The Publisher Lab: Meta’s expensive error, Google to use generative AI to make ads, and Google retires 4 ranking systems

The Publisher Lab is back, and this week we are talking about Meta’s expensive error, Google creating ads with generative AI, and Google retiring a few ranking systems.

If you would prefer to watch the podcast, check it out on YouTube. You can also listen to it wherever you can get podcasts.

Bug in Meta ads overspends advertisers’ budget

Meta recently had a bug in their ad code that caused ad campaigns to spend past their daily caps, costing some advertisers hundreds of thousands of dollars. Most of the overspending was through Facebook via theAudience Network and a small amount on Instagram.

Thousands of Meta employees have been laid off in the past few months, many of which were in account services and technical platform operations, so customer support is slow to those inquiring about the bug.

The advertisers affected by this bug were contacted by email from Meta but the email didn’t provide much concrete information or resolution. Meta implied in the emails to CEOs that advertisers can expect refunds, just not for the entire spend; impressions that were legitimately served and were within the daily cost caps are unlikely to be refunded. Other than the email, there had been no other contact or statement made by Meta (as of Monday, April 24, 2023).

The refund package is likely to be hundreds of millions of dollars.

This bug is in addition to a long series of platform bugs for Meta advertisers over the last year or so. This additional bug certainly doesn’t help Meta’s reputation or credibility. In the earlier days of internet advertising, the world was basically composed of Facebook and Google, with a few others fighting their way in. Now, while Meta still has a leading ad service, it is on a dying platform and experiencing technical difficulties.

Google Ads to use generative AI

Based on an internal presentation for advertisers, it was learned that Google is planning to use AI to generate unique ads. The AI will pull from materials provided by real marketing professionals.

The new technology will be integrated into Performance Max, Google’s algorithm-driven ad placement and copy program.

The generative AI ads will allow advertisers to submit creative content related to a campaign and the AI will create ads that are targeted at specific audiences in order to meet more obectives.

What’s interesting about AI is that for a long time, we have been aware of its ability to replace many people in jobs that are considered more blue collar but involve telecommunications or technology, like call centers. Many, however, did not see AI’s ability to take over creative jobs, like ad creation. And, since ads are a bit formulaic and benefit from testing different combinations and phrasing, using AI to create them makes sense.

There are a few concerns with using AI to create ads, though they certainly offer a lot of potential for improved conversion rates. We already know that AI will sometimes break the rules or give false information in order to carry out a request. In this instance, if AI is optimized to convert new customers, it could spread misinformation or use something copyrighted if it thought it would convert the user.

It seems like Google is eager to get ahead in the AI race and this new use of AI is a reach to incorporate AI wherever possible.

Google retires four ranking systems

Recently, four ranking updates/systems were retired: page experience, mobile-friendly, page speed, and secure sites.

For publishers who wonder what this means for the page experience update from 2021, Google reports that the update was never a separate ranking system, but more of a concept to guide publishers in what to focus on.

John Mueller of Google said, “we’ve seen people hyper-focus on these numbers, it’s not a good use of time and energy.” For any of our listeners or those who keep up with Ezoic blogs, you’ll have heard varying Ezoic employees from over the years say this exact thing.

Before Google ever mentioned anything about site speed, it was a vanity metric between developers. But then, back in the early 2010s, Google began emphasizing the importance of site speed when apps were really becoming popular. Google was working on greater connectivity in places around the world where internet wasn’t as developed and downloading a site or app was expensive and slow. By having publishers focus on site speed, this automatically opened up the world wide web to internet users in those hard-to-reach places.

One of the first issues with the way that site speed was introduced is that Google attached a score to it, which automatically made all publishers obsessed with attempting to reach that target.

It’s likely that we will see a slimmed-down version of Search Console in the near future with the retirement of these ranking systems.

Comments, questions, concerns?

If you have something you’d like to tell us, head to PublisherLab.org and drop us a line. You can also leave us a comment on YouTube.