The "Cached" link is gone from Google results, and the cache: operator has not worked since September 2024. A footer link can no longer be proven by opening a copy that Google keeps. Three sources now share that job: URL Inspection in Search Console shows the indexed version and the rendered HTML, an archive record fixes the page at a past moment, and your own browser capture documents what a visitor sees today. For agencies, footer-link buyers and SEO specialists who built their routine around the cache, the proof is all three, saved together with a date and a URL.
Why Google retired the cache link
What the reports say, in order
Search Engine Roundtable, in its report dated January 25, 2024 (Opens in a new tab), said Google had removed the cache link from search results. A few days later, Google's Search Liaison, Danny Sullivan, confirmed the removal in a post on X. His reasoning was that the feature was meant to help people reach pages back when you could not depend on a page loading, that things have greatly improved, and that Google decided to retire it.
That reasoning is about page availability. It does not mention rankings, link evaluation or SEO checking. Using the cache link as an evidence tool was a habit that sat outside the purpose Google gave for it.
What happened to the cache: operator
In the same post, Sullivan wrote that the cache: operator would also go away in the near future. Search Engine Roundtable's report of September 24, 2024 (Opens in a new tab) says Google confirmed that the cache: search operator no longer works in Google Search, and that Google posted a note saying so. The same report says Google had added links to previous versions of webpages from the Internet Archive a couple of weeks earlier.
If a guide written after that date still tells you to type cache: before a domain, treat it as outdated. To see today's behaviour, run the query yourself and keep the result with its date.
The archive links in results
Sullivan had said he hoped Google would add Internet Archive links where the cache link used to be, within "About This Result", and added that there were no promises. The September 2024 report says those links were added. Interface labels and Google's pages can change, so check the current screen before relying on a label. Even so, do not build your evidence on the link in the result. Open the archive directly, using the address you are checking.
Four gaps the cache copy left open
Practical suggestion: judge the old habit by what it left behind, because the replacement has to cover the same gaps.
The copy hid the layout around the link
Evidence handed to a client or a manager should also show that a reader could find the link. A link inside a collapsed menu, pushed off screen, or added by a script under certain conditions looks the same in copied text.
The copy was stale by design
The URL Inspection help page says the indexed view is not a live test and shows the latest indexed version. A cache copy had the same nature: it was a moment in the past. If the publisher removed the link, the copy could still show it. If the publisher added it recently, the copy might not show it yet. Both errors mislead.
The copy could not be handed to anyone else
"I checked, the link was there" cannot serve as proof between an agency and a client or between a client and a publisher. The method has changed, but the principle has not: proof is a record that anyone else can open.
The copy was not your file
Proof has value when it can be opened later by someone else. Archive services place the capture time inside the address, and that address can be shared. Looking at a cache copy and saying "I saw it" left no durable file unless you took a screenshot yourself.
Alternative 1: URL Inspection and rendered HTML
Property access comes first
According to Google's URL Inspection help page (Opens in a new tab), the URL must be in the property you have open, and each property has a daily limit. If the page holding the footer link sits on a site you do not manage, the tool is not directly open to you. Practical suggestion: ask the publisher to share the inspection screen for that URL, understanding that this depends on the publisher's agreement. If the property is yours, for example your own network of sites, the tool is the strongest link in the chain.
The receiving side is a different place to look. To see links pointing to your own site in Search Console, read the Search Console links report guide. A report and a per-page capture are different jobs and do not replace each other.
Indexed view and live test are separate lines in your log
The help page says "View crawled page" shows the raw HTML and the HTTP headers. It is available only for URLs whose status is "URL is on Google" or "URL is on Google, but has issues". The last crawl time tells you which moment the indexed version belongs to, and it is one of the dates that belong in your log.
"Test live URL" fetches the address in real time. Per the help page, a screenshot appears only in a successful live test, under "View tested page", and none is available for the indexed URL. A passing live test is not a guarantee that the page will appear in Search. Record the two views on separate lines and state which one a given capture came from.
For how Google recrawls a page after an update and how to use the tool for that, see the guide on recrawls, URL Inspection and sitemaps. If the page was not indexed, the cause is its own investigation, and diagnosing pages that are not indexed walks through it.
When a script inserts the footer link
Google's crawlable links page (Opens in a new tab) says links are also crawlable when JavaScript inserts them dynamically, as long as the markup is an a element with an href attribute. It also says that if you use JavaScript to insert anchor text, you should use the URL Inspection tool to make sure the text is present in the rendered HTML. So when a script builds the footer link, seeing it in the page source is not enough, and seeing it in your browser is not enough either. The rendered HTML has to contain an a element with an href.
