Bots create close to half -- or about 48% -- of all internet traffic across all web sites, according to Reid Tatoris, Distil Networks vice president of product outreach and marketing.
And while that number was published earlier this year, he underscored the enormity of fraudulent activity following the bot scheme that Google and White Ops, with the Federal Bureau of Investigation, collectively worked to discover and then reveal this week.
Tatoris said the $29 million lost from the scheme pales in comparison to the totality of the problem.
“During Black Friday and Cyber Monday ecommerce sites saw a huge spike in bots,” he said. “One reason is companies write bots to scrape the prices of their competitors.”
These bots will hit a product page many times per day to scrape the information on the page. Each time the bot scrapes the page, it triggers an ad impression that counts against the total ad impressions served on the retailer’s site if it has an ad on the page, he said.
The Interactive Advertising Bureau (IAB) estimates online ad fraud costs advertisers $8.2 billion annually. Schemes like the one uncovered earlier this week by Google, White Ops and the Federal Bureau of Investigation increase the amount of bad bots, which Distil Networks estimates accounted for about 21% of web traffic in 2017, up 9.5% compared with the previous year.
Ad fraud is only one use for bots. “And even if we caught all the purposeful ad fraud, we would still have massive amounts of waste because all the bots purposely trying to create fraud cause an impression that an advertiser must pay for even when no one is on the page,” Tatoris said.
Tatoris said it’s not about catching every bot operator, because that’s not possible, but rather companies should be thinking about how to build and implement technology that would analyze every ad impression to identify whether or not it being served to a human being.
“That’s not being done,” he said. "Everybody wants to point a finger. This is a side effect of the way the ecosystem is built."
The huge implication of viewing every ad is that revenue from the impression would disappear, he said.
While I fully understand that bots and programmatic ad delivery is here to stay, one thing has changed for the worse that receives no press. Recently one of the very last "Ad Networks" went out of business. For those who never worked with an Ad Network, they were a one on one relationship between company distributing ads and the publishers who placed the ads on they websites. Because there was no computer program involved in the handoff between the ad distributor and publisher, there was zero ad fraud. the bottomline reason ad networks failed is because we have been told that we could trust machines, wires and computer programs more than than publishers and real humans. I know after major ad distribution changes in the past 6 years and the publishers getting hurt the worse of all in the chain, the publishers would be willing to work with advertisers to place their ads with security and location in mind. Furthermore at a very fair price that is better than having dollars stolen from them.
Exactly, Craig. The new kind of "ad network" which I'm wondering about, would seem at first glance like the ones you mentioned that are fading away due to the stupidity of advertisers who think they are getting a better deal if the transactions are handled by computers. As is now painfully evident, once all of the negatives are accounted for---ads not being on screen or sufficiently viewable, ad clutter and disruptive placements, fraud, excessive "tech tax" charges---cutting into the publishers' profits, etc, etc.doing it all the automated way is fraught with problems. However, unlike the original digital "ad networks", the new one---or ones---would have firm advertiser support, though I don't see the advertisers setting these up themselves. As for definitions, metrics, etc. there is no problem getting concensus agreement in TV, providing all of the adbvertisers are of the same type---mostly branding or mostly DR----and they have finally seen the wisdom of planning and buying their media using humans, aided and abeted by data and paperwork processing, but not dominated by it.