“What’s actually going on with GEO?” is probably the question I get asked more than anything right now. More than page speed, more than Core Web Vitals, more than “is SEO dead?” (for the record, it isn’t).
The honest answer is that GEO (Generative Engine Optimization) is still early, and anyone telling you they’ve cracked the code is either selling something or not looking at their data closely enough. But after 6–9 months of testing GEO tactics across our client base, studying real GA4 data from 15 properties and tracking citation patterns across hundreds of monitored prompts and thousands of sourced domains from citations, we’re starting to see patterns emerge from the noise.
It would be easier if the models weren’t also changing in real time. At times it felt like whack-a-mole. But here’s where things stand.
The simplest way to think about it
SEO is linear. A person types a query, Google returns a ranked list of results, and your job is to be as high on that list as possible. It’s a system we’ve all understood for years: match intent, build authority, earn the click.
GEO is more holistic. When someone asks ChatGPT, Perplexity, Gemini, or any other LLM a question, the model doesn’t just pull one page. It synthesizes across multiple sources and topical angles to construct an answer. Your brand’s visibility becomes less about ranking and more about citation likelihood — whether or not the model pulls you into its response.
Same goal (be visible when people look for answers), very different mechanics.
Where the key differences show up
How results are generated. In SEO, results align directly to the keyword searched. It’s a one-to-one relationship between query and SERP. In GEO, models “fan out,” pulling from both direct and indirect sources to generate a citation response. Your content might get surfaced for queries you never explicitly targeted, simply because you covered a topic with enough depth and authority.
What drives authority. In traditional SEO, backlinks are the backbone. Links from reputable sites signal to Google that your content is trustworthy and authoritative. In GEO, brand mentions alone appear to carry more weight than links. The models seem to care less about who linked to you and more about how frequently and consistently your brand appears across the web in relevant contexts.
How many algorithms you’re optimizing for. SEO, for all its complexity, is still primarily one algorithm: Google. Yes, there’s Bing, and yes, Google’s algorithm is a moving target, but the optimization framework is singular. GEO introduces a multi-model landscape. Gemini, Perplexity, ChatGPT, Claude, and others all prefer different content and weight different signals. It’s case-by-case optimization, which is exactly as fun as it sounds.
How you target demand. In SEO, keywords are targeted based on empirical search volume data. You can see the numbers, plan around them, and forecast with reasonable confidence. In GEO, prompts are monitored using estimation tools that are, to be blunt, guessing at wording. We’re using tools like ReachLLM to track citation patterns, but the prompt data is inherently fuzzier than keyword volume. It’s a different kind of measurement muscle.
How you measure success. SEO’s core leading KPI has long been non-brand organic clicks, proof that people who don’t already know you are finding you through search. GEO measurement is more subtle. It shows up through LLM referral traffic in GA4 and branded search lift, which creates attribution gray areas that can be uncomfortable for teams used to clean reporting. We’re getting better at tracking it, but there’s no Google Search Console equivalent for LLM visibility yet.
Not all LLMs are created equal
This is the part that’s been the most eye-opening in our testing: each LLM has its own content source preferences, and they don’t overlap as neatly as you’d hope.
For example, behemoth industry sites (think the major platforms in your vertical) tend to get cited across the board by Claude, Perplexity, ChatGPT, and Gemini. That’s the closest thing to a universal signal we’ve found.
But beyond that, the preferences diverge. Reddit and Quora content gets picked up heavily by Claude, ChatGPT, and Gemini, but Perplexity largely ignores it. YouTube is a preferred source for Perplexity and Gemini but not as much for Claude or ChatGPT. Government and academic sources matter for Perplexity and ChatGPT but barely register with Claude or Gemini. Third-party reviews and publishers? Claude and ChatGPT pull from them, while Perplexity and Gemini don’t seem to weight them as heavily.
The practical implication is that a one-size-fits-all GEO strategy doesn’t work. If your audience skews toward Perplexity (and we’re seeing that in certain B2B verticals), your content mix looks different than if they’re primarily using ChatGPT. This is a fundamentally different optimization challenge than SEO, where Google’s algorithm — however complex — was still one system to study.
What this means practically
Here’s the part I want to be direct about: SEO isn’t going away, and GEO isn’t fully replacing it. We’re operating in two overlapping systems now, and each rewards slightly different behavior.
If you’re a marketing leader, the takeaway isn’t “drop everything and pivot to GEO.” It’s more nuanced than that. Your SEO fundamentals (content depth, intent alignment, technical health, authority building) still matter. In fact, a lot of the content work that makes you competitive in SEO also makes you more citable by LLMs. Good content is good content.
But you do need to start thinking about a few things differently. Your PR and content strategy should account for brand mentions, not just links. Your measurement framework needs to include LLM referral traffic alongside traditional organic metrics. And your team should be aware of which models your audience actually uses — because the content each model prefers varies more than you’d expect.
At E4C5, our approach has been to run this the same way we approach everything: test, learn, deliver, repeat. We’re isolating GEO tactics in the same way we isolate SEO variables, running small pilots, watching the data, and only scaling what shows real signal. It’s not glamorous, but it’s how you avoid chasing noise.
The chess analogy (because of course)
This moment feels a bit like when you realize you’ve truly entered “the open” in chess. Your opening theory and common sequences are over, and now you’re in the improvise stage of the game: the middlegame.
SEO feels more like theory in chess. You can pick your opening on white or black and lean into the long-term strategy you prefer because you already know the moves. The lines are well-studied, the patterns are familiar, and the strongest players have clear advantages.
With GEO, the theory is starting to come into view, but it’s still too volatile to know for sure. We all need to get comfortable navigating the open for the near future: leaning on the strategies we do know, and also being very aware of everything that is still undefined. Still early days, but it’s starting to feel less chaotic. More to come. 🧪
If you’re trying to figure out how SEO and GEO fit together for your business, I’m happy to talk through it.
I offer a free initial strategy consultation where we can look at your current positioning and talk about what’s actually worth prioritizing in this new AI search landscape. Book a time with me here.