April 10, 2005

How Google has Changed Enterprise Search

Vendors selling enterprise search engines have to deal with the new reality, the Google reality, that exists in today's marketplace. Google, and it's entry into enterprise search with the Google Appliance, has shaken up the industry according to Stephen E. Arnold in "How Google Has Changed Enterprise Search", Searcher Magazine, Nov./Dec. 2004. (This article is available in full-text for $2.95 from the InfoToday web site.)

According to Arnold, enterprise search vendors are having a difficult explaining why their complex search products are better than the ease and simplicity of Google. Arnold includes a table comparing public web search and enterprise search, noting the significant differences in how they work and what they're required to do.

Considering Google's positive image, the relatively low cost of the Google appliance, and their good financial position, enterprise search may never be quite the same.

Posted by Cindy L. Chick at 05:54 PM | Comments (0)

February 14, 2005

Search Appliance Matrix

The people that brought you the Desktop Search Matrix, the Goebel Group, have another comparison you might find useful, the Search Appliance Matrix. Search appliances, which are epitomized, but not limited to, the Google Search Appliance, are search engines for your intranet in a package complete with everything that you need to easily implement them.

The matrix compares software from Google, Thunderstone, and Vivisimo and includes the file types indexed, the number of documents, and cost. It's definitely very handy if you're in the market.

Posted by Cindy L. Chick at 04:38 PM | Comments (0)

January 25, 2005

Choosing a Search Engine

When it comes to evaluating search engines, most of us don't know how and why they work the way they do, we just know what we like. And we like relevant search results. Unfortunately, that doesn't quite cut it if you are charged with choosing a search engine for your organization. Kim Guenther lists the important capabilities and features that you should look for in a search engine in her article "Choosing a Search Engine" in the Jan./Feb. issue of Online.

(Note to Reader: Online articles typically appear on Looksmart's FindArticles a few months after publication.)

Posted by Cindy L. Chick at 06:06 PM | Comments (0)

September 07, 2004

20 Questions (With Answers) About Enterprise Search

Steve Arnold answers some of the most frequently asked questions about implementing an enterprise search engine in his article, 20 Questions (With Answers) About Enterprise Search (Online, July/August 2004). This kind of practical information can be hard to come by.
(The article isn't available on the web in full-text, but you'll find it soon on Factiva or Westlaw's ONLINE database within 6-8 weeks.)

Many of us think in terms of the Google search box as the ideal for providing access to information, but according to Arnold, only about half of Yahoo's users use the search box to get where they want to go. The "other half use a point-and-click approach to content....A search box is not a useful tool for about half of the populations's user." It sounds like there's still a place for a controlled vocabulary and browsable categories even in the age of more advanced enterprise search products.

Posted by Cindy L. Chick at 03:55 PM | Comments (0)

August 15, 2004

Selecting an Intranet Search Engine

I've had a white paper from Isys, called "Selecting an Intranet Search Engine" sitting on my desk for some time. It's time to tidy up, and I certainly don't want to lose track of it because there's some good stuff here, including an "Intranet Search Engine Requirements & Evaluation Checklist." So, what else, I figured I'd better blog it.

Posted by Cindy L. Chick at 03:17 PM | Comments (0)

June 02, 2004

More on Enterprise Search Engines from Peter Ozolin

At Paul Hastings we spent considerable amount of effort last year reviewing various search technologies, both in terms of "universal" search (that is finding information/data from a multitude of sources) and more targeted search solutions, such as West KM. This said, we found the problems that the various approaches are solving are different and thus we elected to pursue a two-prong strategy covering both types of knowledge searches detailed below.

1. DMS/Best Practice Database: I think we all understand the approach West KM takes, but perhaps a less known solution we evaluated was from a company called Practice Technologies was ultimately what our attorneys elected to pilot this winter/spring. Bottom line, our attorneys felt West KM's approach was too narrow and the reliance on citation did little to narrow down our most relevant work-product.

Instead, we felt the approach by Practice Technologies was broader and more comprehensive. Essentially, Practice Technologies assumes you have done a poor job of profiling your documents and reclassifies them based upon a rules-based engine. The focus being not only litigation orientated documents, but transaction documents as well. The reclassifying engine augments your already existing profile with substantive information such as jurisdiction, opposing counsel, type of document (motion to dismiss), etc. This data is rarely profiled, nor is it typically part of the profile.

An interesting aside, do a query on your DMS profiles and I think you'll find folks do a poor job of profiling, even if they tell you otherwise. Ultimately, our attorneys recognized it was becoming increasingly difficult to find on point work product, thus this project made sense to pursue. There's other integration points with this product, for example, procedure law (links to treatises, filing laws in CA, etc.).

2. Universal Search: This was a challenging area to evaluate. We probably interviewed upward of 10 companies. Bottom line in this area for us was a commitment to legal, the technology is becoming increasingly difficult to differentiate. However, there will need to be integration points to various legal vendors and we did not want to have to pay for these (as would be the case with Autonomy, which also has a pricing model that needs to be reconsidered). We selected Recommind.

