The Wayback Machine is the Internet Archive's collection of dated copies of web pages. When a backlink or a promotional article you paid for goes live, saving the page in the archive leaves you a timestamped picture in case the page later changes or disappears. Save Page Now saves one page, one time. The short answer for buyers, agencies and publishers: an archive record supports a proof of publication but cannot stand as the proof by itself, and its value depends on knowing what it shows and what it leaves out.
What the Wayback Machine does and does not do
The archive stores how a page looked at a given moment and lets anyone reopen that version through a dated address. It does not report on the present. A record from March says nothing about April, and a record of a page says nothing about whether the publisher still gives the link.
What an archive record is
The help page Using the Wayback Machine (Opens in a new tab) describes the basic use: enter a domain or an address, press "Browse History", and the captures for that address open. The page also says the archive does not promise completeness. If a linked page is missing, the Wayback Machine picks the closest available date, or checks the live web when the link was not archived. A page you view in the archive can therefore be stitched together from more than one capture date.
What it is not
The same page states that the terms of use do not cover backups for the general public. It adds that it cannot guarantee a given site was or will be archived, and that the archive no longer offers a service to pack up lost sites. For a buyer, the consequence is direct: the archive is a second copy of evidence, not the only copy.
| Question you may face | What the archive can answer | What it cannot answer |
|---|---|---|
| Did the page exist on the date in the report? | Yes, if a good capture carries that date | Whether the page was changed hours later |
| Was the link inside the text? | Yes, in the capture that was saved | Whether the link survived the next edit |
| Is the page live today? | No | Requires a fresh live check |
| Was the whole site saved? | Not by Save Page Now, which saves one page | Pages nobody submitted |
| Can the archive restore my files? | No, it is not offered as a backup | Your own copy is the only restore source |
When a record is worth taking
Practical suggestion: take a record when a changed or deleted page would cost something. That covers a delivery check on a purchased package, an agency handing proof to a client, and a publisher who wants to document a publication date. A short-lived notice whose link was not meant to last does not justify the effort.
What happens after a page moves or vanishes is a separate decision, and if the publishing page moves or is deleted covers the choices. The archive record is only the first piece of that file.
Anatomy of an archive address
A saved page gets an address that looks like a normal web address with a long number in the middle. That number is the part that carries the evidence, so it is worth reading correctly.
The fourteen-digit stamp
According to the help page, the number in an archived address is a timestamp in the form yyyymmddhhmmss. The example given is 20000229123340, which means February 29, 2000, at 12:33:40. Copy this stamp exactly into any proof file. Retyping it from memory or from a screenshot invites a transposed digit, and a wrong stamp points to a different capture or to none at all.
Listing every capture
Replacing the stamp with an asterisk lists all the captures for a site. The help page gives the pattern web.archive.org/*/www.yoursite.com and says the best way to see all archived files for a site is the same pattern with a trailing /*. For a single publisher article, the asterisk view shows the days on which the page was captured.
Why the stamp must be checked inside the page
The help page notes that the date code embedded in the address shows when a capture was made, and that it should be checked while navigating an archived site. If you click a link inside an archived page, you may land on a capture from another date. Practical suggestion: after every click inside an archive view, read the stamp again before you quote anything from it.
Archiving one page with Save Page Now step by step
The help page Save Pages in the Wayback Machine (Opens in a new tab) describes the method plainly: put an address into the form, press the button, and the page is saved. The saved page receives a permanent address. The method saves a single page and its images and CSS. It does not save outlinks and cannot start a crawl of an entire site. The page also says the submission is anonymous because the archive does not keep the submitter's IP address.
Before you submit
The record preserves whatever the page looks like at that moment, so check the page first. The link must be visible, the anchor text must read as agreed, and the surrounding paragraph must be complete. Practical suggestion: a flawed record stays in the history, so the fix is another capture.
The submission itself
- Copy the page's address from the browser bar and strip tracking parameters.
- Open the live page one more time and confirm that the link is visible.
- Open the Save Page Now form on the Internet Archive and paste the address.
- Submit and wait for the confirmation.
