We pinned Microsoft search server against Google search appliance to see which one would come out on top. Here's what we discovered.

Earlier this week, I posted a blog comparing the approaches the Google search appliance and Microsoft search technologies take to optimizing the short tail of search. Today's post looks at how you can address the mid-tail of searches by tilting the landscape of the search results.

The Google search appliance (GSA)

Google backed away from their stance that tuning search in the enterprise can set off a tornado of search results chaos after they introduced a "biasing" control in 2007. This feature lets you "bias" the GSA's baked-in algorithm to rank documents of specific kinds or from specific sources more favourably or less favourably. Recently, Google introduced the capacity to bias documents by date of indexing, as well. This simple interface actually provides substantial capacity for modifying results so they reflect your organization's business rules for content creation and storage.

Microsoft search server express, etc.

Microsoft provides an interesting and potentially more powerful approach to the same challenge. Non-technical personnel can identify "authoritative" pages or documents, which is to say the importance of documents "near" these sources of authority are increased. In fact, with this function, you can define pages as having primary, secondary, tertiary or negative authority. 

What counts as authoritative? In our search tuning engagements we look to these sources of authority: 

  • Index pages (which are considered to be authoritative out of the box) 
  • Entry or start pages (as indicated by web analytics) 
  • Pages or content authored by executives 
  • Blogs by in-house thought leaders 
  • The shared social bookmarks of in-house thought leaders (particularly if you have an enterprise 2.0 strategy in place)

Warning

Both of these approaches change the topography of the search results landscape frequently in unpredictable ways. They share a need for careful experimentation. Frequently, modifying these settings to solve one search deficiency will create problems for a whole other range of issues.

Next steps

In a few days, I’ll post the final blog in this series. In the meantime, feel free to download our whitepaper on Google Search Appliance Best Practices - it addresses many of the concepts covered here in additional detail.

comments powered by Disqus

Wondering how to get started?

Ask an analyst

We can share our
real world experiences