Enterprise Search Pitfalls: Believing in Magic

Posted on December 1, 2011 by in Latest Articles

Series Introduction

Regardless how many satisfied enterprise search customers we meet, it seems there are always “irritants”.   Many enterprise search customers have shared with us their frustrations towards their current search engine solution.  For some, frustration grows with their needs.  For others, it seems as though no search engine works quite like they want it to.  It turns out many irritants come from recurrent mistakes or misconceptions, which we have witnessed across several organizations.  Over the years, we have been privileged enough to gather a lot of excellent feedback.  We are introducing this series to bring to light some of the mistakes encountered through the years, hoping it will save you or your organization from doing the same.  Each article in this series will describe in detail an enterprise search pitfall, followed by advice on how to prevent it.

Believing in Magic

Disclaimer: Whenever Google is referenced in this article,  it is as the web search engine that has become an integral part of many people’s live.  When it comes to the enterprise world, Google also offers enterprise-scale products.  This article does not explicitly refer to Google enterprise search products, but talks about enterprise search products in general terms (regardless of search vendor).

We all want to believe in magic.   Some say it keeps us young at heart.  On the other end, I believe not all magic is good for you.  Here is one form of magic I don’t recommend you believe in: enterprise search magic.

There is an unfortunate belief among several enterprise search customers that their biggest challenge is going to be picking the right search vendor.  They believe that somewhere out there, there has to be a search vendor that has THE solution just for them.  After that, they basically should just plug it in and it works. This is particularly true for neophyte customers.  When dealing with search magic believers, everything becomes very hard to justify: hiring search experts, project delays, change requests, implementation defects, time it takes to perform certain updates, etc.   In worst case scenarios, some organizations are convinced to switch enterprise search vendors after a few years only. This cycle is often repeated until they find the magic they are looking for, which of course, is never to be found.  Either they eventually realize there is no such magic and settle on doing a proper implementation with the current search product they have, or it ends up costing them a fortune to start over and over again.

One can blame it on enterprise search vendors claiming their search product can read your mind while brewing your coffee.  I believe it is more profound than this.  I tend to relate this to the “Google effect”.   Google won the battle of web searches for a few reasons, but one certainly is by making search seem effortless and appear to be a trivial piece of technology.  Enterprise search experts very well know there are sometimes realms of differences between general-purpose web searches, and “enterprise” search, where requirements can be very specific, and sometimes conflict from one organization to another.  Many enterprise search customers don’t realize these differences at first.   When thinking about search, most people naturally think of web searches like Google, how it “just works” for them, and how easy search is.  While Google web search deserves its success, these people have the same concepts in mind (“just works”, “easy”) when they approach the enterprise search product they just purchased.  It does not always resonate that to make a piece of complex technology appear simple to use, it can take a great deal of effort.   After all, how can something magical take months to put in place?

You can easily detect search magic believers when at every hurdle encountered during enterprise search development, they are quick to point out: “Why don’t we just go with Google?”.  They are usually making a reference to their experience with google.com, the web search engine.  This can sometimes be a legitimate question.  Going with a web-search or even a turn-key solution can of course be a good choice to many customers. On the other hand, if by leaning towards a turn-key solution you have to sacrifice several of your enterprise requirements, you may need to think again.  What is really meant here goes along the lines of: “Why do we have all of these challenges and why are we investing so much effort when it should just work?”

Prevention

Before being buried by enterprise search, and spending half or your days trying to justify your enterprise search vendor decision, try to educate key stakeholders on the complexities of enterprise search.   Explain the difference between a “web search” and an “enterprise search”, along with why your organization has chosen an enterprise search solution.    When defining your requirements, be sure to highlight what falls under the enterprise search umbrella (indexing from various repository sources, document security, etc).  Point out all of the custom solutions that will need to be put in place to meet those requirements.  In other words, dispel the magic.    Next time you hear “Why don’t we just go with Google?”, try to find out if they are really referring to the Google web search engine and if so, be prepared to list which of your requirements cannot be addressed by a web-based or even turn-key solution, and why these are important to your organization despite the extra time needed to implement.   Watch out for colleagues wanting to drop or adapt requirements to make your organizational needs fit a specific (magic) vendor – this should be the other way around. Finally, remind people that enterprise search, no matter how great it is, is not magical.

Pascal Essiembre has been a successful Enterprise Application Developer for several years before founding Norconex in 2007 and remaining its president to this day. Pascal has been responsible for several successful Norconex enterprise search projects across North America. Pascal is also heading the Product Division of Norconex and leading Norconex Open-Source initiatives.


Comments