A Short History of the Meta Description: From 1990s Keyword Tag to a CTR Signal Google Rewrites

The meta description might be the most misunderstood tag in SEO. People still ask us whether stuffing keywords into it will lift their rankings, whether Google "requires" one, and why the text they carefully wrote keeps getting replaced with a sentence pulled from the middle of the page. Every one of those questions has a tidy answer, and every one of them makes more sense once you know where the tag came from.

So this is a history, not a how-to. If you want the practical rules — the character counts, the pixel widths, the truncation behaviour — we cover those separately. Here we're interested in the journey: how a single line of HTML went from telling search engines what a page was about, to earning clicks, to being quietly rewritten by the very engine it was meant to inform.

Born in the 1990s as a hint to the crawler

In the mid-1990s, search engines had a hard problem: the web was growing fast, full-text indexing was expensive, and the machines doing the crawling were not clever about understanding a page. So the engines of the day — AltaVista, Infoseek, Lycos and their peers — leaned on the page to describe itself. That is what <meta name="description"> and its sibling <meta name="keywords"> were for. You told the crawler, in your own words, what the page was about, and the engine took you more or less at your word.

For a while this worked, because most webmasters were honest and the alternative — parsing and ranking the full text of every page — was genuinely difficult. The meta description was a shortcut that helped both sides. It was a hint, and the engines treated it as one.

The abuse era: when descriptions became a keyword dump

A signal that you control, that directly influences ranking, and that nobody verifies is an open invitation. It did not take long. Through the late 1990s the meta description and the meta keywords tag filled up with everything a page owner wished it ranked for: long lists of terms, competitors' brand names, misspellings, cities the business had never operated in. The tag stopped describing the page and started describing the owner's ambitions.

The engines noticed their index quality dropping. If the description could say anything, it told you nothing reliable. The response was inevitable: search engines stopped trusting self-reported metadata as a ranking input and shifted toward signals that were harder to fake — the actual words on the page, and later, the links pointing at it. The meta keywords tag was abandoned almost entirely. The meta description survived, but its job was about to change.

2009: Google says the quiet part out loud

For years the industry argued about whether the meta description still carried ranking weight. Google settled it publicly in 2009, stating plainly that neither the meta description nor the meta keywords tag factors into its ranking algorithms for web search. Not a tie-breaker, not a small signal — no weight at all.

That announcement is still the single most important fact about the tag, and it is still the one people forget. Writing a keyword-rich meta description does not help you rank. It never will, because ranking is decided by other things entirely. What the 2009 clarification did was free the tag to do the job it is actually good at.

From ranking factor to click-through tool

If the meta description does not affect ranking, why write one at all? Because it is often the sentence a searcher reads before deciding whether to click. On a search results page, your listing is usually a title, a URL, and a line or two of description. The title earns the scan; the description closes the deal. A page can rank in position three and still lose the click to position five because position five wrote a description that answered the searcher's question and yours read like a database field.

This is the tag's modern role, and it is a genuinely valuable one. It has nothing to do with rankings and everything to do with click-through rate — the share of people who see your result and actually click it. Improving a description cannot move you up the page, but it can win you more of the traffic you already earned by being on it. That is some of the cheapest work in SEO.

The "155–160 characters" rule of thumb

The familiar advice to keep descriptions to around 155–160 characters comes from this era — it is a proxy for the amount of text Google will show before it truncates the line with an ellipsis. It is a useful rule and a rough one, because the real limit is pixel width, not character count: a line of wide characters gets cut sooner than a line of narrow ones, and mobile and desktop crop at different points. We go into the practical mechanics — where the cut falls, how to front-load the important words, and how to write for the query — in our guide to title and meta description length. For the history, the point is simply that the length conventions exist to serve clicks, not rankings.

The rewrite era: when Google writes it for you

Here is where the tag's story gets genuinely strange. Having spent a decade telling everyone the meta description is under their control as a click-through tool, Google increasingly ignores what you wrote and generates its own snippet from the page instead. If your description does not match the searcher's specific query, Google will often pull a more relevant sentence straight from your body copy, because a snippet tailored to the exact search tends to earn more clicks than a generic one.

Studies that sample search results have repeatedly found that a large share of the descriptions shown are not the ones the page authors wrote — frequently the majority of them. That sounds like it makes the tag pointless, but it does not. A well-written description is the default Google shows when your page ranks for the query you actually optimised for; the rewrites tend to happen on the long-tail, query-specific searches you could never have anticipated. Writing a good one gives Google a strong option to use, and a fallback it is happy to keep. Leaving it blank just hands the decision to the machine every time.

The AI twist: descriptions in a generative-search world

The latest chapter is still being written. AI Overviews and generative answer engines do not read your meta description and repeat it — they synthesise an answer from the content they can extract from the page, and they cite sources based on what that content actually says. In that world the meta description matters even less as a direct input than it did in 2009, and clear, factual, extractable body content matters more.

That is the same lesson the rewrite era taught, only louder: the words in your metadata are a hint, and the words on your page are the truth. If you want to be understood by both classic search and AI-driven search, the work is in the content, not the tag. We wrote a whole guide on making your pages legible to answer engines — see generative engine optimisation — and none of its advice is about the meta description, which tells you something.

What to do with all this today

The practical takeaway is smaller than the history. Write a real, human meta description for every important page — one specific sentence that reads well and matches the intent of the query you want to win. Do not stuff it with keywords; it will not help you rank and it reads badly to the human who decides whether to click. Do not agonise over it either, because Google may rewrite it anyway. And do not leave your key pages with no description at all, because that is the one choice that is always worse than a mediocre one.

The only hard part is knowing which pages are missing a description, which ones have duplicates copied across a template, and which run past the point where Google will truncate them. That is a job for a crawl rather than a spot check — you can audit the meta descriptions across your whole site in one pass and see exactly which pages need a human sentence and which are quietly handing the decision to Google.

Key takeaways

• The meta description began in the 1990s as a self-reported hint to early crawlers, alongside the now-dead meta keywords tag.
• Keyword stuffing destroyed its trustworthiness, and search engines moved ranking weight to signals that are harder to fake.
• Google confirmed in 2009 that the meta description is not a ranking factor — that has not changed.
• Its real modern job is click-through rate: the sentence that wins the click on a result you already earned.
• Google now rewrites a large share of descriptions to match the specific query, so a good one is a strong default rather than a guarantee.
• In AI and generative search the tag matters even less; clear, extractable page content is what gets understood and cited.