Alternative 2: archive records and a browser capture
Reading a Wayback Machine capture
The Internet Archive help page (Opens in a new tab) explains that the number in the middle of a capture address reads as yyyymmddhhmmss. Its example, 20000229123340, means a capture made on Feb 29, 2000, at 12:33:40. On the calendar, blue dots mark 2nn status codes, green 3nn redirects, orange 4nn client errors and red 5nn server errors. If you need proof that the page opened, a blue capture is the record to cite.
The "Save Page Now" feature saves one specific page one time. The help page says it does not add the URL to future crawls and does not save multiple pages, directories or entire sites. The page also says new captures take about 3 to 10 hours to appear. For link evidence, that means you can start the capture yourself on the day the link goes live.
What an archive cannot show
The help page says some sites are missing because they were password protected, blocked by robots.txt or otherwise inaccessible to the automated systems, and that site owners can ask for exclusion. JavaScript that needs the original server, server-side image maps and orphan pages with no inbound links are hard to capture, while simple HTML is the easiest to archive. For broken or gray images in a capture, the page points to images that were not archived.
For footer evidence, the consequence is this: if the capture shows the link as text with an href, you hold a dated record. If the archive has nothing, that does not prove the link was absent.
Capturing the page in your own browser
Practical suggestion: take three items. A full-page screenshot that includes the address bar, the destination address shown in the status bar when the pointer rests on the link, and a saved HTML file of the page. Put the date in each file name and keep the three together. Reading rel and canonical values with developer tools is outside the scope of this topic.
Open the page in a window where you are not signed in, so the footer appears as any visitor sees it. If a cookie banner covers the footer, keep one frame with the banner and one with it closed. Record the browser and the screen width in the log notes, and keep desktop and mobile captures as separate files.
Which evidence set is enough for a footer link
Decision rule
The recommended order is URL Inspection first, archive second, browser capture as the complement. The reason: URL Inspection gives the raw HTML, headers and rendered version from Google's own view, which is the closest signal of whether the link can be crawled. The archive is a dated record from a party independent of you. The browser capture adds what a human eye sees.
Decision rule, built as a ladder you climb as far as access allows:
- On the go-live day, take the browser capture. Anyone can do it, so it is the floor.
- Start Save Page Now for the same address and log the capture address once it appears.
- Ask for the URL Inspection result if the publisher manages the property, and log the last crawl time.
- If a script adds the link, make the rendered HTML check the deciding record.
- If a rung is not available, write in the report which rung it was and what that leaves unchecked.
No record shows that the link will stay live. Each one says something only about the moment it was taken.
Example scenario: a glassblowing studio and a makers' directory
Example scenario: a glassblowing studio buys a sitewide footer link on a regional makers' directory, with its studio name as the anchor text. On the day the link goes live, its agency opens the directory's home page and one article page in a signed-out window, takes screenshots, saves the HTML and starts Save Page Now for both addresses. Once the captures appear, the agency enters their addresses with their timestamps in the log. The directory agrees to share its URL Inspection result for the two pages, and the agency adds the last crawl time and a note that the link was visible in the crawled HTML.
Months later the studio's manager asks whether the link is still there. The agency repeats the same three steps in the same order and compares them with the first row. Instead of a statement from memory, the manager gets two dated rows and the differences between them.
Evidence log fields
Log columns, in the order a reader needs them:
- Capture number, date, time and time zone.
- Publisher page address, plus the final address if it redirects.
- Href value and anchor text or image alt text.
- Which Google view was used, and the last crawl time.
- Archive capture address with its yyyymmddhhmmss stamp.
- Screenshot and saved HTML file names, with the screen width.
To document whether the page itself is in Google at all, see how to check that a backlink page is indexed. For a step-by-step footer-specific routine, read how to check footer backlinks. How a page appears in results text is a separate subject, handled in controlling search snippets with nosnippet and max-snippet.
A crawlable-link test for the capture
Run this test on each capture. The order starts with what can be read from the markup and ends with what a script and the server add.
- If the link is an image, does the img element carry alt? Google uses the alt attribute as the anchor text for images used as links.
- Is the anchor text placed between the a tags? The page describes good anchor text as descriptive, reasonably concise and relevant to the page it is on.
- Is the link an a element with an href attribute? The page says Google cannot reliably extract URLs from a elements without href.
- Does the href resolve to an actual web address? The page lists a javascript: address among the formats it does not recommend.
- Does the link depend on onclick alone? The page lists an a element with only onclick as a format it does not recommend.
- If a script adds the link, is it present in the rendered HTML? URL Inspection is the tool the page names for this.
- Practical suggestion: did the page return a successful status on the capture day? Read it from the headers in URL Inspection or the blue dot in the archive calendar.
Reading and judging rel values is outside the scope of this topic. Write the value you see into the log as it stands.