- Copy the permanent address that results, including its timestamp.
- Open that address and confirm the paragraph with the link is readable.
- Write the address, the date and your name into the proof file.
Using a browser extension instead
The help page names the Wayback Machine Chrome extension: open the page, click the toolbar icon, and choose Save Page Now. It also lists Firefox, Safari, iOS and Android options, and it says the extension fails on certain pages. Practical suggestion: whichever route you use, open the resulting permanent address yourself. A silent failure leaves you with a proof file that points at nothing.
What the form leaves out
Because outlinks are not saved, the page your link points to is not captured by the record of the article. If the destination page also matters, for example a landing page whose text you want to keep as it stood on delivery day, it needs its own capture. The help page has no statement about advanced options, screenshots or error handling for Save Page Now.
Reading the snapshot calendar: blue dots and delays
The calendar that opens after "Browse History" is the quickest way to see which days have captures. The coloured dots are not decoration. Each one encodes the status code the crawler received that day.
The four colours
| Colour | Meaning on the help page | Use for a proof file |
|---|---|---|
| Blue | 2nn status, good. The page suggests you probably want the blue dots or links | Preferred |
| Green | 3nn status, a redirect | Follow the redirect and open the destination capture too |
| Orange | 4nn status, client error | The page did not load that day; not usable |
| Red | 5nn status, server error | The publisher's server failed; not usable |
The delay between crawl and display
The help page says there is a 3-10 hour lag between the time a site is crawled and when it appears in the Wayback Machine. That sentence is about crawling. For Save Page Now the save-pages page says the page is saved right away and receives a permanent address. Practical suggestion: open the permanent address immediately after saving. If the calendar does not show the capture yet, check again later the same day before taking a second capture.
Checking the capture after saving
A short review catches problems early. Confirm the dot colour. Confirm that the paragraph with your link reads correctly. Look at the images, because the help page explains that broken or gray images indicate the images were not archived, and that you can test an image by entering its address on its own. It also says that with JavaScript off, images and links load from the live web, not the archive. For a link carried by an image, that detail decides whether the capture shows what you think it shows.
A fallback when a page blocks archiving
Some pages cannot be captured, and the proof file needs a plan for them rather than a gap.
Why a capture can fail
The help page lists the causes of missing or incomplete archives: the crawler did not know the site existed, the site was password protected or blocked by robots.txt, the owner asked for exclusion, or technical barriers such as JavaScript that generates links and orphan pages with no inbound links. The save-pages page adds that some sites prohibit crawling and a few have SSL settings that make saving fail.
The manual evidence set
- A full-page screenshot that shows the address bar and the system clock.
- The page saved as an HTML file, or the paragraph containing the link copied from the page source.
- A line recording the address, the check date and the check time.
- Written confirmation from the publisher of the publication date and the page status.
- A second look at the same page after a few days, with the difference noted.
None of these carries a third-party timestamp, so the file must state which tool was used, by whom and when. A screenshot is an editable file on its own. With the address, the date and the source copy next to it, it becomes much harder to dispute.
Google's cache link is not a fallback
People used to open the cache link under a search result to see what Google had stored. The Search Engine Roundtable report (Opens in a new tab) says Google removed that link from search results, and quotes Google's Search Liaison, Danny Sullivan: "Yes, it's been removed" and "it was decided to retire it". In the same post he says he hopes links to the Internet Archive could be added where the cache link used to be, and adds "No promises." So the cache link is no longer a route to evidence, and the archive and your own files carry the weight.
When an owner asks for exclusion
The help page explains that a robots.txt file or a direct request from the owner can keep a site out of the archive. A removal request goes by email to info@archive.org, naming the URLs, the period to exclude and, where relevant, the period during which the requester controlled the site. The request starts a team review, and the page makes no guarantee about the outcome. If a publisher has a page excluded, your own screenshots and source files from before that date are what remain. Keep them.
For a page that is removed rather than excluded, see removing a guest post with a 404, 410 or 301 and link reclamation.
The evidentiary value and limits of an archive record
An archive record is a third-party view of a page on a particular date. Its strength is that the record is not a file either side edits alone. Its limits come from what the help pages themselves say about completeness and dates.
