Simplifying Search Keyword Analysis

Lawson Kurtz, Former Senior Developer

Article Category: #Strategy

Posted on

Search keyword analysis for analytics has always felt more painful than necessary. Such analysis typically involves filtering through hundreds or thousands of subtly-different variations of search queries to form an impression of the relative importance of particular words and phrases. It's a time-consuming and often subjective process that isn't any fun at all.

Since current analytics services lack a tool to digest raw search query data into a list of the most important words, we built one fit for the task.

Introducing Search Words

Search Words is an application that restores some sanity to the search keyword analysis game. Raw search query data goes in, and a list of searched words and their summed search volumes comes out.

Specifically, Search Words is a Rack application (built on Sinatra) that takes a CSV file containing search queries with corresponding search volumes, and returns a combined list of search volumes for each word contained within those queries. The application also combines plural and singular versions of a particular word if both forms are present, and will ignore common words (e.g. "a", "and", "or", etc.).

Example Input and Output

 Input
|---------------------------------------|
| Search Phrase | Total Searches |
| ======================================|
| Lemurs | 1,000 |
| durham bulls | 500 |
| bull fighting | 100 |
| lemur dance | 2,000 |
| Durham | 700 |
| How to catch a lemur | 35 |
|---------------------------------------|

Output
|---------------------------------------|
| Search Word | Total Searches |
| ======================================|
| durham | 1,200 |
| bull | 600 |
| fighting | 100 |
| lemur | 3,035 |
| dance | 2,000 |
| catch | 35 |
|---------------------------------------|

Demo and Installation

A demo of the Search Words application can be found here. To install Search Words, see the project's repository on GitHub. Complete installation instructions can be found within the readme.

Search Words is a very simple application at the moment, so feel free to drop us a line if you have suggestions for improvement.

Related Articles