Google may be signaling that SGE isn't happening anytime soon...or at all. In their recent announcement of Circle, they mention: "But our goal is to make AI helpful for everyone, not just early adopters." I read this to mean their data on SGE is that not enough people found it useful, and it largely remains a tool for the tech community to play with. They also mention: "This week’s launch of AI-powered insights for multisearch is the result of testing we began last year to see how gen AI can make Search radically more helpful, with SGE in Search Labs. We’ve gotten lots of useful feedback from people who’ve chosen to join this experiment, and we’ll continue to offer SGE in Labs as a testbed for bold new ideas." People's feedback may be suggesting that SGE is not as helpful as many had hoped during the early days of the AI frenzy. More: "So moving forward, as we continue to experiment and uncover which applications of gen AI are most helpful, we’ll introduce them into Search more broadly, like we’re doing now with multisearch results." H/T Ross Hudgens. Lily Ray, Marie Haynes, thoughts? Link in the comments as always. #Google #SEO
It all came together for me this morning. SGE is a test bed for features that could be helpful to a searcher. Assistant shows the feature that it thinks you will find the most helpful. In search you might see a maps pack as a part of an entire SERP. In Assistant you would just see the maps pack. On my pixel phone if I tap the microphone that is in the search bar that appears on the home screen of my phone I get different results than if I do a text search. I get assistant. I expect over time more Android phones will get this and possibly even the same will happen for those using the Google app on iPhones. Sundar Pichai said sge was just the way search will work eventually. We have been thinking of sge as something that will get implemented into search, but now it makes sense with the New York times said last year that Google was working on a whole new search engine. I believe this is it. The question is how quickly the transition will happen.
I suspect the impact on search ads and their business model was also at play here. How do you blend advertising with complete answers or fulfillment of the search intent without impacting clicks to advertisers? I'm sure it was a coming of age moment for their product teams.
Does anyone feel like they are just unsure of the best way to monetize the results?
Awesome insights, thanks! This line from the post also stuck out for me: "AI-powered overviews on multisearch results are launching this week in English in the U.S. for everyone — no enrollment in Search Labs required". While "SGE" as we've seen it in labs may not be rolling out to text-based search, a very similar experience is already publicly available for Lens-initiated search. I think we will have to evolve how we talk about SGE as more pieces of it roll out.
SGE seems to be a playground to test new features enhanced with generative AI. Whenever Google feels like a bit of it is adopted by users and revenue neutral or positive, they push it live. They will not push SGE out in full, if they are uncertain how it will affect revenue. Their next earning call and how search revenue as well as cloud revenue developed will be very interesting to see.
Skeptic, Conspiracy take: Testing hasn't produced the significant increase we expected in ad click throughs by pushing organic further down the page...yet.
It's all about the money. Less web traffic = less Display ad income. They're also still figuring out how to integrate Google Ads. And then there's the risk of displaying wrong or copyrighted information... (and the lawsuits that come from it). It's a long time ago it was all about the user experience at Google. The benefit just doesn't outweigh the projected cost.
Founder at North Star Inbound | Organic Growth through Content and Links
3mohttps://blog.google/products/search/google-search-ai-january-2024-update