This document was revised on 4 September 2009. For more information, see the Corrections page on gartner.com.

This Magic Quadrant assesses vendors with capabilities that go beyond enterprise search to encompass a range of technologies. Their capabilities include: search; federated search, content analytics, such as content classification, categorization and clustering, fact and entity extraction, taxonomy creation and management, information presentation (for example, visualization) to support analysis and understanding; and desktop search to address user-controlled repositories to locate and "invoke" documents, data and e-mail.

CIOs, information architects, development managers responsible for information access and enterprise search, legal departments, e-commerce managers, website managers, intranet architects, portal developers, and security and risk professionals must understand the differences between different kinds of information access vendors, and the specific vendors themselves.

We consider all enterprise search vendors to be information access technology vendors. However, those that offer only search capabilities (frequently called "keyword search") are neither Visionaries nor candidates for the Leaders quadrant, although they are part of the market. Finding information, and acting on it intelligently, demands increasingly sophisticated and innovative strategies and technologies.

We recommend that large commercial organizations and government bodies select at least an information access platform vendor for most future projects. Platform vendors offer modular architectures, wide varieties of relevance modeling, multiple vertical applications and significant customizability. These organizations should also typically have a tactical vendor to increase agility for short-term and quick-start projects. Tactical vendors may lack architectural sophistication and customizability, but their products are quicker to deploy and easier to understand. Often, enterprises obtain search as part of a different product, such as a portal or enterprise content management (ECM) application. In such cases, these enterprises may choose to make this embedded search a target for federation. Large enterprises must also recognize the need to explore more specialized products for certain important projects, such as e-discovery, e-commerce search and research science support.

The market for information access technology has been altered significantly and permanently in 2008 and 2009. Google's product has considerably expanded its functionality with no real change in price and, most important, has architecturally changed to address more effectively very large-scale and variable-scale installations. Microsoft's integration of Fast Search & Transfer has made it a formidable competitor in all deal sizes. Specialist capability addressing particular business challenges or, in some cases, challenges in particular verticals, has proved to be the most lucrative direction for vendors to take. "Cloud" architectures are sure to drive a new wave of disruption favoring cloud specialists such as Microsoft and Google, and open-source alternatives with increased credibility are proliferating and gaining visibility. The result is a market in which specialists are favored in deals that match their skills, and generalists are favored when their products have predictable, affordable prices.

Information access technologies read applications such as document management, Web content management and relational database management systems to provide users with insight into their contents. Increasingly, information access technology is also expected to include results from enterprise applications, such as CRM and legacy systems. In addition, it looks outside enterprises as well, to premium sources of information, websites and elements of the social Web. Information access technology is often acquired as an embedded aspect of other applications, and OEM arrangements are significant for information access vendors. Portal, ECM, business application and other vendors frequently include enterprise search as part of their products.

The first and most mature information access technology is search engine technology. It has been common since the mid-1990s, and is typically applied to unstructured data, primarily text, in document repositories. It includes both enterprise and desktop search. Increasