Hello All,
Over the last 18 months we have all been participating in one of the fastest changing digital landscapes since the browser wars of the early 2000s. With the broad adoption of large language models such as ChatGPT, Claud, and Gemini, it is abundantly clear that a new search environment is emerging. As in the past, there will be differences in the behavior of LLM searchers compared with Google and, while there are many points to consider, I’ve noticed that the “quality” of site visitors from LLMs contrasts significantly with that of traditional search engines.
Tracking KPIs such as bounce rate, engagement rate, time on site, has showcased that there is a notable quality element to the visitors from LLMs. To generalize, sites’ visitors originating from LLMs appear to just convert at a higher rate. This comparison is pushing marketers to ask the “why” behind differences in traffic segments and we have several working theories.
Our current hypothesis is that traffic from LLMs have organically moved through the marketing funnel in the process of finding the desired result. We all remember the marketing funnel (at least the first several stages):
- Awareness: Potential customers discover your brand
- Consideration: Prospects evaluate a brand’s product against the marketplace and competitors
- Conversion: Prospects are ready to purchase or take the desired action
Now when we think about a classic Google search, let’s say someone is trying to complete their living room’s new look and needs a couch. Historically, someone will go to Google and search for “couches for sale” which we know there are approximately 90K monthly searches for, and then they will be presented with the classic search engine results page featuring search ads, query refinements, map packs, and organic listings. This can be overwhelming and present many different options for what and where to buy their next piece of furniture.
However, when I conduct a similar search within ChatGPT I am presented with options to refine my search as if I was talking to a salesman on the floor of a large retailer outfit. “Here are your options for budget friendly couches, mid-range, reclining, premium, etc…” with a prompt that then asks me to refine my search by:
- your budget
- room size
- style (modern, cozy, luxury, etc.)
- or even local stores near (you)
As I enter in more information and the LLM provides more specific results, I am inherently narrowing my consideration set and nearing a conversion event. At the end of that journey is the result that ChatGPT thinks is the best for my given needs, which we all implicitly trust and will then purchase. This is precisely why LLM-referred visitors convert at a higher rate; they don’t arrive at your site to begin their research, they arrive to complete it. The whole funnel work has already been done inside the chat window. Additionally, the unknown element, once we make a purchase, or thumbs up a LLM conversation which led to said purchase, how does that then impact our next shopping journey? We know that LLM tools store data and conversation histories, and previous engagement is likely to impact future results, further complicating an emerging channel.
This behavior pattern is expanding across all industries, with some more impacted than others; however, in the long run, every brand should consider their value propositions, their website’s technical configuration, and site content to answer the question of “are we appearing in the right way within LLMs to drive revenue?”
As this channel evolves, the real question is: are you positioned to capture high-intent traffic from AI-driven search?
We’re already helping brands rethink how they show up across LLMs, from content to structure to discoverability. If you’re starting to explore this shift, let’s compare notes.
Thanks for reading and have a good one!