Reading the record when the link seems to be missing
What each situation lets you write
| Situation | Evidence in hand | Wording for the report |
|---|---|---|
| Blue dot, link visible in the capture | Dated archive capture | The link was present at the capture time stamped in the archive address |
| Blue dot, link missing from the capture | Archive capture plus rendered HTML | The link was absent from the capture; a script may add it later, so the rendered view was checked |
| Green dot (3nn) | Archive calendar | The address redirected on that date; the destination page was checked for the link |
| Orange or red dot (4nn, 5nn) | Archive calendar | The page returned an error on that date; a second capture on another date is pending |
| No capture at all | Browser capture only | No archive record exists; the browser capture is the only dated record |
None of these rows establishes that the publisher removed the link on purpose. A template refresh, an address change or a server problem can produce the same observation. When writing to the publisher, send the logged observation instead of a conclusion.
Writing to the publisher
Send the two dated records and the address, and ask one question: what changed on that page between those dates? A message built this way moves the exchange from "was it there" to "when did it change". The publisher can answer from its own edit history without being accused of anything.
Keeping the records usable
File names and storage
Practical suggestion: build each file name from the capture date, the publisher's domain and the kind of record, for example the date first so that folders sort in time order, then the domain, then words such as screenshot, html or archive. Keep one folder per link and one log row per capture session. A log that lives only in a chat thread is hard to hand over, so store it where a colleague or the client can open it without asking you.
Keep the saved HTML file untouched. If you need to annotate a screenshot, save the annotated copy as a separate file and leave the original as it was captured.
Three triggers for a new capture
- Go-live: the first row, taken the day the publisher says the link is in place.
- A complaint: when a client says the link has disappeared, take a fresh capture before you compare it with the first one.
- A renewal or a redesign: capture again before the contract period closes, and again at once if the publisher announces a redesign or a new domain, because a template change can alter the link on every page that uses the template.
Continuous monitoring for the days between those moments is a separate job. The log documents specific moments only.
Mistakes to avoid
Saving the screenshot without the HTML
A screenshot shows how the page looked and hides the href and rel values. Save the HTML file in the same session and give both files the same date prefix.
Capturing while signed in
A signed-in session can personalise the page. Use a window where you are not signed in, so the footer appears as any visitor sees it.
Relying on the archive for a script-added link
The Internet Archive help page says JavaScript that needs the original server is hard to capture. If a script builds the link, the archive record can look empty while the rendered HTML holds the link, so check both.
Reading one template as the whole site
A link seen on one page does not show that it sits in every template. Capture one page per template type, such as a home page, an article page and a list page, and log each separately.
Writing "Google sees the link" without naming the view
A live test shows what Google can fetch now. The indexed view shows the version last indexed. Name the view in the sentence, or the claim outruns the record.
Frequently Asked Questions
What if the publisher blocks archiving?
The Internet Archive help page says some sites are missing from the archive because they are blocked by robots.txt, password protected or otherwise inaccessible, and that site owners can ask for exclusion. Practical suggestion: rely on the browser capture and ask the publisher for the URL Inspection result, and write in the report that no archive record exists.
Which time zone should the log use?
Practical suggestion: write the local time zone of the person who captures next to every time you enter yourself, and copy the archive stamp exactly as it appears. Comparing both on one clock avoids arguments about which capture came first.
Was Google's cache link removed for good?
Search Engine Roundtable reported its removal from results in January 2024, and Danny Sullivan confirmed it in a post on X in early February 2024. A September 2024 report says the cache: operator no longer works either. For anything after those reports, run your own search and keep the result with its date.
Is a Wayback Machine capture enough on its own?
It gives an independent record with a visible date. The Internet Archive help page, though, says that sites blocked by robots.txt or protected by passwords may be missing and that JavaScript needing the original server is hard to capture. If a script adds the link, the archive alone may fall short, and the rendered HTML is needed as well.
Should a sitewide link get one log row per page or per template?
Practical suggestion: one row per template type, with the sampled page address in the row. A link seen on one template does not show that it sits in the others.
Does a passing live test mean the link counts in Google?
No. The help page says a live result is not a guarantee of appearing in Search. The test shows that the address could be fetched in real time and what Google saw in it.
Does the cache: operator still return anything?
Per Search Engine Roundtable's September 24, 2024 report, Google posted that the operator no longer works. A later check is yours to make, and the result belongs in the log with a date.
Conclusion
Losing the cache link did not remove the need to prove that a link is live. It spread the proof over several sources. URL Inspection gives the indexed version and the rendered HTML, the archive gives a record at a specific moment, and a browser capture documents what you saw. For a footer link, the strongest set is three dated records kept on one log row. Repeat the same steps in the same order each period so that any change between two dates can be seen.
If you are planning a footer or sitewide link, you can review the footer link product page. Before buying, ask the publisher for a page where captures can be taken and for the format of the report you will receive.
Post a comment
Please log in to leave a comment. Login