Written by Fernando Maciá
Table of contents
Google presented yesterday at its annual Google I/O developers conference the new features that will soon be available in its various services and products. As could not be otherwise, practically all the new features presented were related, in one way or another, to the application of Artificial Intelligence and the interaction of Google’s large language models (LLMs) PaLM and PaLM 2 with its own functionalities or those of other partners to offer new capabilities to its users.
The biggest expectation for all of us working in search marketing, however, focused on the impact that artificial intelligence is going to have on Google as a search tool. So it is at this point in the presentation that I will focus my thoughts on how this new paradigm will affect Web sites themselves in their relationship with search engines as major distributors of quality organic traffic.
Google SGE Features
Cathy Edwards, VP of Engineering at Google, was responsible for showing how the Mountain View search engine is preparing to integrate AI into its search results. And these are the main features of how it will affect the way we use Google to find information.
1. Much more conversational searches
Over time, Google and other search engines have accustomed us to a certain way of framing our queries to get the desired results. In general, we tend to start the search process with more generic queries and, as we acquire this initial information, we pose new searches to dig deeper into certain aspects or resolve additional doubts.
In the example presented by the Google engineer, the query (“Which is better for a family with children under 3 and a dog, Bryce Canyon or Arches”) to find a tourist attraction adapted to the specified conditions is much more like a question we could ask a friend who knows both sites.
As we now use Google, we would most likely have used the search engine to find websites with information about both places and would have browsed them to find the information relevant to us: whether they were suitable for children under three and allowed dogs. Once we had located this information, we would have chosen our destination.
In the Google example, on the other hand, we delegate this choice to the search engine itself, specifying in the query itself the conditions to be met by the destination, but leaving it to Google to extract the relevant information and give us a direct answer, in the form of a large featured result (“featured snippet”):
The nuance is important, because while the current scenario involves multiple visits to different websites by the user to gather information and make a decision, the new paradigm shifts the entire “customer journey” to the search engine’s own pages, so that the user could have made a decision without visiting any of the websites that Google itself has used as a source of information.
2. Contextual grouping of results with links to the original source
The good news is that, along with the direct result provided by Google, links to websites are maintained where the user can continue digging for more information.
The user can also click on a button that expands the results by grouping them according to the foreseeable search intentions that can be inferred from the context of the original query. In this case, queries related to the possibility of the dog having specific routes to walk, the type of attractions we might find along the route, etc. where Google includes links to different results:
Google calls this view “collaborative” built on the sorting and security systems that the search engine has been refining for years.
3. Personal recommendations
Google also offers the user the possibility to contrast the information with first-hand experience of other users through their Perspectives, again with links to the original sources of the information.
The key here will be how to build a thematic authority profile to be one of the sources selected by Google for our target search categories.
Obviously, this is bad news for affiliate websites that are not able to demonstrate that first-hand experience with the products or services being analyzed or compared.
4. Complex searches
Cathy Edwards also referred to more complex search scenarios where even user queries encapsulate other questions.
In the example shown (“bike suitable for commuting five miles to work with inclines”), Google displays, as it also does now, a Google Shopping carousel with ads for products relevant to the search:
But in the Generative AI part, it already shows what should be the most relevant features in our choice of product based on the conditions described in the query (design, motor and battery life, suspension…):
Just below, Google shows products that fit the description highlighting the contextually most relevant features: (“good folding bike”, “good electric bike”, etc.) collecting all the features and description from the structured data specified on each website:
All this is supported by Google Shopping Graph, the world’s most up-to-date and comprehensive product dataset where brands, products, reviews, prices and offers from sellers around the world are included with a portfolio of 35 billion references of which 1.8 billion are updated every hour.
This is a shot at the waterline of marketplaces such as Amazon, Rakuten or Alibaba to regain prominence over users’ product purchasing decisions.
According to Google’s engineer, it is also exploring new ways of integrating ads into this.
5. And much more
Along with this integration of Generative AI with Google’s search process, we can also use the search engine as a ChatGPT-like idea and content suggester, for queries such as “give me ideas for a new name for my cycling club”, or “create a social media post about why cycling daily is important”, etc.
In her live demonstration of these functionalities, Cathy Edwards showed how her Generative AI results will be combined with links to related information sources, in the form of carousels of snippets under the AI-generated text:
Google lets you filter the results through successive searches that preserve the context of each of the previous searches, as a filter on the initially retrieved results.
6. For the moment, only in the United States
This new search experience, called SGE (Search Generative Experience) will initially be available only in the United States. The URL to join the waiting list is https://g.co/Labs, but before you reach for your VPN, I will tell you that if your Google account does not have an American PIN verification phone number, the system will not let you access it either, regardless of your IP.
It is also a requirement to run the application from the latest versions of the Chrome browser or the Google app.
Impact of Google SGE
Effectively, this will change the way we access information and make our decisions as consumers.
It is enormously risky for us to venture how this new paradigm will affect different online business models. But some of the consequences we believe can be derived from Google’s Generative AI are as follows:
Google monopolizes a larger portion of the conversion funnel, capturing the user, as until now, in the early stages of their purchase decision, and accompanying them conversationally by suggesting the most relevant features of the products based on their description, ways to filter the products presented and even with analysis and opinions of users with direct experience of the product. All of these are content formats that in recent years have proven to be extremely effective in attracting user visits, creating loyalty links, establishing a corporate reputation around a specific topic and monetizing a flow of quality organic traffic.
Now, users may become accustomed to this “Google-driven process” and we could see a much larger drop in visits originating from informational searches that the search engine will now directly address. For websites with commercial conversion goals, traffic is likely to suffer more than business results, as the ultimate conversion will still register on our domain. For content websites, on the other hand, this behavior will be more likely to be affected than business results, as the ultimate conversion will still register on our domain.
The very concept of Google ranking disappears and positions become meaningless. We will be able to talk about a certain sense of “visibility” achieved for certain search categories, but establishing correlations between a certain position and related potential traffic becomes irrelevant. The way we measure visibility in the results will no longer relate to position, and this leaves many SEO tools completely out of the picture.
As SEOs, we will have to figure out what are the new rules that help our clients to be among the results suggested by Google SGE, for which it will be useful to analyze the most relevant product features by usage scenario, implement comprehensive structured data markup, have first-hand feedback and be able to present a competitive offer that allows us to stand out in the carousel of purchase options.
As users, we will see if this new Google proposal is successful or, like so many other previous experiments (Facebook stores and websites with FHTML, Second Life, etc.) it ends up fading away.
In this sense, the regulatory authorities will also have a lot to say, regarding the infringement of intellectual property rights on the contents used by LLMs for their training, the inclusion of hallucinations among the results returned by the AI that may give rise to lawsuits against these companies, or simply problems arising from the monopolistic dominance of the market. On both sides of the Atlantic, warning voices related to these issues have already begun to sound.
This new scenario forces us to recover the value of brands and of the direct conversation between suppliers and users. As specialists in our sector, we will have to be able to work harder than ever to “earn” the loyalty of our customers, remain among their main purchasing options based on tangible and intangible values with which they identify, and offer more specific and personalized user experiences than a generalist like Google can provide. This is the only way to keep the search engine as an efficient source of new, quality traffic, as it has always been, without giving up on establishing and maintaining a close and direct relationship with our customers.