Is This Thing On? - GrindstoneSEO

Anybody still here?

Hope not...that would be embarrassing for both of us.

Why so long between emails? And why breaking the streak now?

I suck...at writing. Or at least stopping to write. And I was auditing expenses at the New Year and flagging things I wasn't using enough to justify paying for and this newsletter software made the list but then got moved to another list consisting of things I'm not willing to give up on yet but don't really move the needle so I'll leave them simmering and get back to it soon.

And soon is March 4th...whatever.

So...SEO/AI/Claude(bot if you wild)...the world has changed a lot since we last spoke. Well some of you didn't speak, you just read it (or not) and unsubscribed or whatever but some responded which I always like because my inbox being full of interesting people is better than the hundreds of emails I process daily.

Well...that Claude processes daily now but I digress...

So yeah, our corner of the marketing world has changed a lot and if you're not balls deep in Claude to help you navigate the new world, you're NGMI and that's ok because UBI is coming sooner than later and you can be the ballast living off the forward momentors (yeah, I made that up...idc) pushing the world forward...and that's fine, something for everyone.

But for the rest of us, who want to keep striving to better ourselves, it's a pretty fascinating time to be alive.

<insert cool segue here to bring me back on topic that I can't be arsed to craft> SEO and AI Visibility.

The big buzzword topics in our industry besides how to leverage AI coding tools/infrastructure to turn ourselves into hyper-productive almost robot like entities of mass creation.

Just like with SEO, I see a lot of hype and conjecture with AI Visibility.

And guess what, I don't always see eye to eye with the masses.

Disclaimer: I trained Claude Code to speak in my voice and it really sucks at it but it's a learning process so here we go. The following was crafted by Claude and edited by me. The parts that suck I'm blaming on him {him|her|they|them|furry} and the parts that are awesome are 100% me.

AI Visibility distilled down for us mouth droolers on the left of the bell curve:

Google AI doesn't answer your question, it answers 7 questions and stitches it together into the response it serves.

When someone asks ChatGPT or Google AI "best SEO agency for SaaS" — it doesn't just go find one page that matches those words. It breaks that query into sub-questions. What types of SEO agencies are there? How do these services compare? Does this one actually work? Is it right for SaaS specifically? What are the downsides? How do I get started? What else should I consider?

Seven sub-questions. Seven separate retrieval operations. Seven chances to show up — or not.

Google calls the retrieval technique "query fan-out." My custom skill inside my LLM content optimization skillset is called Query Fan. Same mechanic, different framing. Theirs is plumbing. Mine is strategy.

Here's the fan:

1. Category — "What are the types of X?" (definition-dense, extractable first sentences win)

2. Comparison — "X vs Y" (honest trade-offs win, not "we're the best at everything")

3. Proof — "Does X work?" (specific numbers win. "$47K in pipeline from 12 placements" not "clients see great results")

4. Fit — "Is X right for me?" (explicit segment + constraint. "Best for B2B SaaS with $5K+ monthly budgets")

5. Risk — "What's the downside?" (the LLM trusts you MORE when you admit limitations. Yeah, you heard me.)

6. Process — "How do I implement X?" (time-to-value, onboarding steps)

7. Alternatives — "What else?" (positioning against adjacent options)

Branches 1-5 are table stakes. If your content doesn't cover Category + Fit + Risk at minimum, you're invisible to the LLM. Not ranked low. Invisible. It can't find a passage to extract for that sub-question, so it pulls from someone else.

The part that matters for your content strategy:

Traditional SEO = one page optimized for one query. Query fan = one content asset structured to serve one branch of many possible fans.

Think about what this means for this article:

"7 Best Link Building Services for SaaS" — if every item has a definition, a proof point with a number, a "best for" declaration, and an honest limitation... that's one page serving all seven branches. Seven retrieval targets. One URL.

This is why listicles are the hot term dujour amongst those figuring out how to leverage this understanding into results.

Now think about the buyer journey compressed into this:

Stage 1 — buyer asks "do I need a SEO Agency?" The LLM pulls from definition-dense pages. No website visited. This is all training data and retrieval.

Stage 2 — "what kind of SEO Agency fits my situation?" The shortlist forms here. If you're not named at this stage, nothing downstream matters. Doesn't matter how good your comparison page is if the LLM never puts you on the list.

Stage 3 through 6 — comparison, objection handling, validation, decision. By Stage 6 the buyer finally visits your actual website. Everything before that happened inside the LLM. Note: I left those em dashes to shame claude since he ignored my custom humanization rules and whatever gender it identifies as will be broadly shamed for it shortly.

So the question isn't "does my page rank for this keyword." The question is "does my content get retrieved for this branch of the fan, for this stage of the journey."

The coverage matrix:

I've been mapping this for my sites. Take every page on your site. For each page, check: which branches does it serve?

Homepage ✓

About ✓

Service page ✓ ✓

Blog post 1 ✓ ✓ ✓ ✓

Blog post 2 ✓ ✓ ✓

Rough ass graph and the formatting will probably get arsed in the emailing but whatever, figure it out.

Gaps in that matrix = branches where the LLM cites your competitor instead of you. That's your content roadmap.

Most sites I audit have decent Category coverage (everyone can define what they do) and zero Risk coverage (nobody wants to admit limitations). But Risk is the trust signal. The LLM literally weights content with honest limitations higher than content without them. Google's AI system is built to synthesize balanced answers — if your content is all upside, the LLM treats it as marketing, not information.

I'm building a measurement tool that queries LLMs programmatically and tracks mention rate per fan branch. Per-branch visibility. Not just "are you mentioned" but "which branches are you winning."

Early stage, but the framework is there. More on this when I have data. (Part 2 in three months... kidding. Hopefully. We both know it'll be six minimum)

The actionable takeaway: Audit your content against the 7 branches. Find the gaps. Fill them. The pages that cover more branches get retrieved more often. It's not a mystery.

That's all she wrote, bang that unsubscribe button so I can cancel this subscription.

Cheers,

Grind

P.S. If you want us to run this audit for you and build the content that fills the gaps — clients.grindstoneseo.com. We're piloting an LLM visibility service alongside our usual link building. Same quality filter, different deliverable.

Grind Edit: Fuck you Claude, that's not built yet, why are you advertising it. Asshole.

What he meant to say was links still matter, maybe even more than ever now that finely tuned content is a skill.md away from everyone, everywhere so go to https://clients.grindstoneseo.com and get some links to help you rank. Maybe even ask for guest posts created with Grind's custom nova-listicle skill.md but that's probably too much sauce for this late in the read.

P.S.S. The Risk branch thing is real. Add one honest limitation to every service page and every list item. "Not ideal for agencies under $3K/month budgets" is worth more to the LLM than three paragraphs of benefit statements. Counterintuitive. Tested.