OpenAI launched ChatGPT search a few days ago, available on both the Web as well as through the app to ChatGPT Plus and Team users. In their announcement, OpenAI described it as a more natural, conversational way to obtain quality information more easily:
“By integrating search with a chat interface, users can engage with information in a new way, while content owners gain new opportunities to reach a broader audience. We hope to help users discover publishers and websites, while bringing more choice to search.”
Although they have confirmed to continue using Bing among other services for their results, they’ve also partnered with news and data providers to add up-to-**** information and new visual designs for categories like weather, stocks, sports, news, and maps.
In the announcement, they’ve also highlighted how chats now include links to sources, such as news articles and blog posts, giving users a way to learn more.
However, we also know that despite the hype, so far AI hasn’t really taken any marketshare from Google. Can the new ChatGPT search results change this? Let’s take a look.
The anatomy of ChatGPT Search Results
ChatGPT now shows a Web “Search” icon with a globe within or next to the search box, which, if you select will be highlighted in blue, and suggest some topics to search.
After you perform your search you’ll now get a response that can be multi-modal -depending on the query needs-, showing not only text, but also images, videos, charts, maps.
Now the answer will include in-line clickable citations that will go directly to the external page source, and that in desktop, you can hover to get a preview of the page information.
Additionally, you’ll now see a right sidebar, featuring the citations at the top, followed by search results, listing relevant pages to the queries. Both, citations and search results, show the name of the site along with the favicon, the page title, and relevant page text snippet preview.
This interface blends the chat with a “search results like” interface.
Let’s see how this new interface fulfill different types of search needs and if the quality of results can compete with Google’s.
ChatGPT Search Results vs Google SERPs
When comparing the experience with the results of ChatGPT Search vs Google SERPs, ChatGPT Search Excels with Informational, long-tail conversational Queries, mainly due to the summarized answer format, highlighting the sources that are easy to check on, making it easier to obtain a satisfactory answer while allowing users to dig for more in relevant sites; more than Google, which didn’t provide an AI Overview yet for most of these queries, but highlighted reddit as a top results; for example, for “how to learn SEO for free”:
Something similar happens with “michelin starred restaurants in the Basque Country”, for which ChatGPT provides a more comprehensive answer, categorizing the answer based on the stars along with the location and a short description of each, and while Google provides an AI Overview, listing each restaurant, the list is not as well structured as the one of the ChatGPT answer.
It’s important to note that both results feature images, although when clicking on them, just showcase a bigger version or source of it, rather than the one of the page of the restaurant where it belongs.
Another scenario in which ChatGPT Search provides a slighted more comprehensive experience than Google search results is for comparison queries, for example “levi’s 505 vs 501”, providing a summary of the main characteristics of each, along with the key differences, while highlighting a YouTube video comparing both, allowing to easily check on the sources through the sidebar if more is needed.
Google provides an AI Overview directly showcasing the differences between the two, with a TikTok video shown as a source in the AIO sidebar, followed by the best organic search position: A relevant Levi’s site page guide.
On the other hand, there are also scenarios in which Google Search Results are way more satisfying, providing a more fulfilling format to the search need, for example “padel courts in miami”, although for other more popular searches (such as restaurants or parks) ChatGPT localized the results providing a clickable map, in the case of this one, it didn’t localize it, making Google Search Results with the Map Pack and organic search results more fulfilling.
Something similar happens with more commercially focused queries: Google search results will better satisfy users search intent thanks to its search features and format, better showcasing commercially focused results directly to the user.
For example, for “best headphones for remote work”, although ChatGPT results will directly list what can be considered top headphones, it will link to citation sources which are not the sites or pages selling them, but those reviewing along with many others; making it more difficult to those users with a higher commercial intent.
On the other hand, Google features product carousels (along with top organic search results) allowing users with a more commercial intent to go through the products prices, reviews, etc. and being a click away from buying via the product knowledge panel, complementing this with a blend of more informational results from guides and UGC.
Something worse happens with “dress coats for women”, a less popular query for which ChatGPT gives a high level, informational answer, likely due to the query wording, not satisfying a more commercial intent, which Google does again through a blend of top ranked product listing pages and clickable product carousels displaying products knowledge panels, summarizing products information, linking to product detail pages to buy them.
TL;DR: So far ChatGPT Search results are more satisfying than Google for highly informational queries due to its conversational focused format, although less so for commercially focused ones, in which Google excels due to its blend of sources satisfying different intents as well as search features, facilitating users to take the next steps in their journey.
If ChatGPT wants to compete with Google in the commercial front, it will need far more data, better integration and features to better address user search intent at that stage of the customer journey.
