The MKG Vantage: AI in the SERPsNathan Stenberg, Anna Powers, Austin Ellis • August 7, 2023 • 10 minutes to read
We challenged MKG’s three SEOs to get opinionated on how AI in the SERPs will impact organic search. How do they see search results and user behavior changing? And how will this change SEO as a discipline and a service?
We Asked: Currently, if you are opted in, how do you feel generative AI impacts search engine results?
MKG Verdict: The current AI-powered search experience is decent, but its helpfulness also depends on what stage of the funnel the searcher is at.
Nathan Stenberg, MKG Strategist and SEO Expert:
The current experience (in Google) is fairly reasonable. In most cases, it tries to provide an answer for you but in some cases, it actually just asks if you would like a response generated.
When an answer is provided, it usually contains links to supporting documentation about the initial search query and it will also provide a few additional prompts to dive deeper into the topic.
Example: If you search for “Avalanche Energy” it will return a response about the energy startup, Avalanche Energy. It also provides three links to specific web pages as well as a prompt to “Ask a Follow Up” which allows you to enter a new custom query and 3 topic specific prompts:
- Who are the competitors of Avalanche Energy?
- Who is the CEO of Avalanche Energy?
- Avalanche Energy stock
Anna Powers, MKG SEO Team Lead:
I would agree with Nathan that the current Google AI-powered experience is decent, especially for TOFU queries like the branded search example he shared above. But my opinion gets more complicated as we go down the funnel.
A middle-of-the-funnel search for “email security features” gives me an AI-generated list of possible features to look for, which is pretty good! A thought starter for me as I scroll through the blue links below. In this use case, the AI-generated results are helpful but not necessary.
A bottom-of-the-funnel search for “email security tools pricing” generates prices in the AI results, without attributing them to a source. Hmmm. This I do not trust, so I scroll right past to the links underneath so I can get pricing directly from websites selling email security tools. In this use case, the blue links capture the click.
For searches that matter (YMYL, cough cough), the SGE results are just not there yet, and they might never be. Google knows this, which is why YMYL searches are, for the most part, not showing AI-generated results.
Austin Ellis, MKG SEO Manager:
I have yet to opt in to the Search Generative Experience, particularly because it’s still in early stages of beta testing after being announced at Google I/O in May. From what I’ve read, it seems as though the AI tends to get even basic facts incorrect in its presentation.
What We Asked: In the moment, as well as in the future, how will generative AI affect traditional SEO practices and strategies?
**MKG Verdict: **We will need to adapt, but we don’t know to what. For now, stay the SEO course with helpful content and sticking to E-E-A-T best practices.
Nathan: In this current moment, the impact is minimal and SEOs just need to stay focused on what has worked, getting great content produced that is topically relevant to the keywords that you are trying to rank for. In my opinion, this will always be the case in some capacity as brands need to build trust and authority around the topics that they want to own even if the search engine experience shifts.
However, in the big picture, I feel that the focus will need to be on making each section of content more impactful as the response in the generative AI experience tends to be about the size of a large website module (100 words or 500 - 700 characters). There is a strong possibility that this will require a shift in the approach to content development to make the content more conversational as well as comprehensive and thoroughly detailed.
Anna: In the current moment, we should continue to focus on making our content as helpful as possible and following E-E-A-T best practices. But in the long term, I am not going to hedge: SEO will need to adapt. We just don’t know what that's going to look like yet.
Easy, top of funnel traffic might very well dry up, and those awful long recipe blog posts that are padded for word count will disappear (thank goodness). The most helpful content, in the human and SEO sense of the word, will still get visibility and clicks even with AI content taking up much of the real estate at the top of the page.
Austin: It’s entirely possible that Google will uphold its promise and its stance that creating helpful content and topical authority will rise to the top of the SERP. However, those who work with smaller niches and markets will potentially have to fight with larger unrelated sites based on what the AI understands, which could be limited.
We Asked: Of the search engines that have implemented generative AI in their SERPs, which is your favorite and why?
MKG Verdict: Split decision here, but it’s clear that Bing’s integration of AI has finally made it a player in the search space.
Nathan: Personally, I am partial to the Google user experience as Google is a far superior search engine and it was a hassle to try and get set up for Bing.
However, seeing as Bing is using OpenAI, which really broke the use of AI into the general market, I do feel that Bing has increased its competitiveness with Google making it more of a viable search engine for users who have Bing installed by default on their home or work computers.
Anna: Bing is the best in terms of accuracy and trustworthiness, in my opinion. I surprised myself with this answer, but Bing runs on OpenAI’s GPT4 and delivers a list of links that looks like the current SERPs. Bing’s AI-generated summary is off to the side and easy to ignore, unlike Google’s, which is the first thing on the page and takes up a huge amount of space above the fold.
Google’s SGE summary also has the unpleasant tendency to scrape content (without attribution!) from other sites, which feels like a content grab and frankly, disrespectful to content creators. This would need to be fixed before I take the SGE at its word for serious research.
Austin: I’ve heard that Bing’s incorporation of ChatAI has been more accurate than Google’s Bard. I don’t know if that’s due to the quality of the training materials or because OpenAI absolutely exploded a few months back, with more users and more queries to understand searchers.
We Asked: How will generative AI impact user behavior and search intent?
MKG Verdict: Search behavior will change, but intent might not.
Nathan: Generative AI will initially change the way users research topics and in time, will impact the way users shop as well.
From the research lens, this supports the idea that the content a website provides will need to be detailed and thorough as well as try to anticipate the next question the user is going to ask so it can be the “source of truth.” This doesn’t really change much as this is the type of content websites should be providing to their visitors.
From the shopping lens, I can see the need for reviews coming heavily into play. It would make a lot of sense to provide a recommendation for the “best hotel in Cancun” or the “best 8k USB-C to HDMI adapter for Mac” by scraping the product reviews from different sites and then combine that with reviews of the company offering the product or service to generate a recommendation.
Anna: I think Google’s new Perspectives filter, introduced to complement AI results, will be big. Perspectives pulls in top results from platforms like YouTube, Reddit, Instagram, and blogs, presumably based on content helpfulness, likes / shares, and overall quality.
You could use Google proper for a search like “best hotel in Cancun” (to use Nathan’s example), comb through the SERPs for a minute after reading the AI-generated results, then pop over to the Perspectives results for a different, well, perspective. The user intent doesn’t change–the user still wants a hotel recommendation–but the user experience and behavior certainly does.
Austin: While much of this is speculative, user behavior may be altered more subtly at first, shortening the attention spans or tolerance for poor website experiences. Meta titles have been and will perhaps be more direct, straight to the point.
We Asked: What are the potential implications of generative AI on website rankings and organic search traffic?
MKG Verdict: Generative AI might create more opportunities to compete for a spot at the top of the SERPs or tank organic traffic. TOFU, e-commerce, and B2C searches might be the most impacted, but it’s hard to tell.
Nathan: It will depend on the type of search that is being performed but for now the generative AI is like a very robust Featured Snippet so while it may provide some solid detail in the actual search page, a user may still continue through to the website to really dig in.
However, in the big picture, we could easily see Organic Traffic decrease significantly so SEOs and digital marketing agencies will need to look to other metrics to show that their efforts are still needed and are driving success.
Anna: Agreed. I think that AI-generated results–particularly those that appear at the very top of the SERPs–will have a huge impact on top-of-the-funnel, informational searches, your garden variety “what is” queries. Try it: turn on the SGE and search for “what is a prime number.” Do you need to go any further? Sorry Wikipedia, but I predict your traffic is going to suffer.
To Nathan’s point, I’m waiting to see how Google will measure and report on appearances in AI-generated results. Clients are going to be very interested in this as well, since there might be more of a chance to compete for a spot in the AI-generated carousel at the top of the SERPs.
Austin: The potential implications are many, depending on how far Google is willing to push it. It seems as though this new layout has been geared towards expediting and enhancing B2C searches and the shopping experience.
This brings potential conflicts if certain sites (Amazon) have a monopoly on the AI generated search spots due to Domain Authority. Google already has control of the SERPs and it seems as though SGE is a further step in the same direction.
We Asked: Are there any best practices or guidelines for optimizing websites for generative AI-powered search results?
MKG Verdict: Not yet. Stay the course with solid content, good site architecture and technical SEO, and on-page optimizations that are enhanced with structured data.
Nathan: All of the research I did around this returned pretty much the current SEO optimization tactics so stay the course for the time being.
Personally, beyond the “standard” optimization approaches, I would focus on providing as much context to the search engines by putting extra emphasis on marking up the content with a clear, easily-navigated HTML structure and, when applicable, comprehensive Structured Data.
Anna: The content carousels included in the SGE’s AI-generated results are probably going to operate a lot like featured snippets and other rich results, which Nathan also mentioned up above. Earning your website a spot in the carousel will require high-quality, helpful content that is proven to get a high CTR in the SERPs. We might even get AI-specific structured data to mark up our content at some point.
Austin: We will have to see once SGE is available on a wider scale but, it seems that the foundational SEO approaches should hold up and remain just as important. The best SEO practices hopefully will continue to be: creating helpful content and structuring sites in a logical manner. Google has already made its negative stance on AI-generated content clear.
So this is the MKG Vantage on AI in the SERPs, based on our expert opinion and the information we have so far. Google has made a big deal in recent days about SEO fundamentals, how first principles like “make content for people, not search engines” haven’t changed in the past 20 years since we’ve been practicing SEO. While this is probably an effort to reassure nervous SEOs about AI in the SERPs, it’s probably also true.
What we know for sure is that Google and Bing’s best interests are served by integrating AI in a way that 1) still allows them to make major money off ad revenue, and 2) still encourages people and organizations to create good content so that the search engines have something to index. Just how that happens is going to be a wild ride.