-
Recent changes in Google Search and further implementation of AI
Hey,
Before I go into what’s this post about and why I’m making it, I guess I’ll give some background. I work in a rather large affiliate-marketing company that builds (in-house) comparison websites (mostly “top 10’s”) in both large and medium markets, mostly in the USA and UK.
We’re profitable in many of our larger verticals (500k+ monthly spend). I’m a team leader in the financial sector and been doing PPC for the past several years.
So. I’ve attended this Google “Search Champions” event thingy. They were obviously talking a lot about automations and AI. After talking a little about the improvements in the current “automated” or semi-“automated” features, like the Broad match type, P.Max, Maximize Conv./ROAS bidding and RSA ads, they’ve dived into how Google Search is going to change in the near future, with the broader implementation of AI.
Changes I found the most intriguing:
1. **There’s no need to go granular anymore**. My company’s been priding itself on how granular and specific we get when creating ad groups and campaigns; very slim with a rather small amount of keywords, so the keyword insertion within ads will be more accurate, and so our campaigns will be more “easily” adjusted (in regards to target CPA or ROAS). They claim that stacking up keywords with the same (or even somewhat close) intent will not only increase traffic, but also provide better results; I guess they’ve been saying this for a while now, but this time they seem adamant, and provided some examples that kinda made sense. We’re now trying it company-wide. ROI’s are looking good. The value for our affiliates – not so much, but whatever.
2. **How people search stuff**. They’re pretty sure that queries are going to change from a few or several words in a search-term into actual questions or even ChatGPT-like prompts (that usually end up being longer than 10 words). I guess what this means is that the Exact match type will become somewhat obsolete. For context, right now – 90%\~ of the traffic in all our verticals is still Exact). You can’t add keywords longer than 10 words into Google’s editor. Thus, in order to get your ads triggered by these “new” search-terms, you’ll absolutely have to run keywords on Broad, as your bread and butter. They also claim these kind of queries are becoming more and more prevalent with time, and what we’ve gotten used to is going to be irrelevant within a year or couple years tops.
3. **Iterative dialogue**. Simply put – they’re going to make Google Search into something resembling the current AI chatbots we love and cherish so much. Basically, an ongoing back-and-forth interaction where people and the searchbot collaborate to refine and clarify the query, leading to more precise and customized responses. I have no idea what it all means tbh. I can’t even fathom how it’s going to affect PPC and ad-triggers.I’m making this post hoping it’ll turn into a discussion revolving around “how-the-hell-is-it-all-going-to-change-our-job”:
* Is PPC going to become something else entirely?
* Is it going to become more “robotic”? just monitoring and making sure things don’t go to shit?
* Will there be a need to analyze the data at all, or do we just trust AI no matter what?
* Is it the time to start being innovative and turn PPC analysts into something broader and more creative, like “Marketeers”?
* In the not-so-long run, will we even need dedicated PPC analysts for each vertical?I don’t know if this post is too long. I promise it wasn’t generated by ChatGPT :(I’m simply concerned about the future of online marketing in general, and PPC specifically.For reference, I’m adding two links from Google that kinda summarize many of the things I’ve mentioned here:
[https://blog.google/products/search/search-on-2022-announcements/](https://blog.google/products/search/search-on-2022-announcements/)
[https://blog.google/products/ads-commerce/google-search-ads-generative-ai/?utm\_source=gads\_facebook&utm\_medium=social&utm\_campaign=io\_genaisearch\_051023&fbclid=IwAR2ndFrpIzk7\_NVQ0humeXg\_bPGrQaSFbJf5hpptXgSdmy0CVF1upxyp95w&mibextid=Zxz2cZ](https://blog.google/products/ads-commerce/google-search-ads-generative-ai/?utm_source=gads_facebook&utm_medium=social&utm_campaign=io_genaisearch_051023&fbclid=IwAR2ndFrpIzk7_NVQ0humeXg_bPGrQaSFbJf5hpptXgSdmy0CVF1upxyp95w&mibextid=Zxz2cZ)
-
1 Reply
-
Thanks for sharing !
-
Exciting changes ahead for PPC with AI advancements and evolving search trends!
-
Mind blowing following for updates!
-
First of all I think each worker will be able to do more work, which means workers will be cut.
In addition, the optimization will be done within the landing pages.
