Practice groups can't find their stuff

Keith Mayfield, the chairman of an AmLaw 100 firm, has just returned to his office from an internal planning meeting with the chairs of the firm's key practice groups.

The chair of the intellectual property group expressed frustration that his group had a rich store of resource materials, but had trouble making them available to the group's lawyers. The group has a paralegal to organize the materials, a resources page on the practice group's intranet and access to the firm's enterprise search engine. Still, group members are constantly e‑mailing around asking for good examples of certain documents, insights into certain issues and experience with certain types of transactions and disputes.

This triggered the voicing of similar concerns by several other practice group chairs. It was resolved that the problem was sufficiently pervasive within the firm that an initiative needed to be undertaken.

Internal ideas don't pan out

Back in his office, Keith calls the IT Director to discuss the problem. After a few minutes listening to the IT Director expound on the benefits of the enterprise search engine the firm has had for several years, Keith realizes the IT Director doesn't have sufficient insight into the way lawyers work. Keith next speaks to the firm's head librarian. She confirms that finding internal work product and other resources is indeed a problem. Lawyers often call the library research desk looking for practice group-specific material. Her researchers are as frustrated as the lawyers about the difficulty in finding relevant material. The head librarian doesn't have any ready answers, though, for how to fix the situation, other than to suggest that the practice groups do a better job of organizing their own material.

Networking leads somewhere

At his next meeting with a group of other AmLaw 100 chairmen, who rotate as hosts to discuss practice management problems of common interest, Keith raises the problem of making internal resources more findable. Many of the other chairmen are uninterested in the issue, either because they believe they don't have the problem or don't understand its significance. One, however, suggests that Keith speak to a consultant who'd been helpful.

A consultant zeros in

The consultant tells Keith that lawyers always ask for their search experience to be like Google, but that what they should be given is one that's like Amazon.

Keith asks the consultant to explain. Lawyers want to type their query in a search box, the consultant states. They then want to receive a list of highly relevant results, as they would in a Google search. The Amazon search experience is more sophisticated, he continues. After the user types the search terms in a query box, Amazon presents the shopper with a list of narrowing filters in the left panel, such as by various price categories, product features, customer ratings, brands and seller companies. Clicking on a filter narrows the results. Each filter also indicates how many items are tagged with that filter.

The advantage of this approach to searching is that it quickly enables the user to narrow the search to a manageable handful of extremely relevant items. This produces the ideal search – a small but complete list of highly relevant results.

This new way of searching is catching on

Shopping web sites have embraced the combination of filtering and word searching for quite some time. Many other organizations have as well, including news publishers, medical organizations and scientific groups. Law firms are starting to take this approach.

Comparing internal law firm search to Google is apples to oranges

Google search mostly looks at the number of other web pages that link to the target page. This works because there are millions of web pages, and millions of links to each other. Searching within a law firm won't produce similar results to a Google search because there are no links among the law firm documents. The search principle must be different (page rank in Google vs. relevancy rank in a law firm search). Relevancy ranking in a law firm search can be based on many factors, but generally reflects the number of times the search terms appear in the document. If the terms appear in the document title, that typically gives an extra relevancy boost. With this brute force approach, it's easy to see why the results list is often long and why so many of the hits are not relevant. Some law firm search engines claim to be smarter by including synonyms or even concepts in the search. These features don't improve results very much for the average user.

In contrast, the Amazon-type search is effective because it allows the results to be narrowed using highly relevant filters. The same can be done for law firm search.

How Amazon search works for law firms

It makes perfect sense that an Amazon-type search relying on filters works much better than a word search. This is because the Amazon search requires real upfront work by people, not mere reliance on computers. First, a list of subjects for tagging has to be developed. There should be many sets of these lists:  for example, one list for legal topics, another for transaction types (or dispute types), and still others for document types, industries and geographic regions. In order to be useful, each list should have a tree structure with many levels. The lists can't be bought from a third party. They have to be developed specifically with the document collection and user group in mind. In a law firm, these types of search systems are best developed for a specific practice group.

The second element of upfront work involves tagging each document in the collection with the relevant items from each list. Not all documents will be tagged with elements from all lists, but many documents will be tagged with many elements from many lists. Tagging can be done only by someone familiar with the underlying content, such as a former practicing lawyer employed for this purpose.

