Google's popularity in the consumer world causes many people to have similar expectations of enterprise search. Beyond the 3000 plus engineers that Google has working on its search algorithms, ranking and relevancy of results, there are many other ways that the two cannot be compared. Taking an off the shelf product like SharePoint 2013 or Office 365 or any number of other specialty search applications designed for the enterprise simply does not work for many reasons. Here are six key differences of why Enterprise Search is nothing like Google search.
1. Google search doesn’t really do refinement
Users learn from Google that their information should be in the first 10 blue links or better. In the enterprise the first 10 documents that match your search terms aren’t likely the ones you were looking for. You’re going to have to refine by date, or narrow the results by site or department or other information.
Amazon's taxonomies and categories -- including product lines, departments and brands -- are a much better analogy of what should happen in enterprise classification. The reality is most users won’t add metadata, so data extraction and automation is a clearer path to consistent metadata for refinement.
2. Google primarily focuses on pages, Intranets search primarily focus on documents
The enterprise stopped producing lots of webpages after Frontpage went out of fashion. Documents need more care and feeding as it relates to search in order to provide relevance, context and meaning, hence the strong need for metadata strategies.
3. Inbound links are rare in the enterprise
There is little in the enterprise to automate page rank. Google has been adding social links into its algorithm. The link backs continue to be what Google search relies on to determine the page rank for its SEO. In the enterprise there are so few links that it’s extremely unreliable to give it much credibility. It's more credible to determine how close the page is from the home page of the Intranet. Click distance and popularity are a better judge for relevance.
4. Google search is designed to create revenue
Those top results are keywords that are fought over by vendors. There’s little comparison in the enterprise. You don’t buy results in the enterprise. Even when you can be added as recommended results, these are managed.
5. Google expects everyone will optimize their sites otherwise … 'Who cares?'
Everyone does Google's work for it. The enterprise requires strategy and partnering to get relevant, authoritative results. People who want their information discovered on Google have to do SEO to optimize it for index, including adding titles, description, keywords, feeding pages and URLs. In the enterprise, the users often don’t know how titles, keywords and file names have any impact if any ... otherwise why are so many files missing titles?
Users must be educated and taught. Have the information managers who manage the sources where the data lives work with the indexer to align the information. Add a process to allow departments or site owners to jockey for recommended results in enterprise search. Without some management you’ll be relying too much on out of the box algorithms for premium placement in relevance.
6. Google is designed around public use with anonymous access. It isn't designed to handle secure documents
Everything is about security in the enterprise. Every system has logins and permissions-based access to information. The difference between accessible and restricted information impacts both search performance and filters results depending on a user's rights.
The next time someone says, “Make our search like Google,” know that it doesn’t work like that.