Purch, publisher of LiveScience and TopTenReviews, added three layers of defense to its anti-fraud system in October. In addition to pre-screening for bots twice, Purch now detects bots, malware and ad quality in real time.
“Better-quality traffic means higher ad performance and higher CPMs,” said Purch CTO John Potter. “This is a trend. Demand sources are getting hard on this, because they are being pressured by buyers. Publishers have to do more to make sure we have good-quality inventory.”
Purch takes extra steps toward protection because its sites include product information and reviews, which attract content-scraping bots. “Because we have so much good product information, we have people scraping us all the time using proxy services or different IPs,” Potter said.
Purch recently discovered that someone had scraped its site Businessnewsdaily.com (which doesn’t even serve ads, using a lead-generation model), hooked it up to programmatic channels and then started sending bots to view ads on the fake site.
“There are bots on your real sites, and also people setting up fake sites to defraud advertisers,” Potter said, adding that Ads.txt (which Purch has already implemented) will help address this problem.
For the first line of defense, Purch’s sites use ShieldSquare. If the tech suspects a bot is visiting the page, users must fill out a CAPTCHA to prove they’re human .
If traffic passes that first test, Purch filters it again using Integral Ad Science (IAS). Purch integrated IAS within RAMP, its proprietary, server-side header bidding solution, which means it only sends out an ad request once IAS determines the traffic is legitimate.
In early October, Purch added a third layer of fortification: real-time bot detection. It integrated another (unnamed) partner to catch bots that slipped through the first two filters. That partner will look for nonhuman browsing patterns. Real-time bot detection also serves as a second line of defense for the 20 non-Purch pubs using RAMP, but not ShieldSquare.
Once ads serve on the page, Purch works with yet another set of vendors to screen for malware and ad quality.
The malware detector will block ads if it observes a redirect – important given that malware operators often run ads on weekends and holidays, when ad-ops employees are likely to be out.
This ad quality system stops autoplay video and expandable ads in their tracks.
When Potter’s team tested the ad quality system two weeks ago, CPMs went up 3.5% when Purch blocked the ads. “These annoying ads aren’t even paying the high CPMs,” Potter said.
Revenue went down 1%, primarily because Purch doesn’t get paid for placement of ads it blocks.
But little blips in revenue don’t affect Purch, which is playing a long game.
“The well-behaved ads paid more,” Potter said. “It’s not going to cost you revenue in the long term, because you are going to end up with better-performing inventory, and that will raise the CPM.”
This post was syndicated from Ad Exchanger.
More Stories
Around the World: AI & Christmas, a Temu crackdown and Aussie influencers
Here’s a Short Roundup for This Week
Melanie Spencer a finalist in Campaign’s Agency of the Year Awards