Study finds under 50% of its impressions viewable and valid

All marketers measured in a new bot fraud study from the ACA had invalid traffic in their campaigns.

A new study from the ACA and fraud detection company Sentrant found that only 44.8% of all the impressions it measured were both viewable and provably valid.

The remainder of impressions studied were either totally invalid (6.4%), human but unviewable (31.3%) or human but un-measurable for viewability (17.5%).

The largest sources of invalid traffic were off-screen browsers and non-human data centre traffic, which accounted for more than half of all the invalid traffic found in the study.

All seven marketers included in the study of more than 235 million impressions were affected by invalid traffic in their campaigns, which ran across multiple publishers, platforms and formats from February to December 2016.  The participating campaigns ranged in the amount of invalid traffic they saw, from a low of 4.5% to a high of 21%.

Canadian news sites had lower fraudulent traffic numbers than the average (1.2% to 2% versus the average 6.4%).

Chris Williams, VP of digital at the ACA, said his organization’s recommendations for battling bot fraud largely comes down to verifying data and knowing who you’re working with.

“Pay for verification services,” he said. “If you’re worried about the cost of verification, then you should be worried about the cost of fraud. And then you want to make sure that data becomes yours. [Marketers] should investigate a direct verification service. The contractual part of who owns the data is a big deal, and we want to see advertisers own it. Think of the verification data as information for an audit, so advertisers want all the information about what’s been at play. You can only audit the information you can see.”

Brands are increasingly seeking to get money back after a campaign has been affected by invalid bot traffic. Williams said adding time into contracts to measure the numbers is key to making that happen. “If the billing cycle is 45 days, why wouldn’t you take at least a week to look at the numbers before you generate bills?” he said. “Take your time and do a good job and understand the amount of the media exposure. The MRC (Media Ratings Council) has put it in into their guidelines. Use the time that you have wisely.”

Image: Shutterstock