So, what has been learned? Unfortunately, we are early in the game. But item (1) above is resonating most with our attorneys. It's premature to give any kind of meaningful endorsement in that it's early, but the key is that the targeted type of applications (whether Practice Tech or West KM) solve obvious problems -- poor profiling of DMS documents and make the task of doing legal research less onerous. We are likely to adopt this product firm-wide the way it is going. Item (2) Universal Search has promise, but I believe the ROI is longer term in that it takes time to work with the data to achieve promising results.

If I had to choose one of the above, it would probably be item (1) and we would choose a less robust search engine, perhaps DT Search. However, the risk there would be falling behind in working with the search technology that looks promising.

Finally, we may have the resources to do some of this in-house, but I would never do so in that the investment/pace of change with this type of technology will be difficult to keep up with.

Posted by Cindy L. Chick at 11:06 AM | Comments (0)

Are Document Management Systems Broken?

Document management systems (DMS) typically promise to provide robust full-text searching across entire document collections in an organization. In larger firms, which arguably have the most to gain by leveraging the knowledge housed in their DMS, that promise hasn't been realized, at least in part due to an inability to search across libraries in different locations, the lack of an intuitive search interface, and incorrect profiling by users. In connection with our discussion of enterprise search engines, Dennis Kennedy asks whether DMS are broken, and if so, what this means given the money spent on such systems. Very good questions...anyone have any answers?

Posted by Cindy L. Chick at 10:44 AM | Comments (0)

May 31, 2004

Enterprise Search Engines - Searching for ROI

Internet researchers accustomed to locating information from a variety of sources using search engines such as Google, are demanding a similar single method of access to their company's resources. In "Searching for ROI", (Econtent, May 2004), Robert J. Boeri points out that though the demand for enterprise search engines is increasing, it is also a hard sell in organizations where the routine question is, "What return on investment will search yield for the enterprise?" He discusses justifying enterprise search systems with several vendors including Verity, Autonomy and Tripplehop, but as he also points out, "Nobody asks for an ROI to prove the value of email systems."

Sorry guys, this one isn't available on the web, but it should appear on Westlaw in the DTBS database within 6-8 weeks.

Posted by Cindy L. Chick at 01:43 PM | Comments (0)

May 25, 2004

West KM/Lexis TotalSearch or Enterprise Search Engine?

According to PC Magazine, "search has become one of the hottest technologies in the computer industry." From products designed to search data on a single PC, to those aimed at small, medium or large companies, there is an ever growing number of vendors promoting full-text search products for document management systems(DMS), litigation support, databases, email, intranets, and some that try to do it all, high-end, enterprise search engines such as Autonomy, Verity and Recommind, just to name a few.

Where does WestKM and the Lexis Total Search products fit into the mix? Lisa Kellar, Practice Automation Manager at Hunton & Williams, explains the difference between these KM products and the enterprise search engines.

"West KM or Lexis Total Search don't meet the same needs as an enterprise search. West and Lexis offer tools, probably most useful to litigators, that allow you to conduct legal research while at the same time "discovering" internal work product that cite the same cases. They offer automatic updating or flagging of citations as well. They let researchers continue to use an environment they are familiar with and will need to continue to use anyway. They can also use either West or Lexis syntax to just search the internal portion of the knowledgebase. Firms are including a subset of their documents for inclusion in these systems and they are limited to documents stored in your DMS (which could include email).

Enterprise search not only would include everything in your DMS (not a subset) it would also search many other repositories within the firm. They offer conceptual searching and other fuzzy logic, but don't have the research orientation that West and Lexis offer. They might also require learning a different search syntax, although most offer natural language searching these days.

I don't think one is better than the other or one is a replacement for the other. I'm impressed with both solutions, but they meet different needs. Enterprise search is probably useful to a much larger audience than something like WestKM but the utility of WestKM might be higher for its users, than enteprise search. The difficulty comes in deciding whether or not its worth the money to have both types of solutions."

Posted by Cindy L. Chick at 08:38 PM | Comments (0)

February 15, 2004

The Search for a Google-Like, Enterprise Search Engine

Valuable information in any organization can be spread across a variety of databases, intranet pages, and of course, document management systems.

I'm convinced that we need to stop requiring attorneys to figure out which data store contains the information that they need, but rather, provide a search engine that will pull all the firm's resources together, and present the appropriate content based on the search terms. Everyone wants something "Google-like", and I don't see why they can't have it! Apparently the law firm of Cleary, Gottlieb agrees. Brenton B. Miller describes their search for just such a product in a recent article in the Legal Times, The Engine that Could.

Among the products considered were Autonomy Inc., IBM (Discovery Server), Recommind Inc. (MindServer), Triplehop Technologies (MatchPoint) and Verity (K2 Enterprise and Ultraseek). And the winner is......Recommind!

By the way, Recommind announced this week that Paul Hastings has also selected their product for their enterprise search engine.

Posted by Cindy L. Chick at 06:13 PM | Comments (0)