The competition will be based on who manages to motivate action within the site -
>There’s no need to go granular anymore. My company’s been priding itself on how granular and specific we get when creating ad groups and campaigns; very slim with a rather small amount of keywords, so the keyword insertion within ads will be more accurate, and so our campaigns will be more “easily” adjusted (in regards to target CPA or ROAS).
Sure, we stopped being super granular (with big clients) since 2020, when I worked for a big agency that had a massive client spending millions of dollars monthly. Google reps weren’t just giving us advice. They were working with us, implementing changes in the backend, fixing issues, testing new things for their R&D teams, etc. That’s in part also what the push for Broad match aims to achieve.
​
>They claim that stacking up keywords with the same (or even somewhat close) intent will not only increase traffic, but also provide better results; I guess they’ve been saying this for a while now, but this time they seem adamant, and provided some examples that kinda made sense.
Now this bit is confusing. What do you mean by stacking up keywords with the same intent? “buy red shoes” would capture “purchase red shoes”, “order red boots” and “shop rust pumps”. Do you mean I will get better performance if I go ahead and add every single one of these as keywords? Doesn’t this defeat the purpose of using Broad match to capture as much traffic as possible? Is this what you’re testing internally?
-
Are these comments generated by AI wtf
-
Interesting stuff, do you think we’re also looking at a change with the social media networks? Do they have a chance here as well? Lots of my closer circles are searching for stuff on TikTok for example.
-
I started to work on Google Ads (AdWords back then) many years ago, and I saw many changes.
The main one is that we have less and less control with what’s going on with our campaigns. We switched from manual campaigns (search & shopping), to automated ones), from manual cpc to smart bidding….With AI, I think you’ll need to have improved products feeds, products pages, …
I don’t think what the future will be, but for my career, I plan to switch for a more strategic job.
-
You know what, I’ve written a pretty long reply, at the very end, when you search for price related things, casino/betting, where to buy the chepeast iphone right now, it will never be able to provide you with a response, PPC/SEO, it doesn’t matter.
When you query Google right now you go to about 10 sites / 2-3 comparisons and compare the price, the AI wouldn’t be able to do so as it will pretty much ddos the entire internet in order to fetch live data.
Let’s say I’m looking for the best casino bonus. 4 PPC results, 10 SEO.
Assuming the 4 PPC results are “ok” in relevancy but may suggest actual brands, the other 10 SEO are more precise in what I was looking for as they provide (hopefully) affiliate websites that compare bonuses.
The AI (as it is right now) cannot provide the PPC results as they don’t actually display bonus comparisons making the entire tool useless, and if it will count on “organic” response, Google won’t make any money.
​
So in short, I can’t think of any better model than what we have right now, the top 10 SEO might change according to a new AI-algorithm but that’s pretty much it.
-
Heeeey, this is right in my wheelhouse! Earlier this year I kicked off building a startup basically offering this style of classification, targeting, and ad-serving tech for non-Google publishers, particularly the nascent batch of generative AI applications or contexts, so I’ve been building in the space and talking to quite a few people around what’s possible / where things are going.
Long story short — I suspect a LOT of things are about to shift, not just targeting being significantly more robust and flexible than manual Google keyword matching. We are, for the first time in quite some time, possibly going to have relevant search boxes that AREN’T just Google — I expect some chunk of the wave of AI-driven search engines (Bing, You.com, Perplexity, Metaphor, Lexii, etc.) to die off, but some of the bigger legacy / traditional media publishers are actively exploring surfacing their content, articles, structured and unstructured data into AI-driven experiences, so this isn’t just a matter of “how does Google Search change?”, it’s “there is a very real chance that a new format that’s something of a hybrid between open web display and current search is in the process of being born.”
CPC is more relevant than CPM in this paradigm, but neither are really fully “proper” for a fully dynamic chat experience, especially as available context signals and understandable user intent expand over the course of a session / conversation, and ESPECIALLY if you’re doing particularly interesting rich media native ads or engaged in followup chat with the ad system itself. (I.E. “advertising chatbot” built / tuned on the brand’s data embedded in the delivered creative payload that users can dynamically explore, whether it’s for brand reviews, other product offerings, or even asking for different brand recommendations, since there’s no strict reason that this experience needs to stick to displaying a single static advertiser.)
tl;dr — things are about to get weird (probably in a good way!) and adtech is actually pretty exciting right now! 🙂
-
Is most of the revenue from these sites generated from PPC? Or other channels as well?
Log in to reply.