Remember the first time you connected to the Internet? I do. Even at the then breakneck speed of 9.6kbps, it was the content that fascinated me. Though the search technologies were primitive, I somehow managed to find The Louvre's Web site, a near magical experience.
Internet search technologies today have progressed far beyond this elementary functionality. Anyone who has browsed one of the many Internet search sites will tell you this. They will also tell you that searching inside the enterprise is a far less gratifying experience.
Historically, enterprise search technology has been expensive, limiting it to the very largest of organisations. And while the technology has become cheaper to implement, the barrier to entry remains high. Here are six reasons why:
1. Search volumes inside the enterprise remain low
Search is often a people issue and private domains are an ongoing part of the struggle.
Grant Hodgkinson is MD of Mint.
One of the ways in which Internet search providers continue to improve their search results is by analysing and reviewing search requests. Internet search vendors have teams of resources dedicated to reviewing search terms, looking at trends, improving the manner in which certain search terms are highlighted, and improving the hit ratios on content.
No organisation can compete on that scale. The number of search transactions inside the enterprise is simply too small to deliver the same volume and speed of improvement.
2. Company dictionaries are difficult to define
Different organisations have different ways of interpreting or applying the same word. The word "residual" is one example: in a financial services company it means something very different from the same word in a manufacturing business. Users expect internally facing search technology to deliver results in line with the company definition of a term, and to incorporate synonyms where necessary. Unless this is defined and carefully maintained, they are likely to be frustrated.
3. Company data is stored in multiple, diverse repositories
When Internet search systems trawl content, the operation is performed across a consistent set of standards - all Web sites deliver content in HTML across http:// or https://. Internet search vendors have been able to capitalise on this consistency.
Inside the enterprise, there are generally multiple repositories of content, each with its own access methods and application programming interface (API). The data in those systems may be related, but the relationships are not defined so they can be deduced programmatically. To get search applications to index this content so search results are meaningful, often requires a lot of data scrubbing, as well as the development of cross-repository indexing plug-ins, which can be costly and time-consuming.
4. Company data is secure
Content exposed to users through an Internet search interface is accessible anonymously - one of the wonders of the Internet. While it is a noble aim to enable all users to have access to all content inside the enterprise, this is simply not possible. As a result, content can be accessed only through very different protocols. Each element or body of content needs a security layer associated with it. And often, the security layers across content are not consistent. To enable an indexing engine to understand these nuances requires much work.
5. Time to meaningful search results
At home, users can browse results at their leisure. At work, unless users are information workers who are expected to spend time searching for information, they simply do not have the time to waste. Imagine a customer-facing person in your business spending an hour looking for information. You would quickly end up with unhappy clients.
6. Content ownership
Content on a Web site has already been approved for publication and anyone can use it. Companies, on the other hand, do not publish their trade secrets online. Also, in large organisations, it is the unfortunate truth that some people try to establish personal empires, and become unwilling to share their department's content with others. This is not necessarily about content security, but about the power that is bestowed by ownership. If department A has to beg department B for content, then department B can feel more important.
Like everything else, search is often a people issue and private domains are an ongoing part of the struggle.
Despite a higher barrier to entry, however, enterprise search technology can deliver enormous benefit to organisations. Next time, I will look at some of the key steps that can be taken to prepare organisations to take advantage of this exciting wave of productivity enhancement.
* Grant Hodgkinson is MD of Mint.
Share