Metadata tips,tools and tricks : Simple explanation how search engines on microstock sites work



When we say search engine, all of us think about the field with magnifier where we insert some text and results magically appear right after clicking on magnifier icon. But what is that field and how does it work.

First of all, the search engine is an application module, part of the program, which uses algorithms to read and display data from the database ("crawling") . Depending on the needs it can be simple, where it can only display data, or more complex, when needs to store data in the database as well ( "crawling and indexing") . All stock sites are using complex search engines, and collect all searched words ( even misspelled). That data is stored in the database and then compared to existing searches. 
When someone performs a search, search engines scour their index for highly relevant content and then orders that content in the hopes of solving the searcher's query.  Every searched keyword is connected to the images that are downloaded, or the user had any other interaction ( adding to the library, cart, etc...) The image that contains searched keyword  if user interacted with it gets "points" ( ranking) again stored in the database . Next time when the same keyword is searched, images with "Points" will rank better and will be shown first.

If you insert more keywords in the search engine, for example, " tropical destination" , the search engine is going to treat that input:
  1. ---- first as one keyword, so if any files have "tropical destination" as one keyword, those files are going to be displayed first. 
  2. ----After that, the words are going to be separated and searched in the database as single words. Second results are files that contain both keywords as a separate word.
  3. ---- And third result will display files that contain only one word.
  4. ----after those attempts , If there isn't any file that contains searched input, the search engine is going to try to find if there was misspelling, by searching for "misspelled" words with "points", comparing input with most similar words also with "points", or just display that there isnt such file in the database.
This is the simplest way to describe "work" that is being done by search engine app. Most sites have extremely complex search engines which are again being updated all of the time, depending on the website needs. 

Here is an explanation of how search engines work on different stock agencies :

Metadata tips,tools and tricks : Understanding Shutterstock search engine



Share on Google Plus

About Negmardesign

0 comments :

Post a Comment