If you run online ad campaigns, you might want to sit down.
That’s according to the first “Robot Traffic Report” (free, no registration) from New York City-based financial native ad platform Dianomi. The data comes from hundreds of publishers around the world on the company’s ad platform, over a period of six years.
The good news: The firm only detected 32 percent of robot traffic for ad clicks overall in the past year, down from last year. But, as recently as April of 2017, that figure was 85 percent.
Bots aren’t all bad guys. Search engine bots, for instance, are benign robot traffic. But that still means you’re sometimes paying for non-human traffic. Here’s how the report breaks down the overall traffic:
CTO Cabell De Marcellus told me via email that US-based publishers have the smallest percentage of bots in their overall traffic, with a large percentage coming from aboard — especially from Ireland, “where nearly all the traffic is bot traffic.”
Some of the higher percentages come from publishers, the report says, who are “buying traffic from questionable sources and consequently delivering more robot clicks.”
There are a variety of reasons why bot traffic is increasingly looking like human traffic, one of which is that 75 percent of the fraud (according to a separate 2016-17 study) was sourced from computers used by both a human and a bot, making it more difficult to filter out traffic by source. Sometimes, the fraudulent clicks are spread out over a longer time frame and several IP addresses, which also increases the difficulty of detection.
As a defense against the robot hordes, Dianomi suggests several key strategies. These include directing good robots to what they should and should not click, employing an outside vendor like DoubleVerify or IAS to analyze your click traffic and avoiding high-risk IP addresses.