Although it is incredibly tempting to lump search into one neat category, we must acknowledge that enterprise and Web search are profoundly different. Ten years ago, search technologies that were originally developed for the Web were applied to the enterprise - yet it has become harder and harder for workers to find and apply the critical knowledge they need to do their jobs. Numerous studies show the average knowledge worker spends at least 15 percent of his or her time searching for the information they need to be productive.1 Why? Although search technologies are born from the same intention, the objective of a person searching an enterprise site is much different in terms of intent and desired outcome than someone using Google to discover information like a local restaurant or movie listing. In addition, enterprise sites have very different requirements to meet enterprise-specific needs that include workflow, security, integration with existing systems and content moderation. When you think about the purpose (and urgency) of someone searching on a wireless companys Web site because his or her phone wont hang up, versus using Google to compile a list of places to visit in Paris, it makes sense that the search tools should be diverse enough to meet a variety of user needs. The industry is catching up to this, and I expect the divergence of enterprise and Web search to continue to the point where the two solutions are classified into entirely separate categories. Lets examine this divide more closely.
Please State Your Intent
One of the main reasons that search offerings on enterprise sites should be designed differently from the Web is because businesses are dealing with a dramatically different searcher with a very specific set of needs. On a corporate site, visitors are usually looking for far more specific information, and this becomes even more specific when they enter the support section. Whether they are on a Web site for a major car dealer trying to learn more about the car that they are considering purchasing, or are trying to learn how to set up their new HDTV, the visitors are looking for targeted answers to specific questions that allow them to take action. This makes it significantly easier to successfully derive user intent - what the user is trying to accomplish based on what they ask and what they access. In fact, on enterprise sites, the most frequent 20 intents typically cover more than 40 percent of all inquiries, according to internal research conducted by InQuira. This makes the pool of options, and therefore the knowledge base, much smaller and easier to manage, which is good, because the repercussions of repeat cases are long lists of relatively useless information - Web 2.0 tools have made this even more salient (more on that later).
This leads us to the actual type of search. The enterprise benefits from knowing its audience and can narrow the search results by anticipating the users intent and eliminating subject matter that is not relevant to its business. Most enterprise sites use a specific industry vocabulary. For example, you don't have to recognize names of U.S. presidents or hybrid minivans for a typical banking site because they are not relevant to the industry. The sheer scope of information on the Web, the wide variety of contexts and concepts, and the different intentions of use all combine to make this need extremely complex.
The Googlization of the World
Different intents trickle into the way that people type their inquiries. When Merriam-Webster added the verb google to its 11th edition, it bestowed a true cultural and business phenomenon status. Googlization is more than just industry chatter; its who we are, and its how we behave. That simple one-line search box has become our window to the Web. Elegant in its simplicity and powerful in its reach, it has had as much influence on how we interact online as any other technology innovation. Because it is keyword based, most consumers know that a Google search is best made with simple words because theyve learned that writing a full question like Where is the closest movie theater? will generate less accurate results than a simpler inquiry such as movie theater plus the ZIP code.
Google is very effective in narrowing the broad universe of the Web to the right haystack, but it isnt as effective for finding the needle within it, in part because every piece of hay looks the same to a keyword-based statistical search engine. Findability and information management are much more entwined in corporate environments. The search experience must return actionable information in context. Because enterprise sites are dealing with very specific information and intents, they are in the unique position to offer semantic or natural language search to produce the most accurate results. If we invest in understanding the unique vocabulary that is used in the context of an enterprise-specific search, we have taken another step forward in making natural language processing genuinely applicable across that particular industry vertical.
Its about Quality and Quantity
Another factor that dramatically separates these two paradigms is content creation and quality. When searching on the Web, you often dont know the origin of the content through which you are culling. The proliferation of Web 2.0 tools has made this especially true. A search for a particular product could generate a ranting diatribe on a popular blog criticizing the item you are reviewing. While it could be helpful in your selection, you could have also stumbled upon a disgruntled former employee or a crafty competitor. Because enterprise Web sites are dealing with more specific information and direct interactions with their customers, they tend to more closely monitor content creation and the formation of a knowledge base from which they are going to offer search results. By populating its database with quality information that will return results that a user can act upon, they will be keeping the user on their site for longer and preventing her from looking for a competitors products.
