Bad ads, Malvertising, Malicious advertising: Whatever you call it, the problem of programmatic ads being used for purposes other than advertising is an ongoing issue for web publishers. As publishers we are judged not only by our content, but by the ads that we run to support that content. When users get malware warnings on our sites, or are the victims of mobile redirects, it is the publisher’s reputation that suffers.
One piece of advice that many publishers receive about tackling bad ads, such as forced redirect ads, is to raise their floor prices. Price Floors are fixed rates set by publishers which determine the minimum price that advertisers must pay for certain ad inventory. The theory is that bad actors bid low and that setting price floors will price malicious campaigns out of the market and keep them off your website.
We’ve never been convinced by this argument at OKO. On a theoretical level we see no reason to believe that malicious ads will always be associated with low bids. If these scams were not profitable then the people behind them would be less attracted to them. If they are profitable and scale-able then why bid low? On a more evidential level, we have also only seen limited success in using floor prices to tackle bad ads in the past.
Now it seems that there is broader data that supports our own experiences of this problem. Confiant, who specialise in monitoring and policing ad creatives on behalf of publishers, have just released their latest Demand Quality Report that sampled over 100 billion impressions to identify patterns in malicious advertising. One of the data points they analysed was the CPM of the malicious ads they detected.
Whilst this shows that you can stop a lot of malicious ads by raising floor prices, it also shows that the loss of revenue from doing so is significant. Based on their Q1 2019 data, 80% of malicious ads could be blocked by enforcing a floor rate of $0.63, but this would also result in 70% of all other ads being blocked. Few publishers are in the position where blocking 70% of ads is an acceptable outcome, suggesting that floor prices are not an effective tool for stopping malicious ads.