How Search Engines Work?
Posted: Tue Dec 27, 2005 1:13 pm
You know how important it is to score high in the SERPs. But your site isn't reaching the first three pages, and you don't understand why. It could be that you're confusing the web crawlers that are trying to index it. How can you find out? Keep reading.
You have a masterful website, with lots of relevant content, but it isn’t coming up high in the search engine results pages (SERPs). You know that if your site isn’t on those early pages, searchers probably won’t find you. You can’t understand why you’re apparently invisible to Google and the other major search engines. Your rivals hold higher spots in the SERPs, and their sites aren’t nearly as nice as yours.
Search engines aren’t people. In order to handle the tens of billions of web pages that comprise the World Wide Web, search engine companies have almost completely automated their processes. A software program isn’t going to look at your site with the same “eyes” as a human being. This doesn’t mean that you can’t have a website that is a joy to behold for your visitors. But it does mean that you need to be aware of the ways in which search engines “see” your site differently, and plan around them.
Despite the complexity of the web, and dealing with all that data at speed, search engines actually perform a short list of operations in order to return relevant results to their users. Each of these four operations can go awry in certain ways. It isn’t so much that the search engine itself has gone awry; it may have simply encountered something that it was not programmed to deal with. Or the way it was programmed to deal with whatever it encountered led to less than desirable results.
Understanding how search engines operate will help you understand what can go wrong. All search engines perform the following four tasks:
Web crawling. Search engines send out automated programs, sometimes called “bots” or “spiders,” which use the web’s hyperlink structure to “crawl” its pages. According to some of our best estimates, search engine spiders have crawled maybe half of the pages that exist on the Internet.
Document indexing. After spiders crawl a page, its content needs to be put into a format that makes it easy to retrieve when a user queries the search engine. Thus, pages are stored in a giant, tightly managed database that makes up the search engine’s index. These indexes contain billions of documents, which are delivered to users in mere fractions of a second.
Query processing. When a user queries a search engine, which happens hundreds of millions of times each day, the engine examines its index to find documents that match. Queries that look superficially the same can yield very different results. For example, searching for the phrase “field and stream magazine,” without quotes around it, yields more than four million results in Google. Do the same search with the quote marks, and Google returns only 19,600 results. This is just one of many modifiers a searcher can use to give the database a better idea of what should count as a relevant result.
Ranking results. Google isn’t going to show you all 19,600 results on the same page – and even if it did, it needs some way to decide which ones should show up first. Thus, the search engine runs an algorithm on the results to calculate which ones are most relevant to the query. These are shown first, with all the others in descending order of relevance.
Now that you have some idea of the processes involved, it’s time to take a closer look at each one. This should help you understand how things go right, and how and why these tasks can go “wrong.” This article will focus on web crawling, while a later article will cover the remaining processes.
Next: Stumbling
You have a masterful website, with lots of relevant content, but it isn’t coming up high in the search engine results pages (SERPs). You know that if your site isn’t on those early pages, searchers probably won’t find you. You can’t understand why you’re apparently invisible to Google and the other major search engines. Your rivals hold higher spots in the SERPs, and their sites aren’t nearly as nice as yours.
Search engines aren’t people. In order to handle the tens of billions of web pages that comprise the World Wide Web, search engine companies have almost completely automated their processes. A software program isn’t going to look at your site with the same “eyes” as a human being. This doesn’t mean that you can’t have a website that is a joy to behold for your visitors. But it does mean that you need to be aware of the ways in which search engines “see” your site differently, and plan around them.
Despite the complexity of the web, and dealing with all that data at speed, search engines actually perform a short list of operations in order to return relevant results to their users. Each of these four operations can go awry in certain ways. It isn’t so much that the search engine itself has gone awry; it may have simply encountered something that it was not programmed to deal with. Or the way it was programmed to deal with whatever it encountered led to less than desirable results.
Understanding how search engines operate will help you understand what can go wrong. All search engines perform the following four tasks:
Web crawling. Search engines send out automated programs, sometimes called “bots” or “spiders,” which use the web’s hyperlink structure to “crawl” its pages. According to some of our best estimates, search engine spiders have crawled maybe half of the pages that exist on the Internet.
Document indexing. After spiders crawl a page, its content needs to be put into a format that makes it easy to retrieve when a user queries the search engine. Thus, pages are stored in a giant, tightly managed database that makes up the search engine’s index. These indexes contain billions of documents, which are delivered to users in mere fractions of a second.
Query processing. When a user queries a search engine, which happens hundreds of millions of times each day, the engine examines its index to find documents that match. Queries that look superficially the same can yield very different results. For example, searching for the phrase “field and stream magazine,” without quotes around it, yields more than four million results in Google. Do the same search with the quote marks, and Google returns only 19,600 results. This is just one of many modifiers a searcher can use to give the database a better idea of what should count as a relevant result.
Ranking results. Google isn’t going to show you all 19,600 results on the same page – and even if it did, it needs some way to decide which ones should show up first. Thus, the search engine runs an algorithm on the results to calculate which ones are most relevant to the query. These are shown first, with all the others in descending order of relevance.
Now that you have some idea of the processes involved, it’s time to take a closer look at each one. This should help you understand how things go right, and how and why these tasks can go “wrong.” This article will focus on web crawling, while a later article will cover the remaining processes.
Next: Stumbling