AEO and AI-Driven Search, Online & Offline

Part 1 of 2

Software tools exist to help people get things done. But to get from point A to point Z you need to use their interface, follow their data logic and flow, and accept their output. The better tools provide degrees of flexibility to enable users to adjust how they work to achieve the output they want. This is why UX is so important: UX is the place where the consumer gets closest to being able to manipulate the outcome. It works this way even when you are adjusting which words you use in the monolithic Google search text field.

What we don't think about is how much our software tools have manipulated our own ways of working. We Internet users have our experience obtaining information molded completely by the development of search over the past 30 years. Although it was introduced in more limited formats much earlier, search for consumers started out in 1994, then it grew into a useful convenience with AltaVista introducing natural-language search in 1995, until what we had from about 2021 was the search field doing double duty as the place where you could also simply type in the URL.

Now there's AI-driven search, which both enhances and confuses the consumer experience. In part 1 of this 2-part series, we'll explore ways that brands and retailers can effectively compete in this new environment by understanding exactly how AI search works relative to conventional search. In part 2 we'll turn the model on its head and explore alternative approaches for both retailers and brands.

You Can't Escape the Venn Diagram

Traditional search results haven't really changed much in the last 30 years. It is a list delivered in a tabular format, ordered from most relevant to least relevant. What determines relevancy is the search algorithm. But it has become less and less pure science: First there were ways to cheat the system, then as those loopholes closed the search engines themselves created their own monetized loopholes. Anyone could agree to pay to be at the top of the search listings, with the results becoming just another high-speed auctioned advertising medium.

This is highly ironic, because search results are supposed to reflect the search terms we use. Search terms are, in effect, an abbreviated brief that are inadvertently incomplete or misleading because we are humans and many words can be both nouns and verbs. Thus, in traditional search the exchange with the search algorithm goes something like this:

• Consumer: Here's what I want

• Algorithm: Here's what you asked for PLUS here are suggestions that brands have paid me to show you based on your data

• Consumer: No, this isn't quite right, let me add some clarifying words

• Algorithm: OK, based on the data I have about you plus your revised search terms, here you go PLUS here are suggestions that brands have paid me to show you based on your data

• Consumer: This isn't for me, it's for someone else, so please just use these revised search terms

• Algorithm: OK, here's what you're asking for PLUS here are suggestions that brands have paid me to show you based on your data

• Consumer: ARRRGH

Previous readers of this column know my wife's hack for this: Search for what you want and then start at the bottom - what the algorithm thinks is least relevant - where you may discover more interesting options. These same readers may also recall my suggestion that a better way to display search terms would be in the form of a 3-circle Venn Diagram, where what the algorithm thinks you want sits at the intersection of all the circles, the next most relevant options sit where 2 circles overlap, and then the other options sit within one of the circles. This solution sidesteps the prioritization inherent in a list format, while encouraging the searching consumer to explore some of more interesting options because they aren't being excluded by the implied hierarchy of page 1 results being "better" than those on page 2 and below.

The real-life analog is going into a wine shop (something I know entirely too much about). Someone who knows nothing about wine and needs to bring a host-gift for a party will seek out the shop owner or a sales associate because they know their selection, and they will steer you toward something at the right price that will deliver the right amount of "wow" from the host. The job of the sales associate is to run the query for you live, in real-time, by asking you key questions to help you refine your brief.

Which brings us to AOE.