How does the rise of LLMs impact how we should approach SEO?

With the use of LLMs for search very much in the ascendency, there’s been a lot of chat about if this will make traditional Google search redundant and if “GEO (Generative Engine Optimisation)” is distinct from SEO.

This article explores some of the ways that the rise of LLMs and generative search engines such as ChatGPT and Perplexity might change how we approach our overall search strategies.

TL;DR - There are some differences between LLMs and traditional search engines such as Google, but it’s about diversifying our approach and adapting tactics accordingly, not replacing one for the other.

How will SEO strategy change?

• Diversified and adapted, not replaced

The rise of LLMs for search represent a huge opportunity for brands, however, it is one of diversification of strategy and acquisition and should not be seen as a replacement of traditional sources particularly Google, which still maintains an enormous market share. In 2024 Google had an estimated 373x more searches than ChatGPT; even with ChatGPT rapidly growing in popularity, this does show we cannot neglect Google.

The rising popularity of LLMs should be viewed as a positive for brands however, who now have an opportunity to benefit from a new channel and reach their customers in new ways (particularly if Google wasn’t forthcoming in this regard for whatever reason!) and for agencies, who can diversify tactics and show their expertise in new and innovative ways.

It does not represent the death of SEO, or Google. Nor does it mean LLMs are now the only thing to focus on. Strategy should be diversified across these channels and adapted to suit the nuances of each. It is not a case of one replacing the other.

• Product > Category

One interesting difference between LLMs and traditional search is that it shifts the focus from category pages to product pages (PDPs).

Category pages for traditional Google search have tended to be the biggest traffic and revenue drivers for eCom sites (since they rank for the high volume generic keywords which also have purchase intent). However, since LLMs have the capacity to skip a particular brands category page and instead aggregate products from a variety of brands, this places the emphasis much more significantly on PDPs for LLM visibility! 

Don’t think Google is standing still though - as has been noted, Google Shopping updates will look to recreate a similar PDP-led experience.

How will Digital PR/link building change?

• Citation-based link building

Whereas currently Digital PR link building is focussed on building links from high authority, relevant publications, with LLMs the specificity of citation used means that Digital PRs targeting LLM visibility will aim to get into particular articles and definitely frequently cited publications.

Below you can see an aggregation of citations for a brand we are tracking LLM visibility for; achieving links from the publications listed will become a priority, both at a domain and a specific article level.

• Links to PDPs > links to category pages

With the increased return of PDPs in LLMs, citations that reference specific PDPs will likely become more important for visibility in these channels, whereas for Google category links still remain the gold standard for ranking for generics.

How will content strategy change?

• Reduced focus on top of funnel content for acquisition

With customers able to ask longer and more complex questions with succinct answers from LLMs without the need to click through to a source site, the number of people searching top of funnel in Google will likely decrease (and is something we’re already seeing). This does not mean that there is no value in top of funnel content, it will still be important in building expertise to rank in traditional Google search and can be very useful in re-engaging customers via other channels such as email.

However, given that top of funnel content is already difficult to directly prove the ROI of, a reduced focus in this area is likely. 

But where should the focus go instead…?

• Increased focus on intra-product comparisons

Comparing and contrasting the pros and cons of a brands product catalogue should become a priority. Some brands already do this, for example the Brooks example below which I found when looking for my next pair of running shoes. Amazon has also been doing this for a while.

However, I think brands will need to go even further; giving honest assessments of the benefits and potential negatives of products, insight into which demographics they may suit best, and deeper product information that LLMs can use to compare and contrast in their own responses.

Although we have given some examples of brands which are doing this, a huge number of brands don’t have this intra-product comparison and analysis at all. This will mark a key change compared to traditional content strategies for a large number of brands.

• Increased focus on brand USPs (implied inter-brand comparisons)

Similarly, ensuring brand USPs are clearly explained on-site, and across all key pages on a site (not just a single informational page which describes brand value, but across PDPs, categories etc), will also help to ensure LLMs have the necessary information at hand to favourable compare brands against one another in results. 

Although direct brand vs brand articles feel a bit too crude (and have obvious bias), implying why your brand is better than competitors by focusing on your own USPs (ie we’re the only brand that does XYZ!) and spreading this across your site should become much higher priority.

The focus becomes ever more brand led.

ChatGPT Shopping actually pulls into the product results tags of the USPs for each product - therefore ensuring that intra-product and inter-brand USPs are clearly assigned across the site and on PDPs will be crucial to gaining a competitive advantage.

How will technical SEO change?

Of the three channels, technical will probably change the least - you still need a website to be optimised, crawlable, and good for customers. However, there may be a some more prominent considerations.

• Server side rendering

It’s been widely reported that LLM crawlers struggle to render JavaScript, so having server side rendered content, or an alternative solution such as dynamic rendering which allows LLM crawlers to directly received the rendered HTML will become priority number 1 once again. 

N.B. It always should have been a priority anyway, but since Google has been able to render JavaScript has been de-prioritised by some brands! Even for Google, rendering takes 9x longer than crawling HTML due to the additional resource required to do so, so it’s still important to optimise this experience for Google as you typically would for a user.

• Structured data priority

Similarly, structured data has for a long time been very important, but with another channel which will likely look to gather product and brand information in the most efficient way possible, clean and clear structured data will shoot to the top of the list. Essentially, the power of GPT crawlers are akin to Google from 2019 so if anything require an SEO to help technically access the site more than Google does!

• Product feed optimisation

Although Google Shopping already allows SEOs to edit product feeds, if LLMs such as ChatGPT offer product feeds for inclusion in their shopping results (as they have teased is likely to come) this will be set to become a much larger part of an SEOs day-to-day work.

Summary

• Google remains a dominant force in search and shouldn’t be ignored
• The rise in popularity of LLMs is a positive for brands, giving Google competition and allowing brands additional ways to reach customers
• Strategies should be diversified to take advantage of growing LLMs and can be adapted in various ways