This is neither a "Google is the devil" rant nor a conspiracy theory but a simple pointer as to why "searching the web" has long stopped to be what you thought it is. My goal is to improve general understanding of the reasons leading to an increase of incorrect, bad, or missing search results during recent years.
Most people nowadays believe that search results are more or less meaningful and cover all results available. That may well be the case as long as the search (and the possible results) fulfills certain parameters. Failure in complying with what is seen as "relevant" (and it is not up to you to decide) inevitably results in either meaningless nonsense or no results at all. People normally do not notice that they are being played; that is until they start searching for specific items not in the main focus of search engines.
For now, let us look at an example which may not appear to have anything to do with searching for china, porcelain items, or marks, at all: imagine wanting to find the original release year of a certain song issued on 78rpm vinyl by a popular World War 2 era singer. Whatever you may be presented with by search engines is often far off, pushing you towards music sites like Spotify, Deezer, Apple Music and similar venues which do not care about correct historic data but want to sell CDs, CD compilations, or music download/streaming.
And so you get the search results of stuff released on compilations, re-issues and whatnot but no hard info, like the year you are actually looking for, even if you include "release" or similar keywords. On top, you also get presented results for artists originating from the same country and/or covering the same genre. Which is like searching for a Porsche automobile and getting search results for everything between BMW and Volkswagen. Not quite what you wanted, right?
From around 2005 onward the big players of the internet (not only Google, eBay, Bing, etc. but also companies like United Internet, etc.) saw that the amount of data having to be stored, indexed and then held available had reached levels which would, even with development of new technologies, soon be impossible to handle efficiently. Just for measure, Google added around ten PetaByte (1 PB = 1024 TB) of raw data (naked information only, no images, audio, or clips) to its structures each day and has to index and sync that data with around of their own 17 server farms (source: ©2020 Google). What at first appears to be "not much" receives more weight when seen from the point of exponential growth, so wonderfully described in the "RICE AND THE CHESS BOARD" story. Hold in mind that Google is not alone and that each added data source has also to be indexed, cross-referenced, and backed up world-wide and in real time.
So the big companies decided to cut back on data compilation, agreeing on plans to reduce data sourcing, overhead, indexing, and of course storage. What could go wrong, right? Well, the process was made dependant on "relevance". And not, as you may believe, per "oh, this could be interesting info" relevance but the plain "money, MORE money, MOAR MONEY!" kind of relevance. Which means that search engines nowadays ignore regular 'plain' or 'historic' data in favor of all kinds of (money-making) shop/store data, presenting you useless and unfitting storefront nonsense instead of true information as provided by fan or collector sites. Even many Wikipedia entries land somewhere in the back rooms (result pages 3+) nowadays.
Before some naysayers move in and claim that I was probably too lazy or stupid to use fitting keywords, hold in mind that presenting song lyrics are NOT omitted from search when one excludes the term LYRICS because there are tons of non-English websites that do not use this term. Not forgetting the fact that sites like Apple Music, Spotify and others pop up even if you have excluded those specific sites. It is actually no secret that keyword handling has become worse in recent years. Another flaw is that you can not create white/blacklists for your searches - search engines want to force-feed you with whatever nonsense they come up with simply because they are paid to do so; disabling their money-makers would seriously damage their business model.
Just to push the issue, some self-proclaimed experts find the following search string completely okay and "easy to use" :
"ARTIST NAME" 78 -anghami -apple -deezer -gaana -jiosaavn -muziekweb -shazam -spotify -amazon -ebay -youtube -lyrics -songtext -reddit -issuu
As if we all wanted to write a bestseller novel each time we are looking for something, right ? As mentioned above, no search engine allows the user to have personal black/whitelist. I therefore use a self-coded search phrase generator which offers a list of presets, depending on item group or context, as I am simply sick and tired of all the commercial crap forced down my throat when looking for something specific (like a certain manufacturer or some historic background).
Years ago it was commonly expected that closed ten year old eBay auctions were available in searches. Google indexed and shared the links, eBay stored the auction data and the images. That procedure is history as the amount of data generated by sites like eBay has long exceeded the amount that can be easily handled on a world-wide scale. Based on complex algorhythms the core data of the single parties (in this example: Google / eBay) as well as the data shared between the different parties (example again: Google / eBay) is continuously monitored and stripped of all data packages deemed not 'relevant' enough as per the rules these parties agreed upon.
This results in all finished/expired/closed auctions older than [x] (the agreed factor as per agreement) simply vanishing. It is for example deleted from the eBay servers and consequently does not appear in search engine results anymore. Observant users will find borderline auction data where the template is still there but all eBay-hosted images are blank. Another great example is the data-mining pest Pinterest which appears to present data even if it is simply not there anymore when you try to click through. Their "service" requires you to sign up (which fuels their true business: getting your data) but once you are hooked they tell you that the data/image you were trying to access is not available anymore ...
It all boils down to the following: search results nowadays are by far not as meaningful or reliable as they were a few years ago. If you are looking for a price to put on your item, a rough estimate on how many other people tried to sell / buy it in the past, or even if you are only looking if there is another image of your item on the web: the last thing you can rely on is a search engine, contradicting its reason of existence.
The number of people that end up on sites like mine, truly convinced that their item must be rare or one of a kind, is constantly rising because they got played like a fiddle by Google, Bing or whatnot. Sadly, that is not all: the typical user is also becoming more and more arrogant and ignorant about it. A family had tried to find out more about an inherited dinnerware set, claiming to have unsuccessfully combed through the whole internet. Now they wanted to know if I could shed some light on the value of their set as they wanted to insure it.
Some readers may remember the time around 2005 when eBay was literally flooded with cheap Winterling china, forcing prices to crash. By 2022 the overall desirability had dropped off the scope and prices were abysmal, with 144 piece sets not even being sold for 20 (twenty!) US$ because absolutely nobody wanted this stuff. Confronted with the truth, the family accused me of lying and then contacted two auctioneers, making real fools of themselves.
Whatever you are looking for, do not take common search results for the godsent truth. Users are being played big time and the only interesting thing for sites like Google or Bing are personalized search profiles. Try to support search sites like DuckDuckGo.com or Qwant.com which do not trace users and provide far less injected commercial stuff. Their results sometimes appear pretty basic, however their search algorithms in team with a little user-supplied brains can lead to far better search results than Google has to offer.
(Version 0.80)
© 2004-2024 C.S.Marshall, all rights reserved