The "Ultimate Guide" Is Dead. Long Live the Focused Answer.
Here's a confession that might get my SEO friends to revoke my marketing card: I've spent years telling teams to "go deeper, go longer, cover every angle." Build the definitive resource. Be the Wikipedia of your niche.
Turns out, when it comes to getting cited by ChatGPT, that advice is about as useful as a QR code on a billboard.
A massive new study just dropped that should make every B2B marketer rethink their content playbook – and honestly, it's one of those rare moments where the data confirms what our gut has been whispering all along.
The Study That Changes Everything
Kevin Indig's analysis at Growth Memo, conducted in partnership with AirOps, examined 815,000 query-page pairs across 16,851 queries and 353,799 pages. That's not a sample size – that's a census.
The headline finding? Pages covering 26-50% of ChatGPT's subtopics get cited more than pages covering 100%. Let me say that again for the folks in the back who just commissioned a 5,000-word "Complete Guide to Everything": moderate coverage outperforms exhaustive coverage.
The "ultimate guide" strategy – the one we've all been running for a decade – produces worse citation results than a focused article that covers two to three related angles well.
Marketing is like dating, remember? You don't propose on the first ad impression. And apparently, you also don't dump your entire life story on someone who just asked what time it is.
What Actually Gets You Cited
So if comprehensive coverage isn't the golden ticket, what is? The study identified two signals that actually predict whether ChatGPT cites your page:
1. Retrieval Rank Is King
A page at position 0 in ChatGPT's web search results has a 58% citation rate. By position 10, that drops to 14%. Pages cited in all three test runs had a median retrieval rank of 2.5. Pages never cited? Median rank 13.
This isn't surprising if you think about it. ChatGPT isn't reading the entire internet for every query – it's pulling from what surfaces first and making decisions fast. Sound familiar? It's basically how your prospects skim your emails.
2. Query Match Is the Content Signal That Matters
Pages with headlines that directly answer the question get cited 41% of the time. Pages with loosely related headlines? 29%.
Here's the kicker: even among top-ranked pages (positions 0-2), higher query match adds 19 percentage points to citation rates. Your headline isn't just a hook for humans – it's a signal to AI that says, "Yes, I'm exactly what you're looking for."
Meanwhile, fanout coverage, word count, heading count, and domain authority? All secondary. Some are flat. Some are inversely correlated.
Data tells you the what, but brand tells you the why. And right now, the data is telling us that AI doesn't care about your domain authority flex.
The Wikipedia Exception (And Why You Can't Copy It)
Before you throw out everything I just said, there's one outlier worth noting. Wikipedia has the worst retrieval rank in the dataset (median 24) and the lowest query match score (0.576). Yet it achieves the highest citation rate: 59%.
How? Wikipedia pages average 4,383 words, 31 lists, and 6.6 tables. They're encyclopedic in the literal sense – exhaustively structured and cross-linked across millions of topics.
But here's the thing: a 3,000-word corporate blog post with 15 subheadings is not Wikipedia. You don't have millions of interlinked pages. You don't have decades of trust signals. Trying to replicate Wikipedia's density strategy with your B2B content is like trying to replicate Beyoncé's stage presence at your company all-hands. Admirable ambition, questionable execution.
The Bimodal Reality We Need to Accept
Here's where it gets uncomfortable: 58% of pages retrieved by ChatGPT are never cited. 25% are always cited when they appear. Only 17% fall in between.
The always-cited and never-cited groups look nearly identical on most content metrics – similar word counts (~2,200), similar heading counts (~20), similar readability scores, similar domain authority (~54).
The on-page signals we obsess over don't separate winners from losers. What separates them is retrieval rank. The retrieval system is the gatekeeper. Everything else is a tiebreaker.
Those "mixed" pages in the middle – the 17% that get cited sometimes and ignored other times? They have the highest word counts, the most headings, and the highest domain authority. They are the ultimate guides. They are also the least reliable performers.
Let's not get seduced by the shiny object syndrome. More content isn't better content. It's just... more.
The Shrinking Citation Pool Makes This Urgent
If you're thinking "I'll figure this out eventually," consider this: ChatGPT Search is citing fewer sites per answer. According to Resoneo's analysis of 27,000 comparable responses, the average number of unique domains cited per answer fell from 19 to 15 after the move to GPT-5.3 Instant.
Four fewer winners per answer. The citation pool is shrinking, which means the stakes for getting it right are rising.
What This Means for Your Content Strategy
As Ted Fay noted on LinkedIn, this research reinforces something we should have known all along: what's good for people is good for the bots.
Here's the playbook:
- Match the query directly in your headings. Not adjacent to the query. Not thematically related. Directly.
- Stay in the sweet spot: 500-2,000 words. Enough to be substantive, not so much that you dilute your focus.
- Use 7-20 subheadings. Enough structure to organize, not so much that you're padding.
- Build the page that is the best answer to one question. Not the page that adequately answers twenty.
Here's a quick test, courtesy of Ted Fay: What is the question that your page answers? If that's not clear – and also answered in the beginning of the page – you're not only losing to the bots, you're losing your readers.
The Return on Imagination
Every CMO loves to talk ROI, but let's not forget there's also "Return on Imagination." The imagination here isn't about creating more – it's about creating better. More focused. More direct. More useful.
The irony isn't lost on me: in an age of infinite content, the winning strategy is restraint. In a world where AI can process everything, it rewards the content that doesn't try to be everything.
Marketing is a marathon with weekly sprints. But this particular sprint? It's about learning to say less, better.
Build the page that answers one question brilliantly. Let someone else write the ultimate guide that answers twenty questions adequately.
The AI will know the difference. And increasingly, so will your audience.
