Google Removed 100 Bad Ads Per Second in 2017—How’s That Hurting Ad Revenue?

Google removed four times as many bad ads in 2017 than it did two years ago. Illustration by Kaitlyn Flannagan for Observer

No tech company (except Facebook) relies more on advertisers to generate income than Google. After winning more than 40 percent of the U.S. digital advertising market, however, the search giant is now blocking advertisers at an unprecedented speed.

In 2017, Google removed 3.2 billion ads that violated its advertising policies, the company said in a blog post on Wednesday. Google declined to disclose what percentage of total ads they had sold, but the number of ads removed in 2017 doubled that of 2016 and quadrupled that of 2015.

At that speed of ad moderation, Google is able to block the majority of bad ads, including fraud, malware and phishing scams, before they reach intended audiences, Scott Spencer, director of sustainable ads at Google, wrote in a company post.

Three billion ads a year translate into more than 100 ads being removed per second. How did Google do that? And, more importantly, does that affect ad revenue, which is virtually all of Google’s revenue?

At such a pace, it’s not surprising that automated detection tools plays a major role.

“Machine learning plays a big part in helping us detect policy violations at scale,” a Google spokesperson told Observer.

Google has 180 filters just to catch invalid web traffic and fraud. And, for specific policy violations, such as “clickbait,” ads that trick people into installing unwanted software, and ads that direct people to malware-laden websites, Google has individual filters designed to catch those particular types of ads.

Humans are still involved, nevertheless.

Google employs thousands of people in various functions within the ad team, ranging from investigators to policy experts, as well as day-to-day reviewers “who are waging a daily fight against bad actors and abusive behavior that tech can’t always catch,” the spokesperson said.

Google declined to reveal further details about its ad detection process, fearing that sketchy advertisers will abuse the system and find ways around it.

Google sell ads through AdWords, an ad auction system to display ads in Google Search, Gmail and Youtube, and AdSense, a program that allows third-party websites to display Google ads on their own sites.

A troubling trend Spencer noticed in 2017: An increasing number of these third-party publishers are either creating scam sites (for examples, a non-news site in the U.S. pretends to be a news site based in London) or stealing content from other sites just to gain quick web traffic and ad money.

We paid $12.6 billion to publishing partners in our ad network last year. But in order to make money from Google ads, you have to play by rules—that means respecting the user experience more than the ads,” Spencer warned in Wednesday’s post.

Advertising accounts for the absolute majority of Google’s revenue, but as it tightens ad policies and the moderation process, the company is looking to generate income from non-advertising product lines, such as broadband service, tech products and even an investing-for-beginners course, to reduce reliance on advertising.

Progress is taking place, but steering the profit model for a $100 billion company takes time. In 2017, advertising made up 86 percent of Google’s total sales, company financial statements show, slightly down from 88 percent in 2016 and 90 percent in 2015.

Google Removed 100 Bad Ads Per Second in 2017—How’s That Hurting Ad Revenue?