What a good record can show
- That the page was reachable when the capture was taken.
- The text, the anchor wording and the visible structure at that moment.
- That the capture carries the date and time written in its address.
What it cannot show
- That the page is live now.
- That the link stayed in place after the capture date.
- That a search engine crawled or indexed the page. The archive's crawler and a search engine's crawler are separate systems.
- That every date is covered. The help page says archived dates may not be 100% complete for every site.
Certified records
The help page sends requests that need a certified record to an affidavit procedure at archive.org/legal. An address you saved with Save Page Now is a working record for a delivery check, not a certified document, and the two should not be described in the same words in a client report.
A layered proof file
Practical suggestion: let the archive address be one layer in a proof file. The first layer is a dated screenshot of the live page with the page source. The second is the archive address with its timestamp. The third is the order record and the publisher's messages. Reading the link's href and rel values belongs to browser tools and is a separate check.
The publisher's side of the story matters too. See the lifecycle of published guest posts for how published posts are updated or archived, and news site archive structure for where content sits after publication.
How to write the record in the file
An address pasted into a spreadsheet is not yet a proof line. Someone opening the file later should learn what the record is, who took it, when, and what it shows, without asking anyone.
| Field | What goes in | Why |
|---|---|---|
| Order or job number | The delivery the record belongs to | Ties the record to one purchase |
| Live address | The cleaned page address | Names the page |
| Permanent archive address | The full address with its stamp | Everyone opens the same version |
| Capture date and time | The stamp translated into a readable date | No decoding needed |
| Dot colour | Blue, green, orange or red | Records the status the crawler saw |
| Where the link sits | Paragraph or heading | Tells the reader where to look |
| Who took it | Name or role | Fixes responsibility |
| Companion files | Screenshot, source copy, messages | Covers what the archive leaves out |
Where to keep the files
Practical suggestion: keep each record in a folder or a spreadsheet row that starts with the order number, add the date to every file name, and do not overwrite the first record when a later check is made. Because the archive is not a backup service, your own screenshot and source copy sit beside the permanent address.
Connecting captures to a content inventory
When one client file holds several publication pages, a table with one row per page keeps the capture dates visible. For deciding which pages to update, merge or retire, the content inventory decision matrix gives a structure to reuse.
When to take a capture and what goes into the hand-off
Decision rule
Decision rule: if the publication page is live and the form accepts it, take one Save Page Now capture and write down the permanent address and stamp. If saving fails, move to the manual evidence set. If the dot is not blue, do not rely on the capture and take another. Bulk archiving is not needed, because the form saves one page and the goal is one publication page.
On the question of the publisher's approval, practical suggestion: for a publicly reachable page, a single-page capture does not need a separate approval step, and bulk captures should not be taken. Telling the publisher can be a courtesy.
Hand-off checklist
- The page address is cleaned and opens live.
- The link and anchor text are visible and correct.
- One Save Page Now capture is taken.
- The permanent address with its stamp is copied exactly.
- The stamp is translated into a readable date and written in the file.
- The captured page is opened and the paragraph with the link is readable.
- The calendar dot is blue and noted.
- A second capture of the destination page is taken when it matters.
- A screenshot and the page source are saved the same day.
- The order number and the name of the person who took the record are written down.
- A live check date for the current state is scheduled.
Applying it: a trampoline park roundup
Example scenario: a trampoline park is named in a local roundup of weekend activities for families. The roundup has a short paragraph about the park and a link to its birthday-party page. The park's marketing contact wants the record in the delivery file.
What the contact does
- Opens the roundup, confirms the link to the birthday-party page is visible, and cleans the address.
- Saves the roundup with Save Page Now and copies the permanent address.
- Opens the capture and reads the paragraph with the link.
- Checks the calendar, sees a blue dot for that day, and records the colour.
- Saves a screenshot and the page source the same day.
- Saves the birthday-party page too, because outlinks are not part of the first record.
