Enterprise Search Sourcebook 2008

COMMENTARY The Importance of Dialogue There is a mistaken belief that the ideal search engine will provide nothing more than an elegantly simple single search box, into which the search term is entered in order to generate a list of relevant results. Let’s call it the Google aesthetic. While your results list may come back in a second, have you ever actually timed how long you spend on those results, wandering off on dead ends, getting sidetracked with ads, and soon forgetting what the original search was all about as you find something really interesting through serendipity rather than effective search technology? This is not a criticism of Google specifically but a reflection of the lack of critical assessment of the search process by specifiers and users. A substantial amount of research into the way in which people search has been undertaken in recent years. Still, I must return to the work of Marcia Bates in the late 1980s, in which she emphasized the role of dialogue in the development of effective search. The best analogy I can give is this: Imagine someone walking into Macy’s and asking the person at the information desk about the location of the birthday present section. The response will be that there is no such section and, knowing Macy’s, the store assistant will try to aid the customer in his quest—asking about the age and sex of the gift recipient, then almost certainly following up with further questions about his or her interests and the potential price range of the gift. A search engine must be able to undertake the same process, which means it must accommodate two basic types of query. The first one is when the users have a good idea of what they are looking for and have some specific search terms in mind, perhaps even a document that they can use as a starting point. At the other end of a wide spectrum, there are users looking for ideas and having no sense of what the search terms might be: “I’m looking for high tensile strength materials that are corrosion resistant to hot sea water.” That’s a very real question if you are in desalination engineering, but try putting that into a basic search query box. Relevant Relevance As we all know, getting fast results is a small part of the effective search formula. We are looking for the right answer. Two ways to enable this are ranking and relevance, which are often used as synonyms but refer to different processes. The results of a search can be ranked by many different parameters, such as date, file type, language, and other metadata elements. Recall is the number of relevant results as a percentage of the number of relevant documents in the collection. Web search engines give a reasonably good recall but poor precision, i.e., they find some relevant documents but they also return too many nonrelevant hits. An enterprise search engine must have a high recall to be admissible in most applications. Providing some indication of relevance in a search result set is an important usability factor. There is strong evidence to suggest that users do not want to go much past roughly 30 results, and they need some indication of whether there are still a substantial number of high relevance items to review in any given search set, which may prompt them to take a different approach in formulating their queries. Between Basic and Advanced While the dream of a simple search interface perpetuates, many users turn to advanced options to help guide them to more precise results. The usual approach to advanced search is to provide a long list of parameters to reduce the length of the results list. Invariably, the parameters make no sense as selection criteria. One of the few examples of advanced search applied intelligently can be found on the website of the Bank for International Settlements (www.bis.org). In general, most advanced search options are poorly developed and often represent a huge leap from the basic search approach. What is often needed is something in between. However, this often involves custom tuning an enterprise search interface to the specific needs of its user group. Categorization and Guided Navigation A search engine has to accomplish two goals in response to an initial query. The first is to provide a set of relevant results, and that is dependant upon technology, not the interface.The second is to provide as much guidance to the user as possible if the search does not come up with a good set of relevant references. One approach is to provide the user with autogenerated categories.Though in principle this can be helpful, I have seen many examples in which the list of categories is so long that there is then a need to scroll the results page, which is rarely a good idea. The website of Nature, the science magazine (www.nature.com), gets it just about right. Build a Dialogue With Users Search is often used under conditions of stress and urgency to locate important content quickly. It is a critical tool used at critical times for critical decision making. It’s time that organizations pay the kind of attention search usability merits and invest in the time and money needed to test and retest search interfaces occasionally. Even as relatively small an adjustment as adding a new set of documents could adversely affect usability. It’s also high time that search vendors demonstrate good practice on their websites and work with their customers to a much greater extent than at present to get the best out of the collection and the search engine. MARTIN WHITE (martin.white@intranetfocus.com) is the managing director of Intranet Focus, Ltd. and the author of Making Search Work (Information Today, Inc., 2006). WWW.ENTERPRISESEARCHCENTER.COM 15 http://www.bis.org http://www.nature.com http://WWW.ENTERPRISESEARCHCENTER.COM

Table of Contents Feed for the Digital Edition of Enterprise Search Sourcebook 2008

Enterprise Search Sourcebook 2008 Contents Editor’s Note Publisher’s Note Findings and Figures Why Enterprise Search Will Never Be Google-y Searching for Search Usability Your Users Are Talking to You What’s Your Search Story? Search Is Dead—Now What? Delivering on the Promise of Enterprise Search Taming Multiple Search Engines in Your Organization Enterprise Search: Trends for 2008 Enterprise Search Seen From the Inside Open Source Search: Elixir or Poison? Avoiding the Big Mistakes in Search Semantic Search Takes Root in the Enterprise E-Discovery Essentials: The Rules You Need to Know SharePoint Search: An Enterprise Contender? Integrating Security Into Your Enterprise Search Infrastructure Engineering a Better Search Infrastructure Letting End Users Ask the Questions, Stat! The Power of Knowledge Legal Research Using Enterprise Search: A Developer’s View From Treading Water to Full Steam Ahead Pulling Out All the Stops With Midas A Natural Search Solution An Incremental Approach to Improving Enterprise Search The Enterprise Search Sourcebook Showcase Directory Index to Advertisers and Companies Mentioned

For optimal viewing of this digital publication, please enable JavaScript and then refresh the page. If you would like to try to load the digital publication without using Flash Player detection, please click here.