The Brief That Should Have Raised a Flag
In Q3 2024, I approved a content brief that landed on my desk with eight target keywords. Four of them were clearly Aristocrat-related: aristocrat, aristocrat slot demo, aristocrat casino playing cards, and a couple of branded variants. The other four were—how do I put this—not. "Video game remake." "Table tennis near me." "How to play golf the card game."
I remember staring at that list and thinking: somebody exported a keyword tool report and didn't clean it. But the deadline was tight, the writer was already assigned, and I signed off. (I should add that I was the one who signed off. That part matters.)
Six weeks later, the page had climbed to position 11 for aristocrat slot demo. Impressions were up 240%. Conversions? Zero. Not a single qualified operator inquiry. And when I pulled the search query report, the traffic was mostly people looking for ping-pong tables in their zip code.
The Surface Problem Everyone Talks About
Ask any content team why a piece underperforms and you'll hear the same three answers: the writing wasn't good enough, the SEO was thin, or the topic was too competitive. From the outside, that looks like the whole story. The reality is almost always upstream of the draft.
Here's what most people assume: if a keyword is in the brief, it belongs in the article. That's the logic that gave us a paragraph about "how to play golf the card game" sitting awkwardly between two sections about Dragon Link mechanics. It read like a hostage situation. The writer knew it. I knew it. Nobody said anything because the brief was the brief.
The page technically answered all eight keywords. It also answered none of them well. Google's own guidance on helpful content is fairly blunt about this—content made primarily to attract search traffic, rather than to help a specific audience, tends not to perform. I'd read that guidance a dozen times. I still let a tool-generated keyword dump become an editorial plan.
The Deeper Problem: Keywords Without a Cluster
What most people don't realize is that a keyword list is not a content strategy. It's raw material. The strategy is deciding which cluster each keyword belongs to, and what job that cluster is supposed to do.
Look at that brief again. There were actually three different clusters jammed into one document:
- Brand and product queries—aristocrat, aristocrat slot demo, aristocrat casino playing cards. These are people who already know or suspect what Aristocrat is. They want specifications, demos, cabinet details, game mechanics.
- Adjacent entertainment queries—video game remake, how to play golf the card game. These might overlap with a casual gaming audience, but they carry almost no operator intent. A casino floor manager is not searching for golf card game rules on a Tuesday morning.
- Pure local intent—table tennis near me. There is no world in which this belongs on a B2B gaming equipment page. It's a separate business entirely.
Three clusters. Three different audiences. One page. That's the structural failure, and no amount of copyediting fixes it. The writer was set up to fail, and so was the page.
I want to be fair to whoever built that brief—keyword tools are designed to surface related terms, and they're good at it. But "related" in a semantic model is not the same as "related" in a purchasing journey. The tool doesn't know that table tennis near me and aristocrat casino playing cards are different businesses. A human has to make that call.
What It Actually Cost Us
I ran the numbers afterward, mostly to justify the process change I was about to push through. The page cost roughly $1,800 in writing and editing, plus about 12 hours of my review time across two rounds. That's not catastrophic in isolation. But it wasn't isolated.
The bigger cost was authority dilution. When a page covers five unrelated topics, search engines have a harder time classifying what the site is about. That page sat on a domain that should have been building topical authority around gaming hardware and casino management systems. Instead we published a page that was one-third local services directory. It took us roughly four months to recover the topical signals on that section of the site, mostly by pruning and rewriting adjacent pages.
And then there's the internal cost, which never shows up in a spreadsheet. My team started treating content briefs as paperwork to route around. Twice, writers asked me privately whether they really had to include every keyword. The brief had stopped being a tool and become a box to tick. That's a quality culture problem, and it starts at review—which, for the record, is my desk.
What We Changed (Short Version)
The fix was embarrassingly simple. Before any brief gets approved, I now require a one-page cluster map: every keyword assigned to one of three buckets—brand/product, adjacent audience, or "does not belong, kill it." If a keyword can't be defended in one sentence as serving the same reader as the primary keyword, it comes out. No exceptions, no "but it has volume."
We also added a second gate: a five-minute blind read of the outline. If a colleague can't guess who the page is for after reading the section headers, the brief goes back. That's it. Two checks, maybe 20 minutes per brief. Our rejection rate in Q1 2025 was 31%—which sounds high until you compare it to the 100% rework rate we had on pages like that Q3 one.
This worked for us because our content volume is moderate—roughly 40 briefs a quarter, with a small team and a single reviewer. If you're running a high-volume operation with automated publishing, the calibration is probably different. I can only speak to the mid-volume, human-reviewed setup. Your mileage may vary.
The lesson, if there is one: the quality problem usually isn't in the draft. It's in the brief nobody questioned. I questioned the draft for two weeks. I should have questioned the list on day one.
