In the 2023 Google I/O developer conference we saw early ideas of how generative AI will feature in search results. We also got introduced to a new term, SGE.
Let's go through what this all means and how we're likely to see very big changes to Google in the very near future.
What is the Generative Search Experience?
SGE stands for Search Generative Experience. It is Google's introduction of generative AI functionality in a search results to make them more personal, meaningful, and tailored to individual user patterns and intent.
When a query is entered, SGE doesn't just pull up the most relevant web pages. Instead, it anticipates the user's underlying questions and needs, and generates unique responses using Google's generative AI technology.
This approach looks very different from the traditional 10 blue links you see in Google. The search results are dynamically generated for each user, taking into account the context of the query, the user's search history, and patterns in how similar queries were resolved for others. This means you may get very different answers each time you make the query.
How does SGE work?
Google's Generative Search Experience (SGE) utilizes a variety of Large Language Models (LLMs), including but not limited to MUM (Multitask Unified Model) and PaLM2. These LLMs are part of Google's most advanced Gemini AI model designed to understand user queries, assess relevant content, and deliver pertinent results more efficiently and accurately.
The integration of these models into SGE signifies Google's commitment to leveraging cutting-edge AI to enhance search experiences for users by providing more relevant, comprehensive, and contextually aware responses.
Can I try SGE?
As of March 2024, the Search Generative Experience is only available in a limited number of countries through the Google Labs early access program.
How will SGE impact search results?
It's unclear exactly how widely the Search Generative Experience will impact search results, but this is likely to be clarified by May of 2024 in the next Google I/O developer conference.
The likely impact is that Search Generative Experience will feature prominently highly in specific search results, either replacing paid ads or organic search functionality with Google's own generated results. In the below example, it looks like the biggest drop will be in visibility for traditional organic results.
Already there are indications of big changes based on beta-testing the new features. According to fresh research by Authoritas:
- Google could display generative results in over 90% of search results,
- Almost all branded searches in the future (eg. "Nike") may feature generative answers,
- 62% of SGE links come from outside of the top 10 search results,
- Transactional queries like ecommerce, travel and insurance are among the most impacted.
Potential downsides of SGE
While the Search Generative Experience (SGE) is likely to bring about the biggest change in how we interact with Google Search for the past 20 years, there are also potential downsides that should not be underestimated.
One significant concern is the risk of inaccurate search results. Since SGE relies heavily on generative AI to produce answers, there's a possibility that the information generated might not always be accurate or reliable. This inaccuracy can stem from the AI’s reliance on existing online content, which itself may contain errors or misinformation. According to a recent report by Bleeping Computer, SGE can include both malware and scams.
Another potential downside is that SGE relies widely on what Google consider as authoritative websites compared to independent small businesses. Recently, user created content websites such as Reddit and Quora have dominated Google search results, with Reddit especially featuring in over 97% of product review queries.
How you should prepare for SGE?
The only good advice I can give to SEO professionals as it stands is "prepare to prepare for SGE roll out soon."
In reality, we do not know how search engine optimization best practices need to be adjusted. We do get some general indication from Google based on recently updated helpful content guidelines.
Here's a few things you can keep in mind:
- Focus on your ideal customer profile: Center your content around the queries of your ideal users, providing thorough, value-driven answers.
- Don't ignore long-tail keyword: Harness conversational, long-tail keywords to meet specific search intents with valuable and unique answers.
- Consider search intent: You can consider the semantics of the search query to understand what exactly are people looking for and answer questions better and more comprehensively.
- Revise your structured data: Apply structured data to make it easier for your website's context for better search result display.
- Don't ignore technical SEO: Focus on a quick, responsive website experience, factoring in its importance for rankings and engagement.
- Emphasize your authority: if you are a subject matter expert, make sure to show it on your website. While E-E-A-T is not a direct ranking factor, let people know why you give authoritative answers.
The bottom line
In my view the Search Generative Experience will bring about the biggest change in Google's search results for the past 20 years. SGE promises to make search results more personalized, actionable, and contextually relevant, leveraging the immense power of generative AI. It remains to be seen if Google can roll out these improvements without exposing users to new vulnerabilities and scams.
While the change will be dramatic, I don't expect it to happen overnight. It will be crucial for Google to refine and adapt SGE, ensuring that it not only enhances the search experience but also maintains the accuracy and reliability that users have come to expect. My hunch is that we'll can see a lot in the next Google I/O developer conference. Only with more information can SEO professionals start to re-define and adjust their approach to a generative AI-centered search experience.