Google Unveils AdSense Auto-Adverts Utilizing Machine Studying to Assist Make Placement and Monetization Selections – TechCrunch
Google today unveils a new ad unit for AdSense that takes advantage of the company’s great urge to add more artificial intelligence to its business and potentially attract more publishers who may be considering stepping up their advertising efforts but don’t have the time or other resources to to manage them.
Google introduces “Auto Ads” – not advertising for cars, but a new ad format that uses machine learning to “read” a page to see and place what types of ads might go there, including where they might be should be placed and how many are running. Publishers activate Auto Ads with a single line of code on the page.
The service was actually quietly announced and launched in a limited beta version in September 2017. Now it is available to everyone. Google tells us that “Publishers who participated in the beta saw an average increase in sales of 10 percent, with sales increasing between five and 15 percent.”
If you follow or use AdSense, you will know that the service already has a reasonable level of automation. The product is used by tens of millions of web publishers to indicate where to place ads (banners and other entities). These ads are then selected by Google based on a crawl of the page to find out which ad may be most relevant. It already makes up a significant portion of parent company Alphabet’s advertising revenue, which accounted for $ 27 billion of the $ 32 billion last quarter.
What’s new with Auto Ads is that Google does the placement selection and does all of the work for publishers to figure out how many ads to show on certain pages, where to place them, and what kind of ads are going to be running.
The use of machine learning is interesting here because it not only helps determine where an ad is going, but also provides analytics on how that ad is performing to teach the system how to better serve ads in the future.
One black hole (and potential danger) is the fact that Google’s automatic ads seem to decide how many ads to show on a page – something you would have had more control over without them. This thread on Webmaster World details how some of the early beta testers were unhappy about how many ads were crowding their pages and what that meant for the user experience on the site.
We ask Google for an answer on this point and whether users can limit the number of units that Auto Ads can place on a page. As far as we know, this will likely be one of the aspects that will be assessed at launch.
It also highlights questions about how good Google’s judgment will be in each case.
The AdSense service came into the spotlight because a lot of nefarious content got into the mix, including ads with “fake news” and other misleading content. The company has made efforts to combat this. The Bad Ads Report published in January 2017 found the company shut down 1.7 billion dodgy ads and banned 200 publishers from AdSense.
For now, the goal seems to be to put this in place and see how many sign up for the convenience of the service that you activate by signing into your AdSense account. Check the global settings under “My Ads”; Copy and paste the existing code and paste it between the header tags for each page you want the ads to appear on (they’ll appear in 10 to 20 minutes, according to Google).
In a blog post by AdSense engineering manager Tom Long and product manager Violetta Kalathaki, the two units contained in the auto ad mix contain anchor and vignette ads as well as ads for text and ad, in-feed and matched content. (Not clear if newer formats like this larger banner will also be included.) You also write that publishers can specify which of these formats to run.
For those who have used page-level ads (showing different types of ads based on the subject of a page, rather than an entire website), all of the code is automatically migrated to serve with automatic ads. And for those using Google’s AMP service for mobile pages, you will need to use code for AMP Auto Ads.
Updated with more comments from Google.