Amazon has invested resources in creating filter lists and tagging its product descriptions with the relevant items from the lists. This involves significant effort. Yet it also produces powerful results.

Whether the effort to create tagging lists and to tag a document collection is worthwhile for a law firm practice group depends in part on the benefits in quality or efficiency expected to be realized. The anticipated benefits are, of course, not objectively determinable. They will be arrived at through interviews with lawyers in the practice group. Whether the effort is worthwhile also depends on the size of the practice group. Larger groups can spread the cost over a larger user base.

How a law firm could get started

The consultant suggests that Keith identify a practice group and run a pilot. The practice group should be reasonably large. If too small, the benefits of the new search system vs. traditional methods such as “asking around” may not be as obvious. Interviews with selected associates and partners in the practice group should be conducted to confirm the group wants to improve the findability of its internal resources.

What documents to include in the collection

As for the documents to include in the collection, that will depend on the nature of the group's practice and what they are seeking to get from the new search system. One practice group may want to improve the findability of substantive commentary about subjects and relevant precedents. Another practice group may want to categorize as many documents in their market as they can get their hands on, whether or not the firm was involved in the transaction. Another group may want to capture the main transaction documents for every matter on which the group has worked.

For groups seeking to capture substantive commentary, the places to look (in addition to the group's own central resource file, if it has one) are the personal resource files of respected partners and senior associates. These are the people generally known within the group as the ones to ask when explanatory material or precedents are needed. This material can be scanned and then tagged.

In all cases, a process will need to be developed to capture new material on an ongoing basis, so the content of the collection remains fresh and relevant.

Create filter lists and find someone to tag

A set of lists to use for filtering must also be developed. This could be done on a one-time basis through a combined effort of two or three senior lawyers in the group who are familiar with the document collection. It may be self-evident, but the nature of the document collection must be decided before the filter lists can be developed.

The person who will perform the tagging, both initially and on an ongoing basis, also needs to be identified. It is unrealistic to expect the practicing lawyers in the group to nominate documents for the collection. It is hopelessly unrealistic to expect them to perform tagging. There could be a part-time opportunity here for a former practicing lawyer in the group who, for lifestyle reasons, has chosen to leave full-time practice.

Identify the software platform

There are two law firm-friendly software platforms. Recommind's enterprise search product can be configured to perform filtered search on a SharePoint-hosted document collection, and perhaps on document collections hosted in other locations. HP Autonomy also has a filtered search product that integrates with its iManage WorkSite document management system. Other software used by larger non-law firm organizations could also be considered, but may involve more development effort and cost to configure for law firm use.

Build and test the system

Once the filter lists are created, the documents are identified and the software is in place, the firm can tag the documents and start testing the system with lawyers in the practice group. During the testing phase there will inevitably be adjustments – modifications to the filter lists, changes in document collection or tagging procedures and revisions to the user interface of the software. Nevertheless, it should be apparent fairly early in the testing process whether the system is achieving the desired results by allowing users to find relevant material quickly.

Conclusion

Encouraged by his discussion, Keith decides to engage the consultant to assist with a pilot for the intellectual property group. They seem sufficiently motivated and are large enough to justify the effort. The chair of the group suggests they bring in a former associate to assist with tagging and other resource organization efforts. She left several years earlier to focus on her family, but wants to return on a controlled schedule now that her kids are older.

A year later, the system is up and running. In terms of actual work, it could have been done in three months, but there were many delays in getting the timely attention of key lawyers in the group to prepare the filter lists, to identify documents for the collection and to provide other needed input. There were also delays in completing the RFP process for the search software vendor and the software consulting firm hired to assemble the system. These delays were due partly to unavailability of lawyers for key meetings and partly to personnel constraints within the IT Dept.

The lawyers in the intellectual property group are extremely pleased with their new resource. They have plans to add additional types of documents to the collection. Several other practice groups within the firm are vying to be the next to get the system. While frustrated (but not surprised) at the overall pace, Keith is quite happy with the way things have turned out.

[Photo credits: © Can Stock Photo Inc. / AnatolyM & texelart]