Found It – Embracing Enterprise Search

By Bert Sandie posted 07-19-2010 23:00


Companies from small start-ups to the mega corporations all encounter the day when they need to find critical internal information at the drop of a hat. This has led to the rise of companies investing in enterprise search solutions to help them more quickly and accurately find the information their employees need to be productive and drive a competitive business advantage.

There are two main enterprise solutions that companies begin to look at when shopping which are Google and Microsoft. Both solutions have their pros and cons; and readers of this article can explore which of these two solutions best serves their needs.

Let’s turn our attention to “What do we want out of an enterprise search solution?” The most common responses from users are:

  1. Speed – results that are returned fast – think about how fast Google and Bing return results; corporate users demand that same speed
  2. Relevancy – high relevancy and accuracy of results based on the search terms provided
  3. Data Source Indexing – corporate users expect as new data sources are added inside the company these are added auto-magically to the search indexing

What users of search really are saying is “I want to find what I was looking for now!”

Search in the past five or more years has become extremely more complicated in the enterprise for a number of reasons. Outlined below are a few of the items that have driven the complexity curve upwards.

  • Amount of Data - It was not long ago we talked in MBs (Mega), then GBs (Giga), now we converse in TBs (Tera) and PBs (Peta) when conversing about storage. To give some perspective on the data explosion, YouTube has 24 hours of video content uploaded every minute by users – based on this it is easy to see how a TB is quickly used up.
  • Data Content Types - Search needs to support the ability to search a wide range of content types include:
    • Text - documents, spreadsheets, etc.
    • People – user profile fields
    • Images/Pictures
    • Videos – many companies have an internal YouTube-like solution
    • Digital/Binary assets – including 3D models, animations, BLOB data, etc.
    • Internal sites and wikis – this can include the corporate portal, wikis, SharePoint sites, etc.
  • Speed – users expect an almost instantaneous response once they have entered in their search criteria and pressed the return key
  • Security – the data must be secure and limited to inside the firewall due the amount of confidential, private and intellectual property that is available to be found; login permissions are required (i.e., single sign-on) to access the data
  • Access Control – different sets of data inside a company need to have access control applied to it as data may restricted by role, location, seniority, etc.

Here are some thoughts on how search can be constructed to help users be most effective:

  • Tags – tag content through the use of folksonomy and/or taxonomy; the search engine can then search and filter based on this meta data
  • Faceting – provide high-level filters to segment images, videos, people, intranet sites, etc.  A second level of filtering may be provided. For example, for images there may be size filter such as small (0 to 300 KB), medium (300 to 750 KB) and large (>750 KB); or by image type (i.e., gif, bmp, jpeg, etc.)
  • Sorting – allow users to select different ways to view the data via sorting:
    • Most recently contributed
    • Most viewed all-time, past 90 days, past 30 days, past 7 days
    • Relevance

Based on my own experiences and the collective wisdom of a number of E2.0 practitioners, listed below are a few recommended best practices to help drive to a better search solution.

  • Invest in at least one individual to staff your search team; it takes time and effort to keep search running efficiently and effectively in medium/large size organization.
  • Regularly ask/poll employees to gain a deeper understanding of what types of searches they are performing and if these are successful; this may result in modifications to indexing, relevancy algorithms, data sources, etc.
  • Provide faceting (filtering and sorting) mechanisms that help lead to query refinements; look at the directions Google has recently provided with their left panel (see example images below)
  • Target a set of key business problems that search can help with and then work to make the data available and easily findable as part of your roll-out strategy
  • User interfaces need to use paradigms people are already familiar with but which can be customized or extended for particular business needs where necessary.


Example 1 - Google Search

Example 2 - Google Search Faceted




Happy searching!

#Search #enterprisesearch #Facetedsearch #search #searchengine