Today Google released their 10th annual Ads Safety Report, which is designed to prevent the malicious use of Google’s ads platforms including those that misrepresent, mislead, or threaten the safety of users.
Their initial report, in 2011, lead to the blocking of over 130 million ads.
In 2020, thanks to 10 years of efforts to clean up the ad space, 3.1 billion ads were blocked or removed.
Of the 3.1 billion ads blocked in 2020, more than 867 million ads were blocked for attempting to evade detection systems and an additional 101 million ads were blocked for breaking misrepresentation policies.
Filtering Out Malicious & Misleading Google Ads Experiences
Google’s analysis helps uncover ordinary-looking accounts that are creating harmful experiences to then filter them out while also creating policies backed by the automation of review signals to help prevent similar instances in the future.
For example, in 2014, Google’s analysis uncovered a vacation rental property scam which helped inform future ad-safety practices.
In 2020, over five million ads referencing the U.S. election were paused while results were being certified.
“The biggest threat and trend that we tracked this year was Covid-19. As part of our work, we stayed abreast of trends around test kits, vaccines, and any misinformation around about Covid-19,” noted Evan Bell, Trust & Safety Strategist at Google.
There were over 99 million bad ads blocked from advertisers trying to capitalize on the pandemic including those for miracle cures, fake vaccine doses, and more.
Ernesto De La Rocha Gomez, Senior Product Manager, Ads Privacy & Safety at Google noted that predicting the future is tough in an ever-changing landscape but it is issues like Covid-19 that help inform Google’s future ad safety models.
Google noted an increase in opportunistic advertising and fraudulent behavior from actors looking to mislead users.
Malicious advertisers are getting smarter, according to Google.
Increasingly, malicious advertisers use cloaking to attempt to hide from detection, promote non-existent businesses, and running ads for phone-based scams to hide from detection or to lure unsuspecting searchers off the platform.
Steps Taken to Improve Ads Safety
In 2020, Google took several new steps toward the prevention of malicious ads by:
- Introducing new policies, including advertiser and business operation verification programs.
- Investing in technology to improve detection of harmful and malicious advertiser behavior.
- Improving automated detection and human review processes based upon an ever-increasing list of network signals, prior account activity, behavior patterns, and user feedback.
The number of ad accounts disabled for privacy policies increased by 70% in 2020, to over 1.7 million.
To prevent future malicious behaviors, Google’s team follows this process:
- Human reviewers identify and label policy-violating ads.
- Engineers create machine learning models using the policy-violating ads to identify signals of malicious behavior.
- The system uses the machine learning models based upon the network signals they’ve identified to flag similar ads.
- Models are continuously improved over time.
A Continued Focus on Ads Safety
Google notes that the work to improve user safety also helps to preserve trust for advertisers and publishers. The work is ongoing and continues to be a priority heading into 2021.
You can read more in Google’s full Annual Ads Safety Report.