We were proud to have our own JR Oakes on stage at SEO Week 2025, talking about something that’s hitting almost everyone in our industry right now: ‘Benchmarking Brand Discoverability: LLMs vs. Traditional Search.’
I got to kick around some early ideas for this study with JR, but the real credit goes to our in-house R&D team. Having a dedicated team focused on research, tool development, and workflow innovation is rare in this space, and it’s a big part of why we’re able to spot major shifts early, build new service offerings around them, and help our partners stay ahead of the curve.
Here’s a quick breakdown of what JR presented, why we went down this path, and why it matters.
Resources Mentioned:
View the dashboards: locomotive.agency/SEOWEEK
JR’s Speakerdeck link: Benchmarking Brand Discoverability: LLMs vs. Traditional Search
The Big Idea: How Discoverable Are Brands in LLMs?
We all know search is changing. Traditional Google rankings aren’t the whole story anymore. You’ve got AI Overviews, ChatGPT, Gemini, Claude, and a flood of other LLMs shaping what people see.
We wanted to figure out: how are brands showing up across different LLMs? And how does that compare to traditional SEO?
Nobody else seemed to be tackling this at scale yet, so we decided to.
The study focused on B2B SaaS brands. We started by building a seed keyword set from structured category data on G2.com, then expanded it with 63 different search intent modifiers like “best,” “enterprise,” “affordable,” “GDPR-compliant,” and more. This gave us a broad mix of high-commercial-intent queries.
We queried across 11 different LLM models, tracking changes between versions like GPT-3.5, GPT-4, Claude 2.0, and Gemini 1.5. For each query, we recorded brand mentions surfaced by the model outputs. Not just ranked lists, but any recognizable brand appearances within responses.
Where possible, we normalized brand mentions across different phrasings, subsidiaries, and aliases to avoid duplicate counting.
Finally, we analyzed the data by model, modifier, geography, company revenue band, and social media presence, looking for patterns in discoverability.
SEO Metrics and AI Mentions: There’s Definitely a Link
One of the first things we wondered: do traditional SEO metrics like domain rating (DR) and backlinks still matter when it comes to AI-driven results?
Short answer: yes, 100%.
Even though LLMs don’t “rank” pages the same way Google does, brands with stronger backlink profiles and higher DR show up a lot more often.
Quick refresher if you need it: DR (Domain Rating) is a measure of a site’s backlink strength, based on Ahrefs.
Affiliate and Aggregator Sites Are Feeling It
In the data, one thing was pretty obvious: affiliate and aggregator-style sites aren’t getting the same visibility in LLM outputs.
JR put it well during his pre-conference Podcast Interview:
“You can’t have people creating 12,000 pieces of junk content a year.
You can’t have AI agents crawling websites, using your content without seeing ads, without returning anything to the creators.
That model just isn’t sustainable long term.”
That reality is starting to show up in the numbers. Traditional aggregator strategies, those focused on monetizing traffic through ads and affiliate links, aren’t translating cleanly into the world of AI-driven search and retrieval. Brands that own their products, services, and content directly are seeing stronger visibility across LLMs.
Some Other Cool Findings
More Brands Are Being Surfaced: Newer LLMs are bringing back 2–3 times more brands per query compared to older models. More opportunity, more competition.
Different Models, Different Results: ChatGPT, Gemini, Claude all show different sets of brands for the same query. Figuring out what impacts each is the next step.
Geography Still Matters: LLMs are getting better globally, but the U.S. still gets the most love. Where your company is based can impact visibility.
Social Media Helps (Maybe): Brands with active social presences seem to pop up a bit more. Could be a correlation, but it’s worth noting.
LLMs Lag Behind on New Stuff: Google still picks up new brands faster. LLMs take a little longer to catch up.
Why This Stuff Matters
If you’re only chasing Google rankings, you’re missing a big piece of the picture.
Search is splitting into conversations, snapshots, agents, and who knows what else next. You can’t just think “ranking.” You have to think “discoverability” everywhere.
That means doing the hard stuff: building a strong brand, earning real links, putting out great content, and giving LLMs something worth pulling from.
At LOCOMOTIVE, we’re built for the hard stuff: helping shape strong brands, earning real links through Digital PR, and putting out content that LLMs and search engines actually want. We saw the wave coming long before it hit, and it was one of the big reasons that we grew our branding and Digital PR services to make sure our partners are ready to win in a world where trust and authority are everything.
Check Out the Data
JR included a link near the end of his talk that gives public access to a version of the dataset we used. You can find the dashboards at: locomotive.agency/SEOWEEK, and his Speakerdeck link here: Benchmarking Brand Discoverability: LLMs vs. Traditional Search.
We hope it helps more people get a handle on how brand discoverability is changing.
Big thanks to JR for driving this whole project, and to the whole LOCOMOTIVE team for making studies like this possible; thanks to Savannah and Erick for the excellent visuals, and Patrick Stox and Kari Sugarman for reviewing.
Reflections & Takeaways
SEO Week 2025 was one of the smartest and most polished shows we have ever been to. It was an immersive and data-rich experience with high caliber attendees from all over the world.
Wil Reynolds delivered a standout presentation highlighting how “LLMs don’t know the truth. They know probability. Search engines penalize authenticity… And what do they reward? Being like everybody f*cking else. Which means that we, as search engine optimizers, reject authenticity.”
Dawn Anderson did a whirlwind tour of current Information Retrieval and how some in the field are out of touch with what SEO is today.
And in a touching tribute to industry pioneers we’ve lost, Mike King paid homage to Russ Jones, Hamlet Batista, and Bill Slawski through specially themed shirts he wore each day of the conference.
The Lavan in Midtown provided an exceptional setting with its three digital projection walls that created an immersive presentation experience. Production-wise, this was one of the best conferences we’ve ever attended.
Huge thanks to Mike King, Garrett Sussman, Charlene Kate Ditch, and others from iPullRank and beyond for putting together an incredible SEO Week 2025. To take a vision and turn it into reality by placing the right people in the right places was just incredible.
As Mike said to all of us there, “You know when you have an idea for something and it doesn’t work out the way you want it to? This is the opposite of that.”
Catch ya’ll down the line!
– Ryan Huser, VP of Growth at LOCOMOTIVE
