跳至正文
技术 SEO 与抓取索引 · Google SEO

Bulk Redirect Audit: Check Old URLs, Every Hop and the Final Page

// / / 光算科技

A useful bulk redirect audit starts with the original URL list and an expected destination, then records every hop and reads the final page. A final 200 alone cannot tell you whether an old product URL reached the right product, a login screen or the homepage.

Use this pass after a URL change, migration or routing repair. It checks what happens to the known URLs now; it does not make an incomplete export into a complete history or decide which content should have been moved.

Freeze the input list and the intended mapping

Combine available pre-change crawl data, old sitemaps, CMS records, relevant logs and search/analytics exports. Record each source and date. Search Console performance page rows, page-indexing reports and link reports describe different things; none should be called a complete list of every old URL.

Keep the original absolute URL, including scheme, host, path, trailing slash and query. Remove exact duplicate records if appropriate, but do not silently lowercase paths, remove parameters or collapse slash variants before checking them. Those distinctions may be the defect.

Add the expected destination and the reason: the same article moved, an equivalent product replacement, an intentional temporary route, or content genuinely removed without an equivalent. A sitewide “all old URLs go home” rule is not a substitute for that decision.

Run the crawl with redirect following explicitly enabled

Screaming Frog’s redirect-audit tutorial describes a reproducible workflow:

  1. Switch the SEO Spider to List mode and upload the old URLs.
  2. Enable Always Follow Redirects in the spider configuration. The tutorial notes that ordinary list mode uses a crawl depth of zero; without the redirect-following option, the uploaded list alone is not a full chain audit.
  3. Start the crawl and let it finish. Record tool version, relevant settings, request conditions and any limits or failures.
  4. Export All Redirects. The documented report includes the original uploaded URLs, fixed start/final columns and the individual hops, rather than only URLs with long chains.

These steps follow the vendor’s published guide, not a claim that this article ran your site through that software. Check the corresponding controls in your installed version. Use modest request rates suited to the site, and investigate 429 or server failures instead of treating a fast incomplete crawl as success.

If using another checker, confirm that it resolves relative Location values, records each response, detects loops and limits, and keeps timeouts as failures. A HEAD-only check may differ from a normal GET; inspect representative actual responses and content. HTTP-only tools may also miss browser-side redirects.

Keep observations separate from the verdict

Report fieldWhy keep it
Original URL and expected destinationMakes the intended content relationship explicit
First status and each hop’s status/LocationSeparates a direct error from an error reached after redirects
Final URL, status and content typeShows where the request ended and what kind of response arrived
Content match, language and applicable product/versionDistinguishes the right page from any page that returns 200
Canonical and indexing controlsRecords further technical signals without claiming Google accepted them
Failure, request time and access conditionsPreserves loops, timeouts, access restrictions and reproducibility

Do not turn an untested content match into “passed.” A crawler’s indexability label is a technical assessment under its settings, not confirmation that Google has indexed the URL.

Two arrows labeled 301 connect Old URL to Intermediate to Final page, with two hops and a final 200 identified as still requiring a content check
Original fictional route illustration, not a crawler screenshot or client result. Two redirects are two hops; reaching 200 does not establish that the content matches.

Read a worked example without flattening the exceptions

The following rows are fictional audit examples. Paths are shortened for readability; a real export should retain complete URLs.

Original pathObserved routeInterpretation
/old-a/301 → /model-a/ (200)One redirect; confirm model A and the intended final canonical
/old-b/301 → /middle-b/ (301) → /model-b/ (200)Two redirects; inspect each mapping before considering a direct route
/temporary/302 → /notice/ (200)Check whether the temporary arrangement is intended
/removed/410, no redirectMay be appropriate for genuine removal without an equivalent
/old-c/301 → / (200)Content mismatch if the expected model C is not represented by the homepage
/loop-a/301 → /loop-b/ (301) → /loop-a/Repeated URL indicates a loop; there is no reached final page
/blocked/403, no redirectAccess denied in this request; not proof of permanent deletion

Google’s redirect guidance treats permanent and temporary redirects differently. Do not convert an intentional temporary route simply to make a report greener. Likewise, multiple hops warrant inspection but do not prove a Google penalty or that the request failed.

Google’s HTTP-status documentation distinguishes successful responses, client errors and server/overload failures. A deliberate 404/410 for removed content is a different case from an important live article returning 404 accidentally. Preserve that distinction when counting defects.

Read the destination and its signals together

Open the destination, check the actual answer or product, and try the next useful step. Confirm the correct language and version. A generic page carrying an error message can return 200; a status-only audit will miss the problem.

Compare HTML and HTTP canonical declarations with the intended destination, internal links and sitemap entries. If the final page points elsewhere, investigate the underlying relationship rather than assuming every multi-URL arrangement is wrong. The canonical-difference diagnosis explains how to compare those signals with Google’s stored choice.

For localized URLs, also check the hreflang target and return relationships. On WordPress, the old-slug and permalink guide (Chinese) explains why some old URLs may already redirect while others need explicit work.

Re-run the same known set after the repair

Fix the responsible routing layer, update controlled links where appropriate, and run the saved input again under recorded conditions. Compare individual rows, not just the total number of 200 responses. Confirm that an apparent improvement did not send more old products to an unrelated page.

Keep both reports, the mapping decisions and unresolved requests. Recheck relevant public-cache behavior and important browser paths. Later URL Inspection observations can establish what Google has recorded at that time; the live audit alone cannot establish indexing or rankings.

For a larger technical and content programme, Guangsuan’s SEO service can work from this bounded input list and the actual findings. State what was covered and what is still unknown before calling a migration verified.