Questions to ask when beginning a search project

By Whit Andrews

Gartner Inc.

Search technology for text and other digital formats remains a vital market. Several visionary vendors may be good choices for your enterprise.

The role of search applications as useful, discrete products independent of infrastructure and other applications dates back roughly to the emergence of the Web as a useful information-location medium. The breadth of solutions that search technology offers to address the seemingly simple problem of how to find useful information guarantees that it will continue to be attractive for the near future. This surfeit of ambition does not, however, guarantee success, or even survival, for its participants. The enterprise search business faces significant price pressures and is unstable.

This is a product of several factors:

– The innovation in search, and the new vendors to provide improved capabilities, has continued. The “problem” of search, which includes advanced mathematics, linguistics and enterprise software design, continues to attract brilliant minds. A glut of visionaries drives down the price these visionaries may charge.

– Major generalist vendors have introduced their own search technologies and bundle them with their own products. Oracle, SAP and Lotus all sell search applications as part of their product suites. At the least, these products are sufficiently effective to delay enterprises’ selection of a specialist. In some circumstances, they may provide sufficiently deep functionality to win strategic deals.

– Having had access to particular technologies and functionalities for a relatively long time, enterprises believe they have grasped the basics of a search installation, and have attached an unflattering price tag to it. They don’t want to pay US$50,000 for “just a basic search engine.” This includes enterprises seeking to make division- or enterprise-level external Web sites searchable. Many prefer a price closer to US$15,000, even if that limits them severely in their selections.

However, improvements in search technology continue. New vendors tend currently to establish vertical specialties rapidly. Enterprises selecting products for strategic projects that are intended to generate clear return on investment in productivity or self-service have many things to consider as part of the selection process.

Gartner’s recent Magic Quadrant for Search Vendors identifies six visionaries and two leaders. In this Spotlight, we examine some of the visionaries.

Leaders Verity and Autonomy are both generalist vendors, which is to say they are appropriate for consideration for a wide variety of projects, including knowledge management and customer self-service. Visionary generalists include Fast Search and Transfer and Convera (a relative newcomer and a longtime vendor with a long history, respectively). iPhrase Technologies, InQuira, the enterprise software division of Ask Jeeves (which Kanisa has agreed to acquire), and Endeca Technologies provide specialty capabilities. Whether these vendors should be considered specialists or generalists is an issue that tends to be addressed in requests for proposals, but none of these vendors provides the broad capabilities of the clear generalists.

The prices of these products vary significantly, and individual vendors are offering incentive-based pricing to capture new business during the current economic slowdown. Enterprises willing to invest staff time in development and maintenance will find less-expensive alternatives that are appropriate. At Progress Energy, the low-priced Google Search Appliance proved sufficient in a Web-facing installation. At Capital One, the low-priced Verity Ultraseek product served as the keystone of a substantial customer service knowledge base project.

Gartner recommends that enterprises begin any search vendor evaluation project by ruling in vendors rather than ruling them out. These questions are appropriate:

1. Does the enterprise desire or accept an application service provider model of search provision?

2. Does the enterprise desire or accept an appliance model for search provision?

3. Will the vendor serve a single project or be an enterprise-wide default for all new projects?

4. What repositories of data will be searched? Will the search product call applications or simply search an index? Will text be the only significant format in which information is stored?

5. What level of security will be necessary, and what means of authentication will be used?

6. What interface will be used for result selection? Will the enterprise desire categorical navigation? Is persuasive merchandising a goal?

7. What interface will be used for query input? Will the enterprise need to use a natural question format, or stick to the familiar keyword input format?

The answers to these questions often are not obvious in the beginning phases of search product selection, and we recommend having a cross-enterprise team of business and IT managers to answer them. The answers should provide enterprises with an initial stage of selection that will allow a significant reduction in the number of vendors they must consider – and clarify their expectations about what problems the selection process will solve.


“Fast Search Offers Architectural Strength, Broad Choices” – Fast Search and Transfer provides effective means for indexing content in static files or in dynamic page assembly repositories, such as Web content management, document management or relational database management systems. By Whit Andrews

“iPhrase Expands on Query Input and Answer” – iPhrase Technologies offers self-service applications and scenario installations from which inexpert users will experience rich results. By Whit Andrews

“Progress Energy Simplifies Its Site Search Strategy” – A simple product for a small group of documents with unremarkable formats can be considered if the search function is not a strategic tool for users. By Whit Andrews

“Endeca Addresses Data Navigation and Revelation” – Endeca Technologies offers a perspective on multidimensional data from which users can select attributes and parameters from a long list of choices. By Whit Andrews

“InQuira Relevancy Includes Flexible Language Modeling” – InQuira’s complex, broad set of methods for determining the relevance of documents should be considered for high-value self-service installations. By Whit Andrews

“Kanisa Repositions Customer Service Search Engine” – Kanisa Site Search should be included in your initial vendor evaluations through at least 1H04 if you are building a self-service installation. By Whit Andrews

Document Type (Strategy, Trends & Tactics)

Entire contents