Menu
  • InShare PlusOne
  • Finding the Value of Search

    The inputs and outputs of modern businesses are often no longer physical products; they are ideas, information and content. Even when a company produces physical goods, it's almost impossible to operate competitively without the right information for decision making. Fast, efficient access to digital information has become an important business need, but the content within many organizations is simply too large, too complex and too volatile to browse manually, absorb, interpret and determine relevance. This information rich environment has made tools like search critical.

    Finding The Value of Search

    The purpose of search is the same as it has always been - finding information. Search bridges the gap between people and content by allowing them to find the information they need and discover related information that they might not have known existed. Although this may sound simple, it's not. A successful search implementation does not just throw content at users and let them figure out the rest. Maximizing the value of search requires:

    • locating the correct content
    • locating the correct version
    • ensuring the content is accessible to the correct users
    • providing clear and accurate rankings to guide the user
    • consistent and contextually sensitive presentation of the information across the organization

    The impacts of missing or inadequate search are substantial. A recent IDC study estimated that the average information worker spends 9 hours a week searching for and gathering information. Half of that time, 3.7 hours a week, is used searching for, but not finding the right content. An additional 3.7 hours is used reformatting data from multiple sources, and 2.5 hours per week are lost recreating information that can't be located. That's a cost of $28,501.82 per year per information worker with a $75,000.00 salary; the cost of inefficient search escalates quickly.

    Internet Search Does Not Equal Site Search

    For many users, Google is search. The brand name has become a verb.

    Google Defined

    This "Googlization" (Yes, I did make up that word.) of search, has given users very concrete ideas about how search should look, feel and work. They want all search to work like Google, but site search is not internet search.


    Internet Search

    Site Search

    Content scope

    Public documents online

    Single site (may include sub-sites)

    Format

    Unstructured

    Structured (e.g. CMS database) and unstructured

    Context

    Generic

    Domain specific

    Taxonomy/Tagging/Categorization

    Generic

    Domain specific

    Security/Access Control

    No authorization required/public

    Identity or role-based access control

    Relevancy

    Keyword and link-based

    Keyword, tags, popularity

    Desired Result

    Highest popularity

    Highest Quality

    Additional Variations

    1. No central authority
    2. Uncontrolled publishing model
    3. Publishers have a high focus on discoverability (SEO)
    4. Limited meta-data
     
    1. Users and authors share a common understanding
    2. Documents publishing may involve approval workflow
    3. No SEO focus
    4. Significant content relationships, context and meta-data
     

    Although the scale of site search is smaller, the security, more sophisticated measures of relevance, information context and varied formats, make it more complicated than its Internet-based sibling. Purchasing a search platform and hoping for the best is recipe for failure - no matter what the guys at Google's enterprise sales division tell you.

    Finding the Right Approach for Search

    A successful search implementation does not begin with technology and tools; it begins with a strategy. Forrester defines a seven-step plan for developing a search strategy.

    Define objectives. Before implementing search, you should understand exactly what you are attempting to accomplish. Is the intent for search to replace navigation or is it a tool to speed information access. Key questions: Who is searching? What are they looking for? What will they do with the content once it's located?

    Define Scope. Search projects can quickly become expensive and complex. Defining a concrete scope will help you avoid the dreaded project death march that sucks in resources, exceeds budgets and tramples schedules. Key questions: Is your project for a single repository or is it an enterprise strategy? Does your search include all content or only certain categories?

    Evaluate and select content enrichment methods. Your site may (hopefully) be built on a content management platform. The underlying content repository can be a rich source of metadata that enhances content findability. Will your search use this information? Is additional metadata required? If additional metadata is required, who will maintain it? Is social tagging acceptable?

    Define requirements, success criteria and identify vendors/products. Key questions: What requirements must be met to make the project a success? How will you measure success? Don't fall into the trap of only collecting qualitative measures of success; quantititative measures can help you clarify the most efficient design, navigation strategy and even what content is included. There are several key metrics you should consider monitoring and setting targets for such as:

    • Zero hits - What percentage searches yielded no results? Zero hits could mean misspellings, a need for synonyms, a lack of content, multilingual issues or a number of other scenarios.
    • Low query click through rate - What percentage of searches resulted in the user not clicking any results? This could be a sign of low quality results.
    • Search vs. overall site usage
    • Search vs. navigation/browsing
    • Total distinct queries - Can be useful for determining if query result caching, which can improve search performance
    • Top queries - Can help clarify what information users are seeking. It can also be useful for implementing features like "best bets" that can speed information retrieval.

    Define a search taxonomy - In addition to understand the content and algorithms, it is important to understand how different search types will be organized and presented to the user. How does this impact user experience? Does relevance change by search type?

    Define the user experience - The search user experience is almost as important as the search results returned. You should engage end users early in the process to determine the best way to display information so that it will not be confusing or create ambiguity.

    Implement, monitor and improve - Once you understand what you are trying to accomplish, you will need to actually implement and tune the search platform. However, your job isn't done when you install a search tool. You must continue to monitor and refine your implementation.

    Although search technology has existed for decades, it's still challenging to get right. The misperception that search is a "solved problem" hasn't helped the matter. However, understanding your user needs, your content, your goals, and your technology options can dramatically increase the likelihood that your efforts towards a search strategy will be successful.

  • InShare PlusOne
  • James Coleman
    By James Coleman
    Tuesday, March 20, 2012
  • comments powered by Disqus