Data from BidSwitch exchange shows video inventory is lagging behind display for ads.txt adoption, and that relatively high numbers of advertisers are still choosing to buy from unauthorised sellers. But more broadly the IAB’s ads.txt initiative is increasingly being adopted by publishers — and warmly received by buyers are showing they’re happy to pay more to buy from authorised resellers.
BidSwitch has been one of the companies leading the charge on supporting ads.txt adoption, and has been monitoring data on traffic coming through its exchange to see how publishers and advertisers are adapting to the tool. The vast majority of traffic coming through BidSwitch now carries an ads.txt file. Data logged last month showed that 72 percent of total traffic carried a file, compared to 56 percent last November, a significant change in a short period of time. These findings are similar to data released by Pixalate on Monday, which found that 71 percent of the top 5,000 programmatic publishers now carry an ads.txt file.
But BidSwitch’s data shows that as well as being less likely than display to carry an ads.txt file in the first place, video inventory is also more prone to being sold by unauthorised sellers (i.e, sellers that aren’t listed on a publisher’s ads.txt file).
BidSwitch say that over 55 percent of video inventory from publishers which do have ads.txt files is being sold by unauthorised companies, whereas for display inventory this figure is only 12 percent, meaning that overall, only 26 percent of all video traffic online comes from authorised sellers, much lower than the 67 percent for display.
Barry Adams, BidSwitch’s VP of commercial development, said that the low proportion of video traffic may be the result of a lot of parties reselling display banner inventory as video. In-banner video (IBV) is a format used by resellers who buy up banner ad inventory for a low cost and sell it on as video inventory, meaning the advertiser is misled about the type of inventory they’re buying, and the publisher ends up with poorly running ads on their domain.
As mentioned above adoption for publishers selling video inventory is also lagging behind display, and growing at a much slower rate. Last November 57 percent of video traffic carried a file, which has grown to 59 percent in March, compared to display which grew from 60 percent to 72 percent in the same period.
This may be due to extra difficulties with implementing ads.txt for video publishers. Oath published a blog post earlier this year outlining some of these difficulties, saying that the syndication and distribution partnerships associated with video content mean that a video publisher has to conduct a full audit of these partners before applying ads.txt. Oath gave an example of HuffPost allowing a broadcaster to distribute and monetise HuffPost video content on its site: if HuffPost hasn’t listed all of the broadcaster’s authorised sellers on its own ads.txt files, then they might appear to buyers as unauthorised.
While it may be difficult to implement ads.txt for video, it seems particularly problematic given the results of tests by Google, Amobee and Quantcast last year which found that video ad inventory was much being sold fraudulently at a much greater percentage than display. A study of 16 large programmatic publishers including The New York Times, The Washington Post and Mail Online found video callouts were overstated by 57 times the available inventory, compared to display callouts which were overstated by four times the available inventory. BidSwitch
VAN reported on a fraudulent IBV campaign last month first exposed on a Reddit thread, where a couple of users claimed that simply holding ads.txt files isn’t enough to stop this behaviour, but rather that publishers should cut off all resellers completely and only work with direct partners.
BidSwitch’s data suggests that advertisers too are less trusting of resellers. They are happier to pay more for authorised inventory, as eCPMs for authorised traffic are on average twice those of unauthorised traffic. But reseller traffic commands a significantly lower eCPM than direct; back in November, direct eCPMs were 44 percent higher than reseller eCPMs, though this gap has now closed a bit with reseller eCPMs now 15-20 percent lower than direct.