- Writes each address, date and colour into one row under the order number.
| Record | It supports | Its limit |
|---|---|---|
| Capture of the roundup | The paragraph and link existed on the capture date | Says nothing about the following weeks |
| Capture of the birthday-party page | The destination's wording on delivery day | Does not prove the roundup's link points there |
| Screenshot and page source | The href and anchor wording | No third-party stamp |
| Publisher email | Publication date and status | Not a view of the page |
What changes later
Example scenario: a few weeks later the roundup's headline changes after a seasonal update. The contact takes a new capture and puts the two blue dots side by side. The link is present in both. Both permanent addresses go into the file, each with its own date, and the first record is not replaced.
Later still, the calendar shows an orange dot for a day when the page returned a client error. That capture proves nothing about the link, so it is not used. The contact asks the publisher about the outage and keeps the earlier blue records. The state of the link today still requires a live check, and the contact schedules one.
Agency hand-off
Example scenario: if an agency delivers the placement, it adds the permanent address and the readable date to its delivery message and attaches the screenshot. The client opens the same address and finds the same paragraph. When a page cannot be saved, the agency says so in the message and lists the manual evidence set it used instead.
Misreadings to avoid
Being in the archive is not being indexed
The Internet Archive's crawler and a search engine's crawler are different systems. A capture in the archive does not show that the page appears in search results.
A blue dot is not a verdict on the link
A blue dot means that the capture came back with a 2nn status. How the link was marked up in that capture is read separately, from the captured page's source or in browser tools.
Treating a gray image as a broken link
The help page says broken or gray images indicate that the images were not archived, and that an image can be tested by entering its address on its own. A gray image therefore points to a missing image capture. Practical suggestion: check whether the paragraph with the link still reads correctly before deciding that the capture is unusable.
Quoting the stamp without checking the redirect
The help page describes a green dot as a 3nn status, a redirect, and notes that the date code in an address shows when a capture was made. Practical suggestion: when the dot is green, open the destination capture as well and read its own stamp before quoting a date in the report.
The archive is not your backup
The help page says plainly that the terms of use do not cover backups for the general public. Your own files are the primary evidence.
Frequently Asked Questions
Can Save Page Now capture a whole publisher site?
No. The help page says the method saves one page, does not save outlinks, and cannot be used to start a crawl of an entire site. For whole sites the page points to Archive-It, a subscription service.
Do I need the publisher's permission to take a capture?
The help pages describe no permission process for Save Page Now. Practical suggestion: a single-page capture of a publicly reachable page is enough, and bulk capture should be avoided. A publisher that wants a page excluded can ask the archive by email.
Which address do I copy into the proof file?
Copy the permanent address that Save Page Now returns, with its fourteen-digit timestamp intact. The help page says the number in an archived address is a timestamp, so an address stripped of it loses the date. Practical suggestion: write the readable date and the live page address on the same line.
Can I archive a page behind a login?
The help page lists password-protected sites among the causes of missing archives, so a capture of such a page should not be expected. Practical suggestion: use the manual evidence set of a dated screenshot, the saved page source and the publisher's written confirmation, and note in the file that the page required a login.
How do I cite an archive address in a client report?
The help page gives an MLA-style recipe: cite the web page in the usual way, then add the Wayback Machine details, meaning the archived address and the date you accessed it, and use the closest archive date when there is no update date. Practical suggestion: in a client report, list the permanent address, the capture date and the date you opened it on separate lines.
What if the page cannot be archived?
A robots.txt block, a password, an owner exclusion or a technical barrier can stop a capture. Fall back to a dated screenshot, a saved page source and the publisher's written confirmation, and keep all three together.
Conclusion
The Wayback Machine keeps a dated copy of a publication page, and the record is not a file either side edits alone. Save Page Now takes one page, leaves outlinks aside, and carries the capture time in the address. A blue dot signals a good capture, but no capture proves the page is live today or that the link stayed in place. That is why the archive address belongs in a file next to a screenshot, the page source and the publisher's confirmation, and why a live check is scheduled separately.
If you want to run purchased placements with a delivery file that holds this kind of record, you can look at the backlink packages.
Post a comment
Please log in to leave a comment. Login