Two reports, released this week and last, demonstrate that marketers often do not have enough visibility into where their ads are going and whether the ads are seen by humans.Krisztian Bocsi/Bloomberg
Groups representing advertisers in both the United States and Canada are urging companies to demand greater transparency from the agencies that help them buy ad space across the digital landscape.
Two reports, released this week and last, demonstrate that marketers are often not getting their money's worth in digital ad buys, and do not have enough visibility into where their ads are going, whether they are seen by humans and how their money is being spent.
On Tuesday, the Association of Canadian Advertisers released a study measuring "invalid traffic," or occasions where views of an ad didn't come from humans but from bots on the Web – sometimes malevolent bots that actively try to leech marketers' dollars through fraud.
The study monitored 33 digital ad campaigns last year from seven companies (who participated anonymously), and watched for bogus views of those ads. It looked at more than 235 million "impressions," or instances in which an ad appeared on a page. Sentrant, one of the firms that have popped up to analyze such traffic and root out fraudulent activity, conducted the sweep.
All seven marketers were affected by "invalid traffic" for their ads – 6.4 per cent of all "impressions" or potential views of an ad were not real human audiences. The participant with the highest invalid views lost more than one-fifth of their traffic to bots. When the audience was human, 31.3 per cent of the ads weren't "viewable" on the page – a standard defined as half the pixels in an ad being visible for at least one second.
Only so many advertisers could participate, because they either didn't have rights to the data on their campaigns to do an analysis, or weren't sure they did. There were other legal or logistic challenges. But while the sample is too small to use the numbers as a proxy for what is going on in the giant digital media ecosystem, it did represent large, well-known and sophisticated marketers, the ACA said. So it does give a view into the problems that exist across the digital ad economy.
Where the ads ended up had some impact: Many Canadian news sites had less than 2 per cent invalid traffic.
All websites will have some level of non-human traffic. That's because bots will travel around to legitimate sites as part of an attempt to look as if they are human Web visitors. There are also bots and "spiders" that exist not for fraudulent means, but crawl the Web to gather information. The Interactive Advertising Bureau (IAB) keeps a list of benevolent spiders and bots to help marketers and their agencies detect them as non-human traffic (but also to know they are not there to commit fraud.)
"We hear people say things like, 'Don't worry, it's priced into the overall cost,'" said Chris Williams, vice-president of digital for the ACA. "Let's put it into the contract … everybody agrees that if we discover [invalid traffic] we're not paying for it. And as we know, it's everywhere."
The ACA is pushing for marketers to build better accountability into their contracts with their media-buying agencies. That should include specifying a "white list" of acceptable sites where ads can appear, setting goals for legitimate traffic and mandating that they will not pay for non-human traffic. It also advocates for publishers of websites to make their ad spaces more open to measurement.
The biggest sources of invalid traffic in the study were "off-screen browsers" – windows of an Internet browser that can open up, invisible to the user, and conduct bogus activity on the sly; and data centres that operate bot networks and other fraudulent programs.
The need for transparency was echoed by another report released last week by the Association of National Advertisers in the United States in collaboration with the ACA and other firms. That study looked at more than 16 billion "programmatic" (or computer-automated) ad buys in 445 campaigns, on behalf of 30 brands owned by seven major advertisers.
This kind of buying has been growing more popular, largely because it's more cost-effective and can be done faster and at a wider scale than human buyers can. It is also lauded for allowing better targeting of ads to relevant audiences. But the process also introduces middlemen in the process of bidding for ad space, and affects marketers' ability to know where their ads end up – especially since some agencies use non-disclosed methods in exchange for giving marketers a better deal. A majority of digital ad space is bought this way.
But there is what the associations call a "tech tax" for that efficiency. In the programmatic transactions the study looked at, 42 per cent of advertisers' money was "consumed by supply chain data and transaction fees," with the rest going to actually buying ad space.
"Based on the range of transparency roadblocks encountered, advertisers may find it difficult to manage, measure, and audit their programmatic media investments with the same rigour as they manage traditional media investments," the report stated.
Buying arrangements where information is not disclosed to the advertiser are widespread, the report said. Even when all information is not available, marketers need some way to verify the effectiveness of their spending.
"It's much easier to deal with problems like this when there is transparency. The more fog there is, the easier it is to get away with nonsense," Mr. Williams said. "… Once you start to see in and have a picture of what's going on, you have to hold people accountable."
Susan Krashinsky Robertson