Google SGE stands for Search Generative Experience. This is Google’s generative AI that will soon be infused into its Search Engine Results Pages (SERPs). Right now, SGE is still in beta and accessible to only select users who have signed up to test it — but it will likely roll out fully at some point in 2024.
Google's goal with SGE is to provide users with more relevant, recent, and helpful information directly in search. It does this by offering more information, links, and content directly at the top of search results — pulled in via artificial intelligence (AI) from various data sources. Google recognizes that asking follow-up questions is a normal and natural way for users to seek information today — so it’s also building interactive features into this new SGE results panel.
Let’s look a bit further into what Google SGE is and how organic rankings will be affected.
Google continuously releases updates to its algorithm — all in the name of improving the user experience. Some algorithm updates never become known to the public, while others cause quite a stir in the industry. Consider Google Penguin from 2011, which focused on demoting spam and forced businesses and SEO professionals to disavow and avoid low-quality backlinks. Or consider Core Web Vitals from 2022, which put the onus on developers to ensure websites were unbloated and fast-loading.
With this new marriage between Google results and generative AI, SGE may disrupt the industry more than any of its algorithm updates have in the past. That’s because it will significantly change how information is presented on results pages — affecting how users discover content and how businesses create content to make it discoverable.
Google uses generative AI to power its new SGE results. For context, generative AI uses large language models (LLMs) to create content-based responses based on user prompts. Content could come in the form of text, audio, video, or code.
The LLMs are created by gathering, organizing, and training massive amounts of data. It then uses natural language processing (NLP) to output user-friendly responses that sound like human language or speech. The specific LLM that Google uses for its SGE is called Pathways Language Model 2, or PaLM 2, developed by Google AI. This model is multilingual and extremely sophisticated in both coding and reasoning abilities.
Google SGE will provide some sort of answer-based result — depending on the type of search query. In testing, it looks like Google uses the left-hand column to showcase information (like short answers or brand listings) and a right-hand column to show supportive source material (like link sources or Maps). The information displayed on the left-hand side is almost like what users have become familiar with in a "Quick Answers” rich snippet. The type of search query dictates the structure of the SGE results panel and the volume of information presented.
Like most generative AI, Google SGE may provide different results each time – even for the same exact search query. This may come in the form of inaccurate information, sometimes called a hallucination, or it just may pull from different sources on different occasions, therefore providing different answers.
Many SGE results also take a while to load. Some results that feature more detailed information, like e-commerce details, take even longer. Google may cache results to improve load time, and in these instances, it may show the same result more than once.
Below, find information on the various SGE result types based on different types of test queries.
This is where Google SGE provides a short answer to a simple or straightforward question. Basic information results look similar to a paragraph or link-based rich snippet. However, it’s likely pulling data from multiple sources (rather than one source/one link as it did in its organic version), so it has the potential for more reliable answers.
When SGE shows a list — a brand name, a thumbnail image, and a one-sentence description appear for each. But upon clicking on the thumbnail image, it displays an image pop-up and link, rather than bringing you to the website itself, which can be confusing. Some brands aren't getting their thumbnail pulled into the SGE listing either, leaving that area blank. Brands should consider correcting this once it’s determined where Google is sourcing those images.
For retail-based queries, SGE displays a listing with more information for each source, including product name, rating, link, price, and a richer description. Plus, upon clicking, there are more product details shown — as opposed to clicks on thumbnail images. Because of these rich e-commerce listings, more emphasis will likely be placed on product reviews for businesses.
If it’s an image-based result, you can click on the image and drill through to the website source. This is one of the more basic and straightforward SGE results.
If it’s a statistics-based result, for example, “what’s the average business call length,” SGE provides a short answer on the left and a few supportive article links on the right. However, for the short answer, it doesn’t include the source of that information. Google may be pulling from the featured articles on the right-hand side, but it's not guaranteed. Plus, users can’t be sure which of those sources is directly producing the answer.
Searching for a local restaurant? Google SGE will provide more information than the former location-based results. SGE stacks listings on the left, places Maps on the right, and shows more context for each location listed. It will pull that enhanced information from user reviews and Google Business Profile data.
For a search that requests some sort of function, like “Fahrenheit to Celsius calculator,” SGE just gives information about the tool rather than providing the tool itself (which it is doing in the traditional organic search listings below it). That’s because, at least for now, the generative AI can only produce content, not functional tools. This is a worse user experience than before since users now need to scroll below the SGE results to find the tool they are looking for.
