A client asked me last month if I could "get the keyword density up to around 3%."
I hadn't had a client ask for something like this in quite a while. Keyword density stopped being a meaningful metric so long ago that Google's own people have publicly rolled their eyes at it, more than once. But it's still floating around out there. Write "SEO tips" fifteen times in an 800-word post. Bold the keyword every time it shows up. Cram the exact search phrase into the H1, the first sentence, and the meta description like it's 2011.
I get why it sticks around. SEO advice ages badly, and most business owners have better things to do than keep up with what actually changed. So I end up giving the same rundown a lot. Here's the version I wish I could hand to every client before we start.
What still actually works
Search intent beats keyword matching. Google got good at understanding what someone actually wants, not just matching the words they typed. If someone searches "best CRM for small teams," they want a comparison, not a page that says "CRM" forty times. Writing to satisfy the intent behind a search does more than any density trick ever did.
Structure makes or breaks scannability. Headers that actually describe what's in the section. Short paragraphs. Lists where a list makes sense, not because lists are trendy but because some information is genuinely easier to scan that way. A reader who can find the answer they came for in ten seconds is a reader who sticks around, and that behavior signal matters more than most people realize.
Freshness is underrated. Updating an old post with current information, current examples, and a new publish date often outperforms writing something brand new from scratch. I've watched a two-year-old article climb back up rankings after a thirty-minute refresh more times than I can count.
Internal linking done by someone who's actually read the other pages. Not an automated plugin suggesting links based on keyword overlap. A person who knows which article genuinely helps with which question, linking the two together in a way that makes sense for the reader first. Search engines pick up on that structure, but it works because it's genuinely useful, not the other way around.
None of this is a secret. It's just less exciting than a hack, so it gets buried under a decade of advice that promised a shortcut.
The AI search wrinkle
Here's what's actually changed things for me lately, and it's not the four points above. It's how many people are getting their answers from AI tools instead of a traditional search results page at all.
A survey from Clear Spark Digital found that 76% of Americans now use AI tools to find information at least weekly, and 48% do it daily. That's not a niche behavior anymore. It's close to the default.
What that means for content writers is a little counterintuitive. The instinct is to start "writing for AI," whatever that's supposed to mean in practice. What I've found actually works is closer to the opposite. AI tools pull from content that's clear, well-sourced, and structured around a real answer to a real question, which is the same content that's always ranked well with human readers. Chasing some separate AI-specific formula tends to produce worse writing for everyone, including the AI.
The clients who ask me to optimize for AI search specifically are usually surprised when my answer is "keep doing the stuff that makes the writing good." It's not a satisfying answer if you're hoping for a new lever to pull. But it's the honest one.
Where I push back
I'll admit I'm skeptical of a lot of the "write for AI" content flooding my feed right now. Some of it is useful. A lot of it is repackaged SEO advice from five years ago with "AEO" stapled onto the front, sold as something new. If a tactic wouldn't have helped a human reader in 2020, adding "AI" to the description doesn't make it work now.
What I tell clients instead is simpler and less marketable: write something genuinely useful, structure it so people can find what they need fast, keep it current, and link it up with the rest of your site in a way that actually makes sense. That advice hasn't needed an update in years, even while everything around it keeps changing.
What's the outdated SEO advice you still hear the most? I have a running list at this point, and I'd love to compare notes.
Top comments (1)
Great point about search intent beating keyword matching — that shift is exactly why we're now seeing a second transition from traditional SEO to GEO (Generative Engine Optimization). AI answer engines like ChatGPT, Perplexity, and Google's AI Overviews are becoming referral sources in their own right, and they cite pages that answer specific questions with sourced evidence, not keyword-stuffed content.
For anyone wanting to go deeper on this, I wrote a breakdown of GEO vs SEO strategies: dev.to/alexzhangai/geo-vs-seo-how-to-rank-in-ai-answer-engines-in-2026-eoo
Also agree on the free tools mention — you don't need expensive SEO suites to get started. I put together a zero-budget toolkit covering 17 free tools across keyword research, technical SEO, and rank tracking: dev.to/alexzhangai/free-seo-tools-2026-the-complete-toolkit-for-zero-budget-search-optimization-1kmg