Google Cache

» see cache:google.com as it lived, on the wayback machine

~ the last words ~

i typed cache:url for years. it was the move when a page returned 500, when the live version had been quietly edited to remove the paragraph you wanted to quote, when a paywall had been wedged in front of an article you knew you had read free a month earlier. by 2010 it was muscle memory. by 2018 it was the only way to still recover content from a few sites that had aggressively deindexed their own archives. honestly, the cached copy was the one piece of google's infrastructure that worked as a service to the people using the search engine rather than to the advertisers paying for it.

the cache shipped with google search itself in 1997. larry page's original architecture stored a copy of every page it crawled, both to support the search index and to serve the cached page back to users on demand. the cached link sat to the right of every result on the search engine results page through almost the entire history of google's consumer ui. the cache: search operator gave you direct access from the address bar without having to scroll to find the link first.

danny sullivan posted the retirement announcement on x in late january 2024 under the google searchliaison handle. the message was short and described the feature as one of google's oldest. by february 2 2024 the cached links were gone from the search results ui. by march the cache: operator no longer worked at all. google later explained the removal by saying the web of 2024 was more reliable than the web of the early 2000s when the feature had been most needed.

~ the record ~

~ what it did ~

the surface use cases were small in number. recover content that had been deleted or edited on the live page. read a page when the source server was down or unreachable. bypass a paywall that had been added after google had indexed the article free. read on a slow connection without the live page's full asset payload. read a non-mobile version of a page on a phone when the css was breaking the layout.

seo people had their own use for the cache, which was to confirm what version of a page google was actually serving as the indexed copy. checking the cache: result told you whether your last content change had been picked up. that workflow had been a daily practice for the seo profession for almost two decades. when the cache went away, the closest replacement was the url inspection tool inside google search console, which only works for sites you have verified ownership on.

~ what got buried with it ~

the small workflow of using cache: to read deleted blog posts was the part that ended cleanly. for many of the older sites whose archives have decayed over the last decade, google's cached copy had been the only remaining accessible version. when the feature shut down in february 2024 a large set of these pages became unreachable in any meaningful way, since the wayback machine had not crawled most of them at the depth google had.

the paywall workaround was the other quiet casualty. for about a decade the cache: trick was widely shared as the move for getting around metered paywalls on news sites, particularly bloomberg and the wall street journal and the new york times. many of those publishers had explicitly disabled the cache by 2020 anyway. but for the long tail of smaller publications that had not bothered, the trick had still worked right up until the shutdown.

~ the long tail ~

google's stated reason for the removal was that the original purpose of the cache had been to help users reach pages when web hosting was unreliable, and that the web had since become reliable enough not to need that service. the replacement plan was internet archive wayback machine links added to search results, which began rolling out in september 2024 about seven months after the cache itself was gone.

the wayback machine integration is partial. internet archive coverage is patchier than google's was, snapshot frequencies are lower, and not every google search result gets a wayback link in the ui. for the daily small uses the cache had served, the replacement has been a downgrade. the cache: operator now returns a search results page for the literal string "cache:url" instead of opening anything. for a feature that ran continuously for 27 years across every redesign of the google search ui, the shutdown was unusually quiet.

@}-->--

              )
             (_)
            .-'-.
            |   |
            |   |
            |   |
            |   |
          __|   |__   .-.
       .-'  |   |  `-:   :
      :     `---'     :-'
       `-._       _.-'
           '""""""

« previous  ·  random  ·  next »