Friday, August 29, 2008

Search Strategies & Law Firm "Information Gravitation"--August 27 at ILTA

Title, Session Link and Slides: Search as Strategy - Creating 'Information Gravitation' in the Firm


Enterprise search is all the rage but much of the talk is about search as an “application.” Can search be more than just another entry on the shopping list of applications firms buy? Can it serve as a foundation element in an overall information strategy? Some firms are beginning to talk about creating a kind of “information gravitation” both inside the firm and across the firewall to clients. With such gravitation in place, the right information flows to where it is needed, when it is needed and often without need for discrete searches. Learn the role such information gravitation play in overall information strategy, the impact such a goal has on choices of technology, security and privacy concerns, deployment timeframes and on levels of investment. If that is the vision, how do you enlist the whole firm in it?


Derek Schueren — Recommind
Felicity Badcock — Mallesons Stephen Jaques

My Take:

Derek of Recommind is one of the more visionary of the vendor representatives out there. And Mallesons is by reputation one of the more innovative firms Down Under, having won the Innovaction award for their PeopleFinder tool. I was looking forward to this session and it did not disappoint.

Derek started with a wonderful quote from Aristotle:

"Suppose that every tool we had could perform its task, either at our bidding or itself perceiving the need…that shuttles in a loom could fly to and fro…and a plucker play a lyre of its own accord.”

(A plucker here is not a person undressing a chicken but presumably the plectrum or "guitar pick" of the ancient world).

Relevancy is the ability to retrieve material that satisfies the need of the user. (I usually think of relevancy in terms of how search result ranking and how well the results compare with the search terms, but this definition appropriately puts meeting business needs over some abstract fitness).

Derek discussed how Google treats relevancy as essentially a popularity contest. The most popular web pages have more incoming links and few outgoing links. This works pretty well on the web, but, because there is no comparable source of relevancy information inside the enterprise, consumers don't have the same experience with search inside the enterprise.

One potential advantage for enterprises, however, is that Google does not take into account who is doing the searching. Search tools might be able to show different types of results for partners, associates, or professional staff.

Derek compared web search and enterprise search.


The Web contains a huge store of simple content. Internet content typically lacks much context beyond what is linked within its pages. All information available is completley public. Who authored the content is generally irrelevant.

Relevancy is based on key word match over layered with popularity score


Enterprises by comparison have small content stores but much complexity. The lack of link structures mean relevancy needs to be computed in a different way. But content in enterprises does relate to other information inside and outside of the information (think for example of matter numbers). The authorship of content really does matter to content validity or utility, (at least to someone like a senior litigation partner looking at a first year associate's research memo).

Enterprise search tools must respect security.

Metadata in the enterprise is much richer. You can provide more context for the document in the enterprise. You can filter down on information post-query, and relevancy can extend beyond the primary object to related content.

Integrating Search

Concept searching allows you to find documents that may not have all the key words, but that are correlated (related) to the search terms.

Content linking will be key. You might want to see a lot more than just contact information for a judge, such as pleadings for matters before the judge, people who have appeared before the judge, and so forth. The Document Management System (DMS) has great content but it doesn’t have everything. Matter information isn’t in the DMS. Enterprise search needs to blend content from diverse locations. Search is an information layer rather than just a box that provides results.

An example of search as information layer is providing information about citations. If you filter and add in the statutes cited, and also possibly some other data (such as case validity) from another source, you get a much richer user experience. (see these posts on Lexis' recent work on adding citation information to search results, and this Caselines post from last year on Recommind's work with West KM on that front.)

Collaborative Filtering

How does Amazon make it easy for me to find what I want?

Amazon knows what I’ve looked for and also what people like me have looked for and found in the past.

Content Linking

Content linking will really take advantage of the information available in the enterprise.

Felicity Badcock / Mallesons

Felicity, a former practicing lawyer like me, has been working in the legal technology for twelve years, including five at Clifford Chance in London.

She addressed information gravitation in three areas, email, enterprise tagging, and people-finding.

Email Gravitation?

At Mallesons, they try to predicting the needs of the user through Decisiv. This application will make a guess at where the person will want to file the email. The rate of prediction success is quite high.

The search will recommend searches based on the metadata in the email. Do you want to search for emails from the person who sent the email you’re on? Or to the company that received the email?

Social Bookmarking

Social bookmarking can help assess which content is worthwhile. You get the benefit of a large number of people who are also tagging and assessing relevancy.

Malleson’s “Scotty” system, in Beta, will generate tag clouds around content. It will have a firm-based taxonomy, but will also allow user-generated tags and ratings. (Tag clouds show the importance of the content related to the tag by the size of the tag word).

Visualization and People Information

Felicity described "visualization" as a quick way to display vast amounts of quantitative data such as business results. For instance, others have joined news mentions of disease outbreaks with Googlemaps in the disease finder, Health Map.

Malleson’s PeopleFinder (which, although Felicity modestly didn’t mention it, won the Innovaction award this year) pulls together data from a lot of sources.

It starts with a small box in the desktop bar that searches google, or Interaction, or what have you by letter codes like “c” or “g”. A Mallesons people result shows their current availability. The goal was to have fewer external calls go to voicemail. There are twelve different icons indicating the status, anything from "on extended leave" to "on the phone."

People are “available” if they are logged in to PC. Secretaries can use this to identify what to do with calls for an attorney. It saves tremendous amounts of time in not making or receiving calls when they can’t be dealt with. A “Communicator” status can be updated. A right-click can lead to a VOIP phone call, email, or IM. A full-screen view of a person record also shows their week's availability, as per their Outlook calendar. A right-click leads to a while range of options for communicating with the person, from dialing the phone, IM, or calendar appointment, to email. A click on a person's "floor" leads to a floor map highlighting the person's location on the floor.

A scaled-down version of the People-Finder was available through mobile devices. Mallesons is also contemplating making some aspects of PeopleFinder available to clients.

The next version will "push" information on a person's own home page or on their contact record, like the person's current matters, clients, documents, and so forth. This is delivering information, not in response to a user request, but in response to an anticipated need.

Information Gravitation

"Pushing" information in response to an anticipated (rather than expressly stated) need is one form of information gravitation. So is enabling navigation between types of information on a document search result, such as a link to a person or matter. Another form of information gravitation is accessing all the possible ways to contact someone from their contact record.

Ultimately "cloud" computing, where vast amounts of enterprise software and data are contained on giant server farms (as with Google or Amazon), may enable even more interlinking, pushed contents, and connections than is possible now. Aristotle's vision come to life?

No comments: