I’ve been working with a client to enhance their SharePoint intranet system in order to improve user adoption.  One of the key selling points of SharePoint is providing greater levels of accessibility to content.  The primary method of making content accessible in SharePoint is Search and since Microsoft acquired FAST Search in 2008, the search capabilities in SharePoint have been going from strength to strength.

However, search in SharePoint is often the area that is given least attention whilst a solution is being implemented.  After all, it’s touted as being the Google (or should that be Bing) of enterprise content, and web searches just work!  Unfortunately, a lot of work goes into internet search engines to provide the correct results.  Search result ranking calculations are constantly being adjusted; indexing algorithms tweaked to ensure all relevant information is indexed (and irrelevant information is ignored); and context is key – have you ever wondered why Google or Bing stores your search history?

With this in mind, do you really expect to be able to point SharePoint search at your enterprise content and expect it to be able to provide better results than if you simply use Windows Explorer to search a file share?  Without customising SharePoint search, that’s essentially all you’re getting.  Yes, it will do some clear work around distance between words in a document, and giving a higher weighting to words in document titles and most accessed files but in the grand scheme of things that’s a marginal improvement.

Microsoft SharePoint search components architecture

SharePoint search components architecture

 

However, the capabilities of SharePoint search customisation extend way beyond the out-of-the-box offering.

  • Authoritative Pages
    • Ensure users get the most relevant information by telling SharePoint which pages in your SharePoint farm contain the most valuable information.  This helps result ranking by calculating the click-distance of results from these pages.
    • Additionally, you can demote non-authoritative pages, e.g. archives of historic documents.
  • Schema
    • This is essential if you are creating custom metadata fields in SharePoint.  Until you update the Search Schema, SharePoint search indexing will ignore your custom fields rendering them all but useless.
    • Changes to the Search Schema also require the index to be rebuilt.  Consideration must be given to the impact on users whenever the search index is rebuilt.
  • Dictionaries
    • The Managed Metadata Service is used to provide dictionaries of words for Company names to be identified and given special consideration during indexing and ranking.  This can be pre-populated with information from other line of business applications, given SharePoint a head start of names the business is interested in.
    • There are also dictionaries for storing common spelling mistakes to allow the search results pages to offer corrections.  It is also possible to exclude words from spelling corrections.  For example of a business uses an acronym that happens to match a common spelling mistake.
  • Query Rules
    • These allow adding some intelligence to users queries.  For example, if a user searches for “business development presentation” a query rule can be specified for the term “presentation” that gives more significance to PowerPoint files.  A search for “absence request” could promote the absence request form page in the intranet site.
  • Search Ranking Models
    • Customising the Search Ranking Models is not for the faint-hearted and should be done with great care.  SharePoint comes with a number of built-in ranking models that will normally provide the results required.
    • SharePoint has a default ranking model that will be used out-of-the-box for all searches.  However, it is possible to switch to alternative ranking model, use Query Rules to select a ranking based on the users query, and create custom ranking models.
    • When creating a custom ranking model, it is recommended to start with an out-of-the-box model and adapt it to the business requirements.  Use Query Rules to try the different system ranking models to select the one to use as a basis for the custom model.

With all the search customisation options available it is important to remember that enterprise search begins and ends with the users.  They are the ones that are creating the content and trying to retrieve it.  There should be an on-going program of education & training for users; monitoring search usage reports; and tuning the search customisations to ensure the results remain relevant for the users.