|Return to MWJ Home
||December 2010, Issue 82|
Semantics Here and Now
By Seth Grimes, Analytics Strategist, Alta Plana Corporation
Semantic technology is hot. It's the brains behind smarter search, a means of taming the information explosion by finding meaning in online and enterprise information sources. Semantics is a path to making information findable, to automating document processing, to turning text into data to facilitate web mining, customer engagement, and social-media analytics. Yet not every technology under the semantics umbrella is here and now, delivering enterprise value. The Semantic Web—a vision of agent-bots that'll automatically book your travel itinerary and the like—remains more dream than reality.
Semantics and Enterprise Business Challenges
Semantic capabilities are already creeping into web search engines and enterprise information-access tools alike. See for yourself—although you probably have already if you've ever sent a query like "map chicago" to Google or Bing. You get back a map of Chicago, Illinois, and not just a list of documents that contain the two words that make up your search. The search engines recognize "map chicago" as an information request given the co-occurrence of two words associated with geography, "map" and a particular geographic area. Bing will also provide a list of related searches, for instance, and Google will allow you to restrict your search to documents published in a particular time range by using descriptive information called "metadata" that is associated with indexed pages.
These engines are great, but they're not much help with information that resides in your organization's own databases and operational systems, whether web-facing or accessible only to internal users. The ranking algorithms aren't designed for enterprise priorities, the crawlers don't reach into restricted-access systems and the interfaces don't suit business workflows well. General-purpose tools aren't top performers for focused business tasks such as supporting online storefronts and searching media sites.
Businesses have highly individualized metadata—ways of describing their products and services—and operate in specialized contexts. Search Amazon.com for "chicago" and rather than a map, you'll see '70s and '80s record albums and copies of the 2003 movie starring Renée Zellweger, Catherine Zeta-Jones and Richard Gere, with the option of narrowing down results according to departments including Music, Books, and Movies & TV, an example of "faceted search."
You don't have to be Amazon.com to tap these capabilities. Information-access tools from providers such as OpenText bring the benefits of semantic computing to enterprise business challenges.
Semantic Search and Navigation Are Well on Their Way
I've been helping OpenText with market strategy surrounding semantic navigation and other capabilities from Nstein, a company OpenText acquired earlier this year to complement its enterprise content management (ECM) solutions. I've known Nstein for five years. The OpenText acquisition is an affirmation of their work in text mining, sentiment analysis and semantic search, which provide tools that enable next-generation publishing, commerce, web applications and data analysis. The market now enjoys a huge opportunity to leverage semantic technology to meet here and now business challenges.
Start with making information findable, for instance, at an online commerce or media site. The goal is to direct website visitors to the product information or content they're seeking which, of course, will help them carry out transactions and create a more satisfying website experience. In this case, semantic navigation means, for instance, providing search facets, which sort information (products, content, etc.) into high-level categories that may reflect a descriptive category (toys, clothes, housewares, books) or that may reflect a property (price tiers, age ranges, materials) or both.
In a media scenario, we may have facets for topics, people, organizations and places, whatever matches the need at hand. Really there are a broad set of possibilities when the interface is powered by assets such as taxonomies, which help publishers categorize information, and content analytics, which is software that identifies names, places, events and facts, and also opinions and attitudes, in text and rich-media sources.
You can see these capabilities in action at sites such as Cyberpresse.ca and OpenText's own site—the company is "eating its own dog food," as the saying goes—and elsewhere on the Web.
Changing the Way You (and Your Customers) Access Information
Interest and uptake will continue to grow as we understand the possibilities offered by semantic search and navigation. Web and enterprise users will become accustomed to interacting with and exploring information, and there will be no going back to plain-old keyword search and low-value hit lists of search results.
Return to MWJ Home
Seth Grimes is an analytics strategist with Washington DC based Alta Plana Corporation, founding chair of the Text Analytics Summit and the Sentiment Analysis Symposium, and contributing editor at TechWeb's Intelligent Enterprise. He consults, writes, and speaks on business intelligence, data management and analysis systems, text mining, visualization and related topics. Follow him on Twitter http://twitter.com/sethgrimes
Marketing Watchdog Journal is a monthly newsletter from Bulldog Solutions, an online marketing agency that changes the way BtoB companies define demand-generation strategy, engage prospects and convert leads to customers. We welcome your feedback on this newsletter's content and design, and encourage you to share your ideas for topics you would like us to cover in future issues. Please send your comments or questions about Bulldog Solutions to Amy Bills, director of demand generation.