Connectors, mavens and salesmen—Human networks and relationships in KM

In the example in Figure 2, an analyst is evaluating the text in a message and wants to find all the entities and relationships it contains. Link Factory processes the text into a table. The visualization component then uses the table to generate a Web-based link chart showing the entities, relationships, geographic locations, etc.

In addition to Attensity, a number of other relational analysis and metadata generation tools are available on the market including: Inxight’s Fact Finder, Clearforest Extraction Modules, SRA International’s NetOwl and Lockheed Martin’s Aerotext.

Search
As anyone who has used keyword-based search engines knows, it can be a frustrating experience. Finding the desired result can be difficult, particularly when the search is on an unfamiliar topic. Enter Endeca, a new form of search paradigm based on guided navigation, which is more like a business intelligence tool than a search engine.

Endeca—from the German word “entdecken,” which means to discover—uses metadata to generate a dynamic system of menus that continually change as you drill into the interface. To generate the menu, Endeca preprocesses source documents to extract entity and relationship information prior to its main indexing activity. In fact, many implementations that deploy Endeca actually preprocess all source data in an external process using a variety of metadata production tools such as entity extraction, categorization,
relational extraction, geotagging, etc.

Endeca is well suited to making use of metadata and other structured data sources, and can ingest information from relational databases and XML repositories, among others. The more metadata that can be made available to the Endeca indexing engine, the better the results will be for the user’s guided navigation information discovery process.

Endeca’s indexing engine is more than just a search engine. It is an analytics tool that is designed to allow the user to search on multiple information facets or dimensions. The core analytics system is designed to find the links and relationships between dimensions, allowing people to discover new facts and perform in-depth analyses that would not otherwise be possible.

In one case study, Endeca installed a solution with an oil and gas drilling company. That company was overpaying millions of dollars a year in replacement parts for equipment and machinery for oil drilling platforms, because it was unable to correlate the relationship between its enterprise resource planning, finance and accounting, and parts inventory systems. Because the oil drilling company couldn’t readily see what it had, it ended up ordering duplicate replacement parts for pumps, valves and other machinery components, rather than using the inventory it had or improvising with similar parts.

Endeca saved the company millions in a matter of months, ultimately giving it invaluable insights into the critical information and relationships buried in its various information systems. The technology allowed the company to see part numbers, the attributes of the part, available inventory and other similar parts and swapping tradeoffs—all in one secure online location. More importantly, Endeca was able to help the company find the right information the first time, every time, raising productivity and performance.

Endeca also provides support for hundreds of Fortune 100 retailers on the Web including such stores as Home Depot, Barnes & Noble , Lowe’s, Wal-Mart, Circuit City and many others. One of my favorite retail sites is Nike, which has been developed using a Flash front end that talks to the Endeca guided navigation system. Figure 3 shows the Nike store Web site with a product filter on the left side.

Endeca’s guided navigation interface allows the user to quickly narrow the search in as few clicks as possible. For example, when searching for men’s, apparel, shorts, cycling, I was able to find specifically what I wanted in four clicks and narrow my search from almost 3,000 items to just five. Along the way, I could visually explore a wide range of available Nike products of interest to me.

That kind of core processing allows the user to quickly drill down to a goal while also allowing further discovery. The approach is central to Endeca’s focus on information discovery. Figure 4 (page 22, KMWorld, Vol 16, #5) shows how I narrowed my search using the guided navigation menus on the Nike store Web site.

If you want to see more examples of Endeca-powered applications, go to the Endeca Web site.

Although no other search company appears to have come close to duplicating Endeca’s unique business intelligence and search qualities, other search companies are making efforts to compete. Companies that continue to develop advanced search technologies include Google, FAST, Autonomy and IBM.

We’ve seen that understanding the entity relationships in networks can dramatically increase productivity. There is, however, one question remaining: Are you a connector, maven or salesman? Try this: Check out Wal-Mart’s Endeca-powered e-commerce site for sales and new products, then call all your friends and family and tell them about it. Voilà! The power of relational software has the ability to make you a connector, maven and salesman, all at the same time. Now that’s what I call productivity.