What does `<meta name="description">` actually do for a page in search results, and is it a ranking signal?

## What the tag is ```html <meta name="description" content="A step-by-step guide to configuring TLS on nginx, with a working config and the three settings people get wrong."> ``` A `name`/`content` meta pair that declares a page-level summary. It is markup with no effect on rendering, no effect on the page's behaviour, and no effect on how the page is indexed. Its only consumer is software that wants a one-line description of the document. ## What it does and does not do for search **It is not a ranking factor.** This is the point interviewers are usually probing. Search engines rank on the page's actual content, links and a long list of other signals; the description tag is an author-supplied claim about the page and is therefore trivially gameable, which is exactly why it was removed from ranking consideration long ago. Stuffing keywords into it moves nothing. **It is a snippet candidate.** The two lines under the blue link in a result listing are the snippet, and the description is one input to it. The engine picks whatever it thinks best answers the query, so when the search terms appear in the body of your page, the snippet will often be assembled from that body text instead — with the query terms bolded — rather than from your description. This is normal and not a defect to fix. Expect your description to be used most often for navigational or brand queries where there is no specific passage to lift. **It affects click-through, not position.** Two results ranked identically can perform very differently depending on whether the snippet reads like a promise or like a template. That is the whole business case for writing one. ## Writing a good one - **Be specific to the page.** "Acme — quality solutions for business" on 4,000 URLs is worse than nothing, because it guarantees no result is distinguishable. - **Keep it around 150 characters.** There is no hard limit in the spec, but result listings truncate at roughly that length on desktop and less on mobile, so front-load the substance and treat anything past it as expendable. - **Describe the payoff, not the site.** "Includes a working config and the three settings people get wrong" earns a click; "This page discusses TLS" does not. - **Do not duplicate the title.** The title is already on screen directly above; repeating it wastes the only two lines you get. - **Do not keyword-stuff.** It buys no ranking, and a description that reads like a list of terms reads worse to the human deciding whether to click. ## Where else it turns up Because it is the only standard page summary in plain HTML, other software falls back to it when nothing more specific is present — browser bookmark descriptions, reader modes, feed and preview tooling. That is another argument for making it a real sentence rather than a slot filled by a template. ## The tag next to it that does nothing ```html <meta name="keywords" content="tls, nginx, https, ssl, security"> ``` Major search engines stopped using `keywords` for web ranking in 2009 and have said so publicly. It survives in templates because it is easy to add and nothing visibly breaks. Being able to say plainly that it is inert — and that you would delete it rather than maintain it — is a small but real signal in an interview, because it separates people who learned SEO from mechanisms from people who learned it from folklore. ## A generation habit worth naming On a large site, descriptions are usually generated rather than hand-written: derived from the first meaningful paragraph, from a product's short description field, or from structured content fields. That is fine as long as the generator produces something distinct per page and truncates on a word boundary. What fails an audit is the fallback path — the template default that fires when the source field is empty and quietly stamps the same sentence across a whole section of the site.