Lets look at that stat about knowledge workers spending 15 to 20 percent of their time searching for the information they need to be productive. Why is this? Because critical enterprise knowledge tends to exist in an unstructured form that makes it hard to harvest without technologies like natural language search solutions, which can understand the intent behind the users search and can index the searchable information sources for their semantic meaning.
Keyword-based search tools often fail in an enterprise environment because they are designed for the sheer magnitude of content that exists on the Web. On the Web, there are so many documents available that virtually any combination of keywords will match a set of indexed documents. On the enterprise, there is no such luxury. Because there is much less content available on an enterprise site, a better understanding of the intent and the content itself is needed to bring users the information that they need.
Another huge issue that Web search engines have had to deal with is that the relationship between search and content is more or less antagonistic. People make Web pages that bully their way to the top of the search results by employing a whole grab-bag of tricks (search engine optimization, advertising, etc.), one of which is not a content-quality filter. With enterprise search, on the other hand, the people who make the content are on our side. They want the right content to be shown in the right context, to the right user. They understand what the content is trying to say, and they are willing to revise content or write more if that turns out to be useful. A virtuous cycle that includes automated analysis of content needs as expressed through search queries, which in turn triggers a workflow process for content changes, is a big advance, and one with effects we will be witnessing in the next several years.
I Need Help, and I Need it Now
The final reason that enterprise and Web search should be on different development paths is because they have different business models. Web search is primarily in the business of serving ads to drive people to the correct (and more profitable) Web sites. Enterprise search aims to enable transactions and provide customer service. There is a growing trend toward Web self-help, especially among people under 30.With so many Web-based products and services, this is often the only customer support offering available. The Web 2.0 era is raising the bar on our expectations for quick information access, increased service levels and an improved customer experience. People want accurate, immediate responses - not pages of largely extraneous information - when looking for support or product answers online. Web search engines are designed to sift through and recall an enormous number of documents quickly. In the enterprise, while recall is vital, the depth of knowledge and understanding of queries makes all the difference. By populating your knowledge base with useful information, the search engine will serve its purpose.
But what about Web sites that are not meant for the consumer? Many enterprise Web sites serve as a resource for employees, partners or B2B customers. In this day and age, when everything is available on the Web, any enterprise that is founded on intellectual property as its primary selling point must have a useful Web site that its users can refer to so they can effectively do their jobs. A software company, for example, may have an intranet site for its sales team that contains all of the information about its products and services. While having all of this information in one place is helpful to the sales team, it is not useful unless it is properly managed and organized, and there is an effective way for them to search it. If this resource is not user-friendly, employees will find their own ways to organize their information, probably by saving information on their PCs and email, or *gasp* writing it down in a notebook, thus defeating the purpose of the collaborative nature of an intranet. A company can significantly invest in building a foundation for its employees to help them do their job, but the bottom line is that if they cant easily find information, they will not use it. Enterprise sites in this category must go the extra mile to include a search mechanism that delivers results, and this type of search tool cannot be retrofitted from larger Web search technologies.
So Now What?
So, how do we ensure that search engines on our Web sites are generating information that our customers need? First of all, we must recognize that the search needs of enterprise Web sites are much different than Web search, and we must move away from using the same old Google model in the enterprise. Web search serves a wonderful purpose as a catch-all of information, a resource to discover new products, services and destinations. But when people are looking for specific answers from a company or their own employer, they want the right answer immediately. Yes, search tools are a big part of making that happen, but companies must address the information that supports that search, otherwise they will not be successful. Companies must make knowledge and content management across the enterprise a priority. If you can harvest and apply the insight gleaned from customer interactions, whether it be through searches, discussion forums or live interactions, you stand an excellent chance of providing even better service to your customers. Updating product information, editing support documents and flagging issues raised in discussion forums are activities that might take a little work up front, but they will pay off tenfold. Finally, we cannot forget that the goal of enterprises is to please the customer, and we must develop technologies that hone in on the details that our customers need, rather than offering them too much information that leaves them lost and frustrated.
- Susan Feldman. " The High Cost of Not Finding Information." KMWorld.com, March 2004.
Register or login for access to this item and much more
All Information Management content is archived after seven days.
Community members receive:
- All recent and archived articles
- Conference offers and updates
- A full menu of enewsletter options
- Web seminars, white papers, ebooks
Already have an account? Log In
Don't have an account? Register for Free Unlimited Access