Why isn't react-helmet working for SEO?

It is working — in the browser. Head-management libraries update the document after JavaScript runs, which is after the crawler has captured the server's response. Unless your setup renders those tags on the server as part of the initial HTML, crawlers and social scrapers capture the page from before the library executed. The tags must be in view source to count.

Why do my links unfurl without a preview on WhatsApp or Slack?

Social and chat scrapers fetch the raw HTML and execute no JavaScript, so Open Graph tags set client-side are invisible to them permanently — there is no later rendering pass. The tags must be server-rendered, and after fixing them you need to refresh each platform's cache through its debugger tool, because platforms store the previous scrape.

Does the meta description affect rankings?

No — it is not a ranking factor, and Google replaces a majority of descriptions with query-specific snippets from the page anyway. Its job is click-through: a good description is an ad for the result. Write it for humans, keep it accurate, and don't spend engineering effort forcing Google to display it verbatim, because no mechanism guarantees that.

Is setting document.title with JavaScript enough for Google?

It is unreliable. Google may pick up a JavaScript-set title in its rendering pass, but the first-pass capture, the version many systems keep, contains whatever the server sent — often a placeholder. Social scrapers never see it at all. A title that matters belongs in the server response, where every consumer of the page receives it unconditionally.

I fixed my tags weeks ago and Google still shows the old title. Why?

Two possibilities. The page simply hasn't been recrawled — use URL Inspection to prompt it and confirm the new tags appear in the crawled version. Or it has been recrawled and Google is rewriting your new title too, which means the title itself still reads as a poor match. Check it against the page's H1 and shorten it before assuming a technical failure.

HomeFixMeta tags not showing

Google isn't showing your meta tags. There are two ways that happens.

Short answer

Two different problems produce this symptom. If your tags are set by JavaScript, crawlers capture the HTML before the tags exist — a rendering problem, fixed by moving metadata to the server. If the tags are in the HTML and Google displays something else, Google is rewriting them — a content problem, fixed by writing better tags.

These two failures look identical in a search result and need opposite fixes. Teams routinely spend weeks on the wrong one — rewriting titles that were never reaching the HTML, or re-architecting rendering to fix a title Google was always going to rewrite. One check tells them apart.

The fork: is the tag in the raw HTML?

In the raw HTML?	Diagnosis	Fix lives in
No — tags appear only after JavaScript	Problem A: rendering. Client-side head management is applying tags after capture.	Your codebase
Yes — tags present, Google shows different text	Problem B: rewriting. Google judged your tag a poor match and substituted its own.	Your copywriting
Yes in raw, but different values after rendering	A metadata mismatch — two versions of your tags disagree. Related but distinct.	The hydration mismatch page

Problem A: the tags never reach the crawler

Client-side head management — react-helmet and its descendants, document.titleassignments, metadata effects in client components — updates the DOM after JavaScript runs. It works perfectly in the browser tab, which is why it survives code review. But the crawler's first-pass capture happened when the server responded, before any of that executed. The tag exists everywhere except the one artifact that gets indexed.

The tell: Google shows a generic title, your homepage's title on inner pages, or the placeholder from your index.html— because that's what the first response actually contained.

Problem B: the tags arrive and Google overrules them

Google rewrites a substantial share of titles and most descriptions — by design, not as an error state. Titles get rewritten when they're too long, keyword-stuffed, boilerplate across many pages, or when they disagree with the page's visible H1. Descriptions get replaced with whatever page snippet better matches the specific query. If your tags are in the raw HTML, no rendering work of any kind will change this — the lever is the tag itself: concise, specific, consistent with the on-page heading.

Social scrapers execute no JavaScript. Ever.

Google eventually renders JavaScript; the scrapers behind link previews on Facebook, WhatsApp, Slack, LinkedIn and X do not. There's no delayed second pass and no "eventually." Client-set OG tags don't work late — they never work, which is why a site can look fine in Google and still unfurl as a bare URL in every chat app. And because platforms cache scrapes, fixing the tags isn't enough: re-run each platform's card debugger to evict the cached tagless version.

Confirm which problem you have

The check below fetches your page both ways and compares the metadata fields directly — title, description, canonical — in the raw response versus after rendering. That comparison is the fork.

Check which tags are in your raw HTML

https://

Reading the result

On this page, read the metadata fields, not the verdict color:

Metadata pattern	What it tells you
MISSING Title/description empty or generic in raw, correct after render	Problem A confirmed. Your metadata is client-applied. Steps two through four are your path — and note the social-card consequence applies in full.
SERVED All tags identical in raw and rendered	Rendering is clean. If Google still shows different text, you have Problem B — rewriting — and the fix is editorial. Compare your title against the page's H1 first; disagreement between them is the most common trigger.
MISMATCH Tags present in both but with different values	Your server and client disagree about your metadata — a mismatch, and the worst pattern here, because you don't control which version gets used. Fix determinism before anything editorial.

Image slot

Screenshot: the metadata panel with title missing in raw, present in rendered — Problem A's signature.

Problem A in one screenshot: the title exists — just not in the version that gets captured.

How to fix it

Six steps. Step one is the fork; A-path runs through steps two to four, B-path is step five.

Check whether the tags exist in the raw HTML

Everything branches on this. Fetch the page without JavaScript and look for your title, description, canonical and OG tags in the response. Present means Google is choosing to display something else. Absent means they're being applied client-side, after the capture.

If absent: find what's setting them client-side

The suspects are client-side head libraries, document.title assignments, and metadata effects inside client components. All of them run after the server response is sent — which is after every crawler and every social scraper has already captured the page.

Move metadata into the server response

Use your framework's server-side metadata path: the metadata export or generateMetadata in Next.js's App Router, server-fetched props feeding next/head in the Pages Router, useHead in Nuxt, or your SSR layer's equivalent. The test is brutal and simple: the tags must be in view source.

Treat social cards as the strict case

Scrapers for Facebook, WhatsApp, Slack, LinkedIn and X execute no JavaScript at all — client-set OG tags don't work late, they never work. Once the tags are server-rendered, run the platforms' card debuggers to purge their caches of the old, tagless version.

If present: fix the tag, not the rendering

Google rewrites titles it judges poor matches — too long, stuffed, boilerplate, or disagreeing with the page's visible H1. Rewrites are fixed by writing a concise, accurate title that matches the page's content and heading, not by touching your rendering stack.

Verify and request recrawl

Confirm the raw HTML carries the final tags, then use URL Inspection on the affected pages. Displayed titles update on recrawl — give it days, not minutes, and don't chase it with rapid successive edits, which just create more versions to reconcile.

How long updates take

Surface	Update speed
Social card debuggers	Immediate — the debugger re-scrape is the update. Run it per platform; their caches are independent.
Google, inspected URLs	Days. URL Inspection prompts a recrawl; the displayed title changes when the new capture is processed.
Google, site-wide	Weeks, on the normal recrawl cycle. Frequently crawled pages update first; deep pages last.
Rewritten titles (Problem B)	On recrawl after your edit — if the new title is actually better matched. A rewrite persisting after a recrawl means Google still prefers its version; revise again rather than waiting harder.

Common questions

Either your title never reaches the raw HTML because it is set by JavaScript, or it reaches the HTML and Google rewrites it. Google rewrites titles it judges too long, keyword-heavy, boilerplate, or inconsistent with the page's visible heading. Checking whether your title exists in the raw response tells you which case you are in, and they have opposite fixes.