Tuesday, January 10, 2012

Search School: Words and Phrases, Not Sentences

Searchers usually take one of two approaches to doing a search. Beginning or less experienced searchers usually try typing a whole sentence or question into the search box This is called a natural language search because it mirrors how we naturally speak. For instance, a student using a natural language who is looking for information on the causes of the Second World War might use this query:

What were the causes of World War II?

On the other hand, a more experienced student searcher might use a keyword search. This type of search uses just one or two important (or key) words that the search engine will match against it index of pages.

causes "World War II"

A keyword search is usually more effective for a few reasons.

Non-essential Words: Remember that search engines look for all the words in your search query. If you said what were the major underlying causes of World War II, Google would assume that the words major and underlying have to be somewhere on each of the pages in your results.

Stop Words: Words like a, an, the, of, who, whom and similar words that do not have any meaning by themselves are usually ignored when you include them in your search. These are called stop words.

Precision: Focusing on just the essential concepts in your question allows you to be more precise in your thinking. Look at your search. Are your search terms too vague? Can you add a few words to tweak the search? This is harder to do with a sentence instead of a handful of words.

Phrases Searching: By default, a search engine will look for all of your words anywhere in the page. For instance, a search for World War II causes might return a page where all of your words show up different places in the page. So, the page might be one that says

The European powers divided up the world into various colonies…They were constantly at warTwo causes for this were…

This page might come from a web page about European Exploration instead of World War II.

For well-known phrases like World War II, the search engine will return relevant pages, but we want to eliminate irrelevant pages if we can. Putting quotations around any concept that has more than one word tells the search engine to make sure that only pages where the words are right next to each other on the page should be included in the results.

Practice Exercise

Here is a simple technique for modifying your search so that it only includes what you need for a good search.

1) Write out what you want to know in question form (natural language). Make sure to be as precise and specific as possible.
2) Go back and circle the most important concepts in the question—the 2 or 3 words that are the main ideas of the question. 
3) On a piece of scrap paper, write down the concepts from your question. Using your background knowledge, or after skimming a reference article (an encyclopedia article, Wikipedia, etc.) or textbook, brainstorm any synonyms or more general/specific ways to state the main ideas or relevant related concepts.

You now have several different ways to search for the same question.

 For more posts in this series, check out: How Search Engines Work and Crafting a Better Keyword Strategy

No comments: