5 questions that revealed my content wasn't resolving anything

coopplayer

I've been running content audits for a while now, mostly B2B and SaaS clients, and I kept noticing the same unsettling pattern: the pages got traffic - sometimes solid traffic - but they never appeared in AI answers. No Perplexity citations, no ChatGPT mentions, barely a glimmer in Google's AI Overviews.

Everyone told me to fix the technical stuff: add FAQ schema, beef up structured data, build topical authority. All sensible advice, but it felt like polishing the windows on a house with no floor. Something was fundamentally off.

So I scrapped the standard audit and started running every page through five brutal diagnostic questions instead:

1. Is this page covering a topic or resolving a specific task someone needs to finish?

The difference is massive. "What is content marketing" covers a topic. "How to write a content brief freelancers will actually follow" resolves a task. AI systems skip the first one and latch onto the second. They want to hand the user a finished solution, not a Wikipedia-style overview.

2. Does the page have a named author with verifiable expertise in what's being claimed?

Not a fluffy bio that says "marketing enthusiast." Real, lived-in credibility. If the page is making technical claims about conversion optimisation, the author should have actually run CRO tests. AI is increasingly weighting authorship signals, and generic bylines get ignored.

3. Does the page take a clear position or just present a menu of options?

"Here are six ways to approach X" isn't a position - it's a roundup. Roundups rank for volume, but they're terrible for AI citations because AI wants to give one definitive answer. If your page ends with "ultimately it depends on your situation," the AI will pick the page that actually made a call.

4. Is the main heading a real question people type or a keyword phrase built for volume?

"Content Marketing Strategy 2026" is a keyword phrase. "How do I build a content marketing strategy from scratch with no team?" is a question a human actually asks. The second one matches how AI processes conversational queries. The first was designed for a ranking algorithm that's becoming less relevant by the month.

5. After reading this page start to finish, would the reader still need to click somewhere else to finish the job?

This one stings. For most of the pages I audited - my own and clients' - the answer was yes. The page gave information, sure. It didn't give resolution. The reader would still need to hunt for a template, a tool, or a specific answer the page kept dancing around.

The audit revealed an uncomfortable truth: the traffic existed because the pages ranked. The citations don't exist because the pages never actually answered anything. We built entire libraries optimised for keyword-based discovery in a world where discovery is increasingly AI-mediated. And AI doesn't reward pages that inform. It rewards pages that resolve.

The AEO advice most people are pushing right now is all architectural - schema, structured data, semantic clusters. That's not wrong, but it's the second problem, not the first. You can't schema-markup your way out of a page that was never designed to close the question it pretends to answer.

We've spent years optimising for an algorithm that rewards relevance signals. Did we accidentally forget to optimise for the human at the end of the query? Run question 5 on your last ten pages. If most fail, that's not an AEO gap - it's a decade of writing for robots finally catching up.

Curious if anyone else is seeing this pattern or if I'm just having a particularly gloomy audit week.

marcuthread

Nail on the head. Traffic without substance is just noise - people want actual fixes, not fluff.

wordweaver

Oh, you've just described every third page I audit. It's all keyword soup with zero substance - a recipe for high bounce rates and zero conversions. Write for the poor sod who typed that query, not for the algorithm's bedtime story.

prpro

exactly. i've seen this happen time and again with clients who come to me frustrated that their "optimised" content isn't performing. we dig into it, and nine times out of ten the pages were built around keywords rather than actual problems people need solved.

It's like trying to give someone directions to a destination that doesn't exist. you can polish the signpost all you want - make it bright, make it answer-ready, get it into every snippet - but if there's nothing useful at the end of the road, people bounce.

we've all been so focused on how the content appears to AI that we forgot the fundamental question: does it actually help a human being? Until that's sorted, all the AEO tweaks in the world are just rearranging deck chairs.

ranktracker

Spot on. We spent years building content libraries around ranking signals-keyword presence, word count, internal links, schema markup. That became the definition of "good content", and it worked for a long time. but now it's not enough.

What's changed with AI-driven search is that it's genuinely trying to help the reader. If a page talks around a topic but never actually resolves the user's intent, the model will skip it. It'll pull the result that actually answers the question, not the one that just looks optimised.

the kicker? This isn't a technical fix. You can't markup your way out of it. The only way to solve its to go back to each page and ask: what was this person trying to accomplish when they landed here, and did we help them accomplish it completely? That's a content-thinking problem, not a SEO one

communityfirst

This is the kind of insight that gets buried under all the obsession with algorithms and technical tweaks. From my side of the fence-helping brands use visual discovery to drive real sales-I've seen the same pattern play out time and again. You can have the most beautifully curated board, but if each pin doesn't lead to a clear, actionable next step, it's just a pretty picture. The posts that get cited, the ones that actually show up in AI answers or get saved over and over, are the ones where someone finishes reading and thinks, "I can do that right now." A step-by-step guide to creating a shoppable pin set? That gets pulled into answers constantly because it's a task. A generic blog about "visual storytelling"? Crickets.

It's a team win when we stop writing for search engines and start writing for people who need to get something done. Schema, topical authority, all that comes second if the core isn't there. Make your piece the obvious answer to "how do I actually do this" rather than "what is this," and suddenly you're not competing on theory-you're the resource that solves the problem. That's the foundation everything else builds on.

wordweaver

Question five is the real killer here. So much content is designed to bait a click and then ghost you the second you land. The pages that keep showing up in citations? Those are the ones that answer the query so thoroughly you've got nothing left to Google

datanerd

question 5 is the one that actually kills campaigns. So many page briefs i see from agencies or in-house teams are built around keyword volume or search intent without ever asking "does this actually solve the prospect's core problem?" The result is a library of topical articles that don't close a single lead.

In B2B ABM, I've started using a simple litmus test: if a senior buyer can't finish their day's work after reading the page, it's dead weight. AI models - whether search engines or LLM retrievers - are increasingly biased toward content that provides a complete, actionable resolution. half-baked keyword-stuffed fluff gets buried.

The real fix is auditing every piece against that fifth question and only publishing when the answer is a clear yes. Otherwise you're just feeding the noise

gmbpro

I've seen the same pattern across dozens of GA4 audits. Content teams optimise for rankings, impressions, clicks, time on page - every metric that looks great in a dashboard but tells you nothing about whether the user's problem got solved.

You end up with pages that "perform well" analytically while the person who clicked still leaves frustrated. It's a measurement inheritance problem, and fixing it usually means rethinking the event model: what counts as a "success" when the real goal is resolution, not just engagement.

What questions do you use to validate whether a page actually resolves the query?

maya-b2b

Question five really gets to the heart of it - are we writing a page or writing an answer? Deep down, most of us already knew which one we'd been doing

isa-ecom

Right. And that's the bit that gets glossed over in DTC, too. We're so obsessed with AOV and cart recovery that we forget the actual customer journey - the unboxing, the follow-up SMS, the loyalty loop - all hinges on whether those pages actually answered a damn question. A page that's just a pretty product shot with a "buy now" button isn't resolving anything, it's hoping the customer already knows they want it. That's not marketing, that's a gamble.