Search Generative Experience, Bard and Gemini: SEO into the future

SEO is undergoing the biggest transformation in a long time. There was never any doubt that all the major players would jump on the AI wave, and now Google is ready to fully enter the market.

Google Bard is already well-known to many, so AI is not new for Google. After all, they have been claiming to be an "AI-first" company for 10 years, and they have already used AI in many products. In fact, ChatGPT itself is largely based on research conducted by Google.

But they have never made AI as accessible to the general public as OpenAI did with the release of ChatGPT. Bard was their attempt to keep up while they worked on their own AI initiatives.

What is Bard?

In short, Bard is Google's version of ChatGPT. But Bard is actually just the user interface on top of the AI model itself. The AI model is the engine that Bard is built upon, and it was initially called LaMDA, which was later replaced with Palm 2.

These models have been inferior to ChatGPT in most ways. Bard is a bit better at maths than ChatGPT, but that doesn't help much when most users need help with text or images.

I have tested both Bard and several competing models from companies like Facebook, but none of them hold a candle to GPT-4.

Is a shift on the way?

But now Google is introducing a new model to support Bard: Gemini. This model is (just barely) better than GPT-4 in most metrics and will apparently be a real competitor.

If you're wondering why Google is switching to a new model so quickly after starting to use Palm 2, we're in the same boat. No one is completely sure, but Google itself says that Palm 2 and Gemini are not competitors, so they probably serve different purposes.

Gemini is not just a language model - it's what's called "multimodal." Multimodal AI has so far been the description of an AI that can process images and videos in addition to text, as you have probably seen with new versions of GPT-4 and Bing. But Gemini includes much more: graph processing, image analysis, audio processing, language models, programming, and 3D models.

An illustration of the three versions of Gemini: Nano, Pro and Ultra.

A lot can be written about how great Gemini will be, what metrics Google has used to claim superiority over GPT-4, how much easier Bard will be to use, and so on. But what interests someone like me is: How will this affect my job as a content producer?

And that's where we come to Search Generative Experience (SGE).

Search Generative Experience

Bard will be just one of many Google products that use Gemini. Google has several launches in the pipeline, including a completely new version of its search results page. The top of that page, where the paid results are currently located, will look quite different:

An illustration showing the new google SERP: Paid ads will be replaced by SGE on the top of the page

And Gemini will be in charge of the top of the new search results page.

What will SGE include?

The purpose of the blue box in the image above is to provide Google users with the best possible search result with minimal scrolling. This is achieved by allowing artificial intelligence to generate text and find helpful web pages that best answer the user's query.

An example from SGE where a user has asked for "bluetooth speaker for a pool party". SGE gives a short generated text, shows 3 sources on the left, and relevant products underneath.

Next to the text, the sources SGE used will be presented in a carousel, and if the user specifically asked for a product, it will be followed by some product pages.

By expanding the results in SGE, the user can see exactly where the AI sourced the information in the generated text.

SGE and ads

You might be asking, "Where did my ads go?" SGE does not include ads right now, but since Google earns a lot from ads, it's quite certain that they will be included in the future.

No one knows exactly how ads will be integrated, but Google says they will "experiment with search and ads that are directly integrated with AI."

See how we work with ads

SGE and SEO

If the AI-generated search results are so good that users hardly need to scroll to find a good answer, SEO practitioners face a challenge: What's the point of investing in ranking high on Google if you can't rank all the way at the top?

It's quite clear that most of the traffic will go to the top 3-5 ranked results for a search query (i.e., the sources SGE used to generate an answer). It has always been important to rank high, but being in the top 10 has been fine so far. Now, it will be extremely advantageous to be in the top 3 (even more so than before), and investing in anything less will be a waste of money.

So what does it mean for SEO? Well, fortunately, it means what I have been saying for a long time: We need to focus more on making content so good and specific that it cannot be ignored. And it's not just about being good at writing - included in the definition of good content is (and has been for a long time) writing meaningful content about a very narrow topic that helps readers solve a specific problem or make a decision.

We must focus even more on building content clusters around the topics we are experts in and structure these pages in a way that both Google and readers understand how everything is connected.

Building your knowledge bank around a given topic will give your website higher authority on that topic, which (just like before) will push you up in the rankings.

If you continue to focus on E-E-A-T (Experience, Expertise, Authoritativeness, Trustworthiness), you will be able to continue ranking high on Google. After all, the changes with SGE are meant to provide real users with better search results, and the E-E-A-T criteria will naturally be part of what makes a website rank high.

And the return on investment will be even higher: if you get into the SGE results for specific queries, you will almost monopolize certain topics.

Back to the classroom

Do you remember the test questions in school that started with "elaborate on..."? It's time to bring back those skills and write comprehensively about a topic because it will probably be harder to rank high by answering questions simply.

Long articles that explain multiple aspects of a topic will rank higher than dry factual articles.

But SGE doesn't have all the answers

Sometimes, SGE won't provide any results at all. Cyrus Shepard has created a list of things that SGE won't be able to answer:

  • Most navigation-related searches
  • Recipes
  • "Adult" queries (NSFW)
  • Some health and financial-related queries
    News
  • Quick answers; song lyrics, weather, etc.
  • Sensitive questions related to gender identity, the Holocaust, and similar topics.

So for those working in those areas, just keep doing what you've been doing.

Will Bard and SGE be competitors?

It's not easy to keep track of models and products, but one thing is certain: Bard and SGE will coexist.

Perhaps it's a tactic to see which product resonates more with users? After all, people have started replacing Google with ChatGPT precisely because they get comprehensive answers instantly, without having to search as much.

At the same time, one still needs to fact-check the answers on Google. But if they get their models to stop hallucinating, it's hard to see the significant advantages of using SGE over Bard or ChatGPT. Unless one is specifically looking for a particular product or service.

It is therefore possible to envision a future where the two products are merged: where you get both comprehensive answers and ads and product recommendations in the same place.

Do you need help preparing your business for SGE?

We have been working with website development, content, and SEO for over 20 years, and we have no plans to rest on our laurels when SGE becomes part of everyday life on the internet. If you would like to have a conversation about how you can increase your visibility online, feel free to send us a message through our contact form. Alternatively, check out our blog or sign up for blog notifications right below this article to get what you need to succeed with SEO.

Author

Sven is part of the Experience Team in the role as content producer and project manager. He writes about customer experience, content, trends and digital marketing from his experience and creative work.

Subscribe to our blogs