Google hasn't announced when SGE will be fully implemented, so it’s hard to say when the official rollout will occur. Based on that research, it seems like the official implementation will take place at some point in 2024. Once Google SGE is fully implemented, users may see four result types when they perform a search. It’s still unclear how and when Google will decide which kind of SGE result to display.
With this type, once the user enters a search query, Google generates and shows the user a full SGE result. This result gets generated automatically, i.e., the user doesn’t have an option to manually produce the SGE result. It also takes up a significant amount of SERP real estate above the traditional organic listings.
Once the user enters a query, Google automatically generates a snippet of an SGE result and presents a "Show more" button that the user can click to view the rest of the generative result. This will initially take up less real estate than the full autoload result unless it is expanded.
With this result type, Google doesn't automatically show the user an SGE result — but rather asks the user if they'd like to generate one via a "Generate" button. If the user clicks the button, then an SGE result is generated and displayed.
Finally, Google can choose not to present an SGE result at all. In this case, users will only see rich snippets, paid ads, Google Maps, and/or organic search results — just like before the Google SGE rollout.
Google holds over 90% market share for search engine usage, so that’s where most businesses and professionals will continue to focus and why understanding SGE is important. However, two potential threats to Google SGE usage are ChatGPT and Bing’s generative AI results — both of which are powered by the Open AI’s LLM called GPT-4.
ChatGPT is a much more conversational chatbot that users tend to use for idea generation. Suggesting recipe ideas, email subject lines, or brainstorming blog topics are a few common use cases. However, these habits may change as users continue to explore the tool and become more comfortable with it. Uniquely, Bing has trained GPT-4 data to be used within its own search engine results. Like Google SGE, Bing allows users to see generative AI results, prompt the AI for further information, or explore deeper results based on the presented links.
It may be wise for business owners and SEO professionals to monitor and test results in both Chat GPT and Bing before deciding whether it’s worthwhile to put in any extra optimization efforts.
Google SGE will create new real estate opportunities in search while displaying information in a new structure. Therefore, it will naturally affect listings and the level of traffic websites have come to expect from both organic and paid searches.
Google Ads may be directly affected by Google SGE. Right now, in beta, SGE results are showing above paid ads for some tests while showing below them in other tests. Google may still be experimenting with how they would like to display ads on their new pages.
If paid ads appear below SGE results, it may entice businesses to put more emphasis on content and SEO and less on paid advertising — shaking up organic results and the competitive landscape even more. If it chooses to present paid ads ahead of SGE results, it would make ads even more critical for businesses looking to drive search traffic — increasing ad competition and driving up bid rates.
It’s best to monitor how Google decides to incorporate ads into the new results and where it makes the most sense to invest resources.
The organic landscape will see an even more dramatic change than paid ads. Experts warn businesses that they may experience a 10-20% drop in organic traffic once SGE fully rolls out. Of course, this could be higher or lower, depending on the industry. Yoast SEO also suggests that users may see fewer clicks with Google SGE — affecting the total volume of organic search traffic. There are two specific ways things could shake out.
First, organic listings will be pushed further down the page to make room for SGE results. Not only will SGE often answer questions directly without users needing to click through, but it could cause some businesses to curate content differently to accommodate that change.
Second, SGE will provide new opportunities for more businesses to appear in search inside of this new generative results panel. Search Engine Land conducted a study and determined that there was very little overlap between the organic results vs the SGE results. In fact, 94% of links differed from those two areas. This could open up an opportunity for small business websites or smaller publishers to find their way onto page 1 while skirting around the need to try and outrank the organic top 10.
To prepare for Google SGE, businesses should continue to follow white-hat SEO best practices, create quality content, and follow Google’s recommendation of EEAT. This stands for Experience, Expertise, Authority, and Trustworthiness, where content creators should focus on appropriate keywords in the industry they belong to and generate deeply rich content that answers specific user questions. With SGE being able to gather and output much more specific responses, it’s also possible there will be a trend toward creators focusing more on longer-tailed keywords.
Businesses should pay close attention to updates from Google to better understand when it plans to fully release SGE and the best practices it offers for ranking, indexing, and being discovered in light of these major changes.
Of course, once you've mastered SEO and Google SGE, you'll be getting more leads coming in for your business as part of your efforts. Fortunately, you’re not on your own to deal with them. Partner with Smith.ai’s Virtual Receptionists and you’ll get a 24/7 call answering to ensure you never miss a single business call, chat, or text. Plus, we can even take care of appointment scheduling and lead intake. Schedule a consultation to learn more.
Disclaimer: Since Google SGE is still in beta at the time of this article’s writing, the above results may change as Google continues to make updates ahead of its official product launch.