided her once she noticed the measurement companies were only measuring 5 percent of the impressions across some of her clients’ mobile campaigns. But 22squared doesn’t consider a sample valid unless it includes at least 30 percent of a campaign’s impressions.
“They’ll still give a [viewability] number, but it is misleading,” Farmer said.
The reason only 5 percent of the mobile impressions got measured is because the vendor didn’t have access to measure publishers’ app inventory. Since the campaign was broadly set up as a mobile campaign and didn’t differentiate between app versus mobile web, the vendor report generalized the campaign’s metrics from the few impressions that came from mobile web.
An ad buyer requesting anonymity said it had a mobile campaign that averaged only 35 percent viewability because much of its app buys were rated as non-viewable. This was problematic for the buyer because the brand client, which requested the agency to buy viewable inventory, perceived that the agency was buying remnant inventory and not following the brand’s instructions. These kinds of stories are troublesome for the overall industry because users now spend 58 percent of their time with digital media on apps, according to comScore.
Ad buyers concerned about in-app measurement issues can set up their deals with vendors to split out app and mobile web. But doing so requires the buyer to implement two sets of tags and monitor an additional data set for each client. It also demands that the buyer be cognizant of their vendor’s limitations in-app.
There are two ways for verification companies to measure in-app viewability, and each has limits, said Amit Joshi, director of data science at verification firm Forensiq. One way is for vendors to get publishers to adopt their software development kits, which allow third parties to integrate their features. The constraints on this are that each vendor has its own SDK, and each SDK causes the app’s size to grow, which prompts users to have to download updates. The publisher also has to deploy a developer to integrate the SDKs, and implementing SDKs can contribute to latency.
The other way is for vendors to drop their tags onto the publisher’s mobile pages, Joshi said. But this tactic only works in apps that pull their articles from the mobile web. Many publishers do not use this approach because natively uploading articles within the app cuts page-load times. So even if a vendor has the technology to accurately measure viewability in-app, the tech will only be useful to ad buyers if publishers structure their apps to give vendors access.
“If you run verification in a browser, you are in a universal system, regardless if you are tracking CNN or The New York Times,” said Rob Auger, vp of media technology at DigitasLBi. “But in-app, you are tied to the way the app is developed.”
DoubleVerify, Moat and Integral Ad Science declined interview requests for this story, but these companies are well aware of the challenges that plague in-app viewability measurement. Hoping to spur publisher adoption of in-app verification, IAS developed an open-source SDK that other vendors can use. The idea is that by giving publishers a single SDK that is flexible between vendors, pubs won’t have to upload as many SDKs, which will increase their likelihood of implementing third-party verification in their apps. While the initiative may seem promising, once again, its power is limited without adoption from the supply side.
The clumsy state of mobile viewability is a real problem for advertisers given that in-app ad spend in the U.S. reached $33.1 billion last year, which was greater than the $24.9 billion that was spent on desktop inventory, according to eMarketer. Although mobile consumption continues to grow, Farmer said her agency may “have to pull away from in-app inventory if we can’t track it.”