Google automates its ad ecosystem, but experts stay critical.

Google automates its ad ecosystem, but experts stay critical.

The Wall Street Journal recently published an article about two people who lost their jobs to AI titled “ChatGPT took their jobs. Now they walk dogs and fix air conditioners.” [link]

The piece is misleading in two ways: first, two examples - of which one is questionable - mean very little. Second, low-level tasks were always at risk of becoming redundant (I don’t want to presume the two individuals mentioned in the article did poor work).

The Copywriter’s Handbook by Robert W. Bly illustrates the four functions of a headline:

  1. Get attention
  2. Select the audience
  3. Deliver a complete message
  4. Draw the reader into the body copy (website, in this case)

AI allows Google to automate all four of these steps. Does that mean copywriters are redundant? No.

The way to think about work in the age of AI is in low and high-level tasks. Low-level tasks are easy for AI to do because they can be standardized: writing headline variations or a/b testing. High-level tasks, like market conducting research or evaluating customer feedback, are tasks where humans excel due to their ability to empathize and pick up subtle signals.

Google has announced several automation features that roll out over the next 12 months across its ecosystem and are likely to take over low-level tasks in advertising - with implications on SEO.

One golden rule of copywriting is that the headline is the most important element of the ad because it catches attention.

Google automates ad campaign creation and testing with AI

Google is in the process of automating and simplifying its whole advertising ecosystem.

From Google (bolding mine):

Last year, we started rolling out automatically created assets (ACA) for Search ads, which use content from your landing pages and existing ads to generate headlines and descriptions. Soon, we’ll be supercharging ACA with generative AI to more effectively create and adapt Search ads based on the context of a query.
For example, with a search for “skin care for dry sensitive skin,” AI can use content from your landing page and existing ads to create a new headline that aligns even more closely with the query, such as “Soothe Your Dry, Sensitive Skin.” This helps you improve ad relevance while staying true to your brand.

Google evolves ACA (Automatically Created Assets) to generate AI images and crawl your site to find images and text that fit better to what searchers are looking for. So far, Google only combined uploaded assets to find the best fit, but the process of generating and testing copy and creative becomes a lot simpler. [link]

Soon, Google generates images and ad copy for you with AI

An area where Google has used automation for almost two years to provide a better experience is SEO titles. Most sites are not aware their titles are being rewritten, and many rewritten titles are worse than their human pendants. It was just a matter of time until title rewrites would spread to ads as well, but Google takes it a step further.

But automating ad copy might work for several reasons.

  1. Google has tons of data about its customers across Maps, Youtube, Android, Gmail, Search, etc.
  2. SGE gives Google even more context because searches use more longtail queries.
  3. Google can a/b test headlines much more efficiently than humans

Ads get better because Google has more context. Google's Performance Max campaigns already test ad copy and creative, but in my humble view, Google is setting the stage to show better ads within SGE.

Google already tests combinations of copy and creative automatically (opt-in)

Performance Max campaigns show ads across the whole Alphabet ecosystem: Youtube, Discover, Display, Search and Maps. Long-term readers remember Platform Confluence, the combination of all platforms within an ecosystem, which is the force at play here.

Google also simplifies advertising in e-commerce:

Merchant Center has been a manual process consisting of setting up a product feed along with all required products, pricing, imagery, descriptions and more. Merchant Center Next will detect all of this from your website and pull it in automatically. There’s good news, though: merchants can edit said products or turn the feature off entirely.[link]

Merchant Next, the evolution of the Google Merchant Center, soon won’t rely on a product feed anymore and pull product information straight from the website instead. Optimizing the product feed is an underestimated way to gain visibility in organic product listings, but it’s also inefficient (a low-level task). By relying on structured data (which becomes a lot more important) on the site, marketers won’t have to worry about the disparity between feed and website data anymore.

Shifting focus from leg work to creative/strategic work

Google’s long-term goal is for you to simply enter your credit card details and let Google handle the rest: target audience definition, campaign creation, campaign optimization and conversion reporting. Ad automation can lead to less transparency and more wasted ad spend, but early tests show good results. [link]

The idea of simplifying advertising to such a low level is attractive to both Google and companies. Small businesses that can’t afford to hire someone full-time or work with an agency to advertise can run ads themselves. The longtail of companies (=small businesses) is much more important for Google than large companies that spend billions of dollars on Google ads. Of the 33 million small businesses in the US (as of 2022), only 7 million are estimated to advertise on Google. The potential is massive. If all small businesses in the US would spend $1,000 USD annually on Google, it would sum up to $33b. And that’s just the US alone.

Automation doesn’t end with making advertising more accessible; it also leads to better ads. Bard gives Google much more context about what users are looking for than a single query because the experience is conversational. Every turn (Ai chatbot request) is a better signal for Google. As a result, Google can display better ads to searchers.

Better ads = bad for SEO

Google ads are a bidding system that depends on context from searches. Since Alphabet still makes the absolute majority of revenue with ads, Google will always prioritize them for monetizable queries. When ads aren't good, users tend to click on organic results. The better ads are, the less of a need there is to click on organic results.

More SERP Features and ads are already leading to organic traffic erosion, especially in ecommerce. If ads get a lot better due to more context, organic traffic might erode even further.

An automated ad ecosystem provides another opportunity for Alphabet: thriving in the cookieless world. Google and Apple are phasing out cookies in 2024, which will hurt behavioral advertising platforms like Meta or Youtube much more compared to contextual advertising platforms like Google or Amazon. Conversations, meaning users indicating their needs through queries, can significantly improve the signal that matches ad inventory with buyers.

All this together - more potential advertisers, better ads, and less sensitivity to cookies - set Google up incredibly well from an ad revenue perspective. Together with SGE, which is much better at answering longtail and top-of-the-funnel queries, SEO is squeezed from both sides.

Where I remain confident, though, is that someone still has to create assets like images or copy in the first place before Google can automate campaign creation and testing. Someone has to talk to customers and identify their pain points before AI can write 100 headline variations in a few seconds. And most keywords don't show ads.

There is a saying that "AI won't take your job, but someone who uses AI will."

The same can be said for tasks: "AI won't replace jobs, it will replace tasks." That's a good thing because it will allow us humans to focus on higher-level tasks that demand empathy, taste, judgment, creativity and a fine-tuned gut.