Aleyda Solis covers OpenAI's new shopping research feature in ChatGPT, which pulls from Google Shopping, plus John Mueller's clarifications on AI content recovery and a case study showing traffic gains from removing low-value pages.
Chapters
Let's go through this week's biggest search stories in today's SEO FOMO to London read, going through the most popular news from the latest edition of the SEO FOMO newsletter. Today's edition is sponsored by Similarweb. Thank you very much for your sponsorship. The biggest search story by far this week was OpenAI release of the shopping research in ChatGPT, which is a new feature that allows ChatGPT to research on your behalf to help you to find the right product. As you can see here, after you search for something and recommending the initial round of products, it will showcase this new option to research further while asking you for questions that will allow ChatGPT to refine the results and make them more accurate. This is a shopping assistant, and this feature, along with their e-commerce instant checkout integration, will pretty much allow the user to perform the whole purchasing journey within the ChatGPT platform. Platform, which is one of their goals. The amazing Alexis Rilko has already done an initial but very in-depth research going through the major characteristics of this new shopping assistant/guide feature from ChatGPT. And one of the things that he discovered will surprise you or not, that is that pretty much ChatGPT Shopping is being powered by Google Shopping. He explains how the shopping research goes through a 4-step shopping guide. Understanding the user intent, reasoning, querying external sources, and producing the final shopping guide. And he explains how ChatGPT has moved forward with the usage of Google SERPs. We knew that this was already the case by some additional findings from Alexis that I also was able to replicate, but it's not only that. ChatGPT Shopping goes even further. The product information is exactly the same as in Google Shopping: product names, images, average rating, number of reviews, the URLs of the pages referenced in ChatGPT Shopping regularly contain the Google Shopping markup. Surprise, surprise. The merchant icons in ChatGPT come from the Google Shopping API and not the standard API. Check out Alexis's guide. I highly recommend it. All of this should make us realize that AI search and LLMs is not only a branding channel anymore, but an actual performance channel, at least for e-commerce sites. For this particular vertical, their experience and features are definitely moving to a different path that will actually enable to develop itself as a performance channel. And if you are a retailer, you should optimize for it accordingly. The next top story is this coverage by Barry Schwartz going through this clarification by John Mueller saying that rewriting AI content with human content won't directly lead to a recovery. To give a little bit of more context, this was an answer given by John to someone who was asking about if a site with low-quality AI content can recover and be indexed if they rewrite everything manually and switch languages. He mentioned that the changes that they will need to do will likely need to be strategical and in-depth, rethinking their whole content strategy and the purpose of their site rather than just rewriting it, which makes complete sense. The other top story is again another John's clarification about the need of LLM-only Markdown pages. In particular, building LLM-only Markdown or JSON pages for LLMs, because that's why we have already HTML and structured data that is already in place. John asks, why would they want to see a page that no user sees? And if they check for equivalence, why not use HTML? And makes complete sense. There's no need to reinvent the wheel. And last but not least is Stephen Muschietti's post in LinkedIn that attracted a lot of visibility and shares as well, showing what happened when they deleted over 50% of a website content. He explained that this was an energy comparison site that had more than 1,000 outdated, off-topic, and low-value URLs dating back to 2015. And after an in-depth content audit, they removed everything that didn't align with their core proposition anymore. No redirects, no rewrites, no second consolidation, just removal. And at the end, that has resulted in an increase in organic search traffic, as we can see here, confirming once again that quality beats quantity. An important clarification here before moving forward and doing this with your own site. As you can see here, an in-depth content audit is very much needed to assess Which are those actual pages that are not needed anymore. So it's important that you do that as well. And finally, I want to thank Similarweb, who are today's sponsor. Similarweb is one of my favorite tools to analyze AI insights, AI traffic. It has this very comprehensive dashboard showcasing the evolution of AI traffic from all AI platforms. The top pages attracting that traffic, and a sample of the prompts, not only for your own site, but also your competitors. If you haven't taken a look yet, go and check it out. And remember to subscribe to the SEO FOMO newsletter, now with more than 40,000 subscribers. And if you want to keep up with the latest in SEO without waiting for the newsletter each week, go and take a look at the SEO FOMO news sites. Thank you very much. And until the next week, bye-bye.
