The meta keywords tag is a dead ranking signal. Google has publicly ignored it since 2009, Bing treats it as a spam indicator rather than a relevance cue, and no AI Overview or generative answer engine reads it the way it was designed to be read. If you’re still maintaining a comma-separated keyword list in your page head, you’re spending effort on a field that does nothing for rankings and, worse, hands your target terms to any competitor who views your source.
Meta Keywords
Meta keywords are an HTML <meta name="keywords"> tag listing comma-separated terms meant to describe a page’s topics — a deprecated signal that major search engines no longer use for ranking.
What the meta keywords tag actually is
The tag lives in your page’s <head> and looks like this:
<meta name="keywords" content="running shoes, trail running shoes, breathable trainers">It was introduced in the mid-1990s as a machine-readable hint: a quick, structured way to tell primitive crawlers what a page was about before search engines could reliably parse body content. At the time, early engines like AltaVista and Infoseek leaned on it heavily because their content-analysis was weak and a tidy list of terms was easier to trust than messy HTML.
That trust didn’t survive contact with spammers. Because the field was invisible to users and accepted unlimited terms, it became the single most abused tag on the web — stuffed with irrelevant high-traffic words, competitor brand names, and repeated phrases. The signal-to-noise ratio collapsed, and search engines learned to ignore it. By the time Google formally confirmed in 2009 that it does not use the meta keywords tag in web ranking, the tag had been functionally dead for years.
If a field is invisible to humans and free to fill with anything, it will get gamed into uselessness. The meta keywords tag is the textbook case — and the reason modern ranking leans on signals you can’t fake in your head.
How modern search engines treat it
Here’s the current, post-AI-Overviews reality across the engines that matter:
| Engine / system | Treats meta keywords as a ranking signal? | Notes |
|---|---|---|
| Google Search | No | Ignored for web ranking since 2009; reaffirmed multiple times |
| Google AI Overviews / AI Mode | No | Generative answers synthesize visible content and entities, not hidden tags |
| Bing | No (as relevance) | Bing has said it may use the tag as a spam signal — a flag, not a boost |
| Yandex | Largely no | Historically read it; now downweighted to near-irrelevance |
| Internal site search | Sometimes | Some CMS/search plugins still index the field |
| Niche / legacy directories | Rarely | A shrinking set of low-traffic engines may still parse it |
The pattern is consistent: every engine with meaningful traffic either ignores the tag or treats filling it aggressively as a red flag. AI Overviews and generative answer engines make this even starker — they extract meaning from the entities, structured data, and prose a user can actually see, so a hidden keyword list contributes nothing to whether you get cited. If you want to influence generative results, that work happens in your visible content and your structured data, not in a 1996-era meta tag.
The hidden cost: you’re publishing your keyword strategy
The case against meta keywords isn’t just “it doesn’t help.” There’s a real downside we flag in audits: the tag is visible in your page source to anyone who looks. A maintained, accurate keyword list is a free competitive-intelligence handout. Any rival can open view-source, read exactly which terms you’re targeting per page, and reverse-engineer your content map without spending a cent on tooling.
So the realistic options are: leave the field empty (no value, no leak), or fill it accurately (no ranking value, plus a leak). Neither beats simply removing it.
What actually drives rankings instead
If you’re tempted to “do something” with keywords at the markup level, redirect that energy to fields and signals search engines genuinely weigh. Here’s the priority order we work in:
- Title tag — still one of the strongest on-page relevance signals. Lead with the term, keep it under ~60 characters, make it unique per page.
- High-quality, intent-matched content — comprehensive prose that answers the query, built around your focus keyword and validated through real keyword research.
- Heading structure — one clear H1 tag plus a logical hierarchy of header tags that maps the page’s argument.
- Meta description — not a direct ranking factor, but it drives organic CTR, which is the click you’re actually competing for in the SERP.
- Semantic depth — cover the topic’s entities and relationships, the foundation of semantic SEO and how modern engines judge true relevance.
- Structured data — schema markup that earns SERP features and feeds entity understanding.
- Authority — relevant, earned backlinks remain a top-tier signal.
- E-E-A-T — demonstrated experience, expertise, authoritativeness, and trust, which Google leans on hardest for YMYL topics.
None of these can be faked in a hidden tag, which is exactly why they survived and meta keywords didn’t.
The 30-second cleanup
For most sites, the right move is a one-line decision:
- Using it for internal site search? Keep it, but treat it as an internal index field, not SEO.
- Not using it for anything? Strip it sitewide. It’s dead weight and a strategy leak.
Don’t agonize over it. There’s no penalty for having an empty or absent meta keywords tag — removing it simply closes a door that was never doing useful work. If you want a structured way to find pages still carrying legacy cruft like this, that’s the kind of thing a proper SEO content audit surfaces in bulk.
How this fits a real SEO workflow
We treat meta keywords the way we treat keyword density — a relic of an era when SEO was about gaming markup, not earning relevance. The genuinely useful keyword work happens upstream of the page: identifying intent, mapping terms to pages, and writing content that resolves the query better than the competition. That’s the engine behind programmatic SEO, where thousands of pages are built around real query data — and not one of them needs a meta keywords tag to rank.
If your team is still carrying habits from the keyword-stuffing era, the meta keywords tag is a good tell. Clear it out, then put the time into the signals that move 2026 rankings.
Frequently Asked Questions
Does Google use meta keywords for ranking in 2026?
No. Google publicly confirmed in 2009 that it does not use the meta keywords tag for web ranking, and that position has never changed. Its AI Overviews and AI Mode also ignore the tag, drawing instead on visible content, entities, and structured data when generating answers.
Should I remove the meta keywords tag from my website?
If nothing on your site reads it — like an internal search plugin — yes, remove it. The tag provides zero ranking value and exposes your target keywords to competitors via page source. There’s no SEO penalty for deleting it, so cleaning it sitewide is a low-risk housekeeping win.
Can meta keywords hurt my SEO?
Not through ranking algorithms directly, but they carry two real risks. Bing has indicated it may treat the tag as a spam signal if stuffed, and the field publicly leaks your keyword strategy to anyone viewing your source code. The safest posture is to leave it empty or remove it entirely.
What’s the difference between meta keywords and a focus keyword?
A meta keyword is a deprecated tag value search engines ignore. A focus keyword is your strategic working concept — the primary query a page is built to win. The first lives in dead markup; the second guides your title, headings, and content. Only the focus keyword matters for modern rankings.
Do AI Overviews or AI search engines read meta keywords?
No. Generative answer engines, including Google’s AI Overviews, synthesize responses from visible page content, entity relationships, and structured data. A hidden comma-separated keyword list contributes nothing to whether your page is cited or summarized. To influence AI results, invest in clear prose and schema, not the meta keywords tag.