I want to touch on a number of themes in this article.
- The first is that being pragmatic and using software that might just be 'good enough' instead of perfect may allow you to make improvements to business processes in a more rapid, if iterative manner.
- The second is to highlight a real world scenario where the introduction of a 'search engine' has been studied to show the real benefits it brings to users.
- Third and finally, after all the positive elements, I want to re-iterate that simply installing a search engine is absolutely not the answer to all your information management or knowledge management problems.
A Case Study on SharePoint 2007 Search
When I joined my current company, one of Canada's biggest and leading retailers, I was quite surprised to find that we had not previously deployed an “enterprise search” facility. The company headquarters in Toronto is the home to approx. 5000 staff, the majority of whom can be characterized as 'knowledge workers' -- staff who generate, synthesize, analyze and distribute information, or items of 'intellectual capital' rather then making physical products, such as those that we sell.
Our intranet environment is based on SharePoint Server 2007 and utilizes the built-in search functionality to provide a capability to search on the news articles, information pages and documents published to the intranet. It also provides people search via some custom backend integrations with HR systems, as well as the standard connection to MS Active Directory.
This facility is not characterized as enterprise search however, as it does not index other content stores and information repositories across the organization, it is purely intranet and staff directory in scope.
Too Many Documents, Not Enough Search
My colleagues in our small “IT Knowledge Management Group' had not been idle. Although having only been in existence for about a year as a group, they had conducted comprehensive interviews at all levels of the IT division to provide some insight into peoples pain points, and one of the biggest issues was the inability to search the massive corpus of documentation generated within the IT division itself.
The division has around 500 staff, plus contractors, and the IT project management environment is based on a waterfall type Software Development Life Cycle (SDLC). This amounts to a corpus of well over a million documents. Given this, it's hardly surprising that it's difficult to find things by browsing folder structures on file shares. Oh yes, did I forget to mention, there is no document management solution in use?
The Decision to Use SharePoint Search
As you might expect, deciding to implement a search solution was not a controversial plan. Taking the pragmatic approach they decided to use the SharePoint search facilities. This was decided for two primary reasons:
- We already had SharePoint and all the licenses required, and as additional 'virtual infrastructure' was to be used for the proof of concept (POC), this was an almost zero software cost option.
- Although SharePoint might not be the best or most heavily featured enterprise search system it was definitely 'good enough' for the POC and probably for production use too.
Initially, slightly over 100 users were invited to take part, particularly Business Analysts who were struggling the most when it came to hunting down documentation. Many more users were added individually as they heard about the project and asked to participate.
A slight enhancement was soon made to the POC system by adding the limited Faceted Search features from the MS CodePlex project of the same name. This adds capability for users to refine search results based on facets such as file type, author or source repository.
SharePoint 2007 -- Faceted Enterprise Search
A Successful Result
All in all the POC has been a roaring success. 93% of the participants reported real time savings. Benefits analysis has shown that the Business Analysts are actually saving an average of 9 hours per week (24% of their time) while other staff are saving on average around 2.6 hrs/wk.
90% of the participants agree that the search system is improving their or their teams efficiency, by making them more self sufficient and less reliant on subject matter experts. So if we take a very conservative approach and say that the search facility might save 50% of IT staff 2.6 hrs/wk the potential ROI in dollars is equal to 13 Full Time Equivalent (FTE) staff.
A Search Engine Alone is Not Enough
However simply installing a search engine will not bring you to a state of information management nirvana. For example, every project has to create a Business Requirements Document -- searching on 'BRD' will turn up thousands of results, all helpfully called BRD.doc. This is not all that helpful really (yet this is where faceted search adds value) -- but it is to be expected when no metadata is captured with the documents.
Unfortunately in my previous consulting experience, the answer to this is often to cry “we need a better search engine” and to spend lots of money, for no particular gain. In our case we are introducing a SharePoint based project management system which introduces document management and includes a 'Content Type' for each of the 50 plus templates, with a comprehensive metadata schema to improve both categorization of the documents, and increase findability.
The pragmatic and iterative approach has worked for us. Enterprise search -- or department search for our POC -- for the IT division is about to be moved to into production, for use by all IT staff. And while rolling out the initial features we continue to evaluate third party add-on's as alternatives to the CodePlex Faceted Search, and begin to study the new features and improvements to search in SharePoint 2010.
Of course once we have demonstrated the ROI for the IT division, the next logical step will be a full enterprise search implementation for all users. Done right -- through tested and refined increments -- this is where the real ROI begins to show.