The High Cost of Assuming Enterprise Search Works Like Google

by Mindy Carner

 

A common mistake when seeking to improve access to content across the enterprise is assuming that intranet systems will act as “Google for the enterprise.”  Unfortunately, this assumption is incorrect because the technology that Google employs is designed for the World Wide Web, not for an intranet. This means that the algorithms, measures and tricks that search engines use to rank your website cannot be mirrored for internal content.

The following list describes several elements of the internet search engine puzzle, and why they do not apply to internal search applications. 

Inbound links: The value of a web page is a function of the number of other websites that link to it. Arguably, the more sites that link to yours, the more likely it is that your content is good and trusted. The more credible a web page’s content, the higher its ranking in search. This algorithm simply does not apply to intranet content that does not have such wide variances in quality of content.


Click-throughs: Web search engines also use click-throughs to determine ranking for a page. The more times that people click through your link on a given search, the higher that link will rise in the ranking. This algorithm is also absent from intranet searches.


Structured data: The collection of web pages that constitute the bulk of content on the World Wide Web is called structured data. That’s because the data is organized using HTML (HyperText Markup Language), which categorizes the web data using tags that can tell a search engine what’s a title, what’s a header, and what’s plain old paragraph writing. HTML also includes keywords, descriptions and much more structured markup that a savvy page creator can use to tell search engines what their web page is about. Because intranets rarely, if ever, have sophisticated organizing mechanisms, they produce copious amounts of unstructured content. This includes your Office documents, PDFs, OCRs, and other digital assets that have no formal structure to their content. This lack of formal structure makes for unstructured data, and poor search results.

The danger in not understanding this vital difference between enterprise search applications and internet search engines can lead companies to invest thousands of dollars in a technology solution that will underperform unless they also invest in the people, processes and metadata necessary to transform unstructured data into structured data. Without these additional investments, an off-the-shelf technology application is likely to yield nothing more than a very expensive keyword search.

For an in-depth look at how taxonomy fills in the gaps of an enterprise search solution, please read the Optimity Advisors Orange Paper “Enterprise Search and Taxonomy: Filling the gaps of an out-of-the-box enterprise search solution.” 

 

Mindy Carner is a Senior Associate in the Media and Entertainment practice at Optimity Advisors