Dans cet article
Good GEO IS GOOD for SEO
Good GEO is good SEO. This is not me saying it, but Danny Sullivan, a spokesman for Google.
For those who don’t know, GEO means Generative Engine Optimization, like Gemini, ChatGPT, Perplexity, Claude, etc
Why talk about generative engine when talking about Google AI Overviews and AI mode ? Because the AI overviews (and Google AI Mode) is based on the Gemini under the hood.
This might suggest that a good SEO is enough to be visible on the AI Overviews. But this is only part of the truth… so, What are the techniques that really work for ranker on the AI Overviews and other LLMs ?
How to optimize for AI Overviews
Several techniques are very popular in this time. They are not all White Hat, so use it at your own risk ! A man warned is worth two to…
The ‘best’ listicles
The creation of listicles “best” is a simple technique to put in place… and that really works. The concept : write an article with the best X solutions in your niche, by putting your solution in the first position. Simple, but effective.
Attention : some SEO consider this technique as Spam. Some will go upto say that cis of Black Hat.
Lily Ray has published an article on his LinkedIn to denounce this practice and to show how easy it is to manipulate the AI Overviews ⬇️
Lily Ray is gone even a step further by publishing an article listing the best SEO (with a little anecdote). It was found that the LLMs had taken into account his article in less than 24 hours ! Mind-blowing no !
✅ Strong Impact.
✅ Easy implementation.
❌ Gray Hat – Black hat
Press relations (PR)
Google has confirmed that the press relations improve the visibility on the AI Overviews. If it works for Google, you can be sure that it also works for the other LLMs.
Even without official confirmation of Google, many SEOS have found that the press relations play a key role in the visibility on the LLMs. In contrast to the first technique, this strategy has no negative effect, so you can put there without worry.
✅ Strong Impact.
‼️Implementation costly.
✅ White hat
Chunking
Google AI Overviews take parts of the articles thatthey use for Grounding. Youhave probably note. The chunking (cut content) plays on this distinctive.
Theobjective of the chunking nis that of cutting your content in part, independent, clear and right to the purpose to promote theuse of your content by the ai overviews and AI.
‼️ Impact intermediate.
✅ Easy setup.
✅ White hat.
PS : I have a detailed article on how to do chunking properly for SEO / GEO.
Aim for the query information long trains
‘VE Overviews, as well as the LLMs as ChatGPT prefer naturally the query information literacy long tail. Logic : think of the way you ask a question of your LLM preferred. You do not just say : “I want a work space in Paris”. You describe, rather : “I am looking for a working space not too far from my neighborhood, with a good connection and not too expensive”.
✅Strong Impact.
✅ Easy setup.
✅ White hat.
Put in place structured data
What is theinterest of the markup scheme for GEO ?
Models ofIA recognize the content with greater precision anduse in the generated responses.
The web sites that appear most prominently in the search results, even without clicks.
The information is displayed directly in searches without a click.
The chances ofbeing included in the extracts optimized or the panels of knowledge are the highest.
‼️Impact intermediate.
✅ Easy setup.
✅ White hat.
Optimize for Query Fan Out
Query fan-out is a technique used by the LLMs to respond quickly and efficiently to users ‘ queries. I talked about it in my article on the difference between ‘VE Overviews and the AI Mode of Google.
The technique is to decompose the query into several facets and then into sub-queries, in order to prepare the answers more quickly and provide complete responses.
The best way ofoptimizing your content for the Query Fan-Out is to know the sub-queries related to your topic. For this, there are tips and tools. I am planning to do a full article, but in the meantime, you can take a look at Otterly, which is rather effective.
✅ Strong Impact.
✅ Easy setup.
✅ White hat.
Embeddings research (vector)
The embeddings to transform their text into digital vectors for as LLMs can understand its meaning , and find relevant information quickly.
What is the interest ofusing the embeddings ? RankBrain and RankEmbed BERT using theintegration vector.
Should be noted that John Mueller of Google reminds us that the optimization by embeddings near the keyword stuffing, and therefore can be counter-productive.
‼️ Low Impact.
‼️ Complicated set-up.
‼️Grat hat / Black Hat.
Creation ofan entity
Engines generative as ChatGPT, Perplexity, or Gemini do not work such as Google : they do not just index the pages, but to synthesize the responses from entities and relationships that they have learned or retrieved on the Web. That is why it is important to create an entity for its brand and ensure consistency of all the information on all channels.
✅ Strong Impact.
✅ Easy setup.
✅ White hat.
Matrix techniques GEO
In conclusion, it should be borne in mind that the techniques presented are not ‘hacks’ are isolated. They are the components ofan integrated strategy. Building your approach on the pillars of theAuthority, Content and Semantics, you do notoptimize for Al Overviews oftoday,today, you const-truisez a solid basis for the research of tomorrow.
Leave a Reply