How Do Search Engines Work?

praetserge

Elite Member
Executive VIP
Jr. VIP
Joined
Apr 4, 2014
Messages
6,453
Reaction score
4,860
I've been wondering for some time and I still dont understand it.
How do search engines work?
Let's say you use google and type in "best chicks online", surely google will not analyse trillions of pages before returning you the results. So it means search engines analyse the pages in the background and have a set of results for every possible query?
So if there are enough searches for a keyword, search engines will create a set of results and suggestions?
 
I've been wondering for some time and I still dont understand it.
How do search engines work?
Let's say you use google and type in "best chicks online", surely google will not analyse trillions of pages before returning you the results. So it means search engines analyse the pages in the background and have a set of results for every possible query?
So if there are enough searches for a keyword, search engines will create a set of results and suggestions?
You already have the idea. Google has pages it has already indexed sorted for each keyword according to whatever metrics they use. Since they get a fairly large amount of unique searches everyday, the algorithm will have to go through their 'list' of sites that closely answers the question the searcher is asking.

This is how I see it. Certainly, there are aspects not none to the rest of the world asides the guys at Google.
 
You got the basic idea.

They don't "create" anything when you search. They just sort their already existing index of sites and show you the results.

surely google will not analyse trillions of pages before returning you the results

Actually they do but the ones already processed and sitting in their index.

It might seem like an impossible task but it really just comes down to computing power. And they have no issues there :)
 
I've been wondering for some time and I still dont understand it.
How do search engines work?
Let's say you use google and type in "best chicks online", surely google will not analyse trillions of pages before returning you the results. So it means search engines analyse the pages in the background and have a set of results for every possible query?
So if there are enough searches for a keyword, search engines will create a set of results and suggestions?
yes that is basicly how it is done. thx to the super computer.
Google spent billion $$$ on the server annually to make sure their service up and running.
 
Well, now they have RankBrain, their ML algorithm testing the behavior of users clicking the search listings in real time. Based on the behavior of users (whether they stay on the site or hit the back button asap and come back to the Google SERPs, then Google knows that listing is not really relevant to the user's search query. This happens when several users behave this way. A listing in position 5 that performs better will slowly move to position 1.
 
That's exactly how it works. They crawl all the websites and keep updating their index (their huge database) and change rank for each website based on various factors for a particular keyword. When a user searches, they just do a check for the keyword and produce the result for the searched keyword from their latest index.
 
They USED to work by viewing backlinks as 'votes' about the context/content of a site

Now, they use backlink signals AND AI-powered predictive systems (RANKBRAIN)

User click behavior now "trains" Google on how to rank results and determine context
 
Let's say you use google and type in "best chicks online", ...

Your title is spot on. And the answer is, this is what full time advanced SEOs try to answer: how google works.

As Google employs hundreds of Science PhDs from top unis, why is it is a surprise that results are found so fast?
(speed as always been one of their main foci, as they know this improves UX)
 
Search engines work by taking list of known URLs, which at that point go to the scheduler. The scheduler chooses when to creep every URL. Crept pages at that point go to the parser where essential data is separated and listed. Parsed joins go to the scheduler, which organizes their slithering and re-creeping.
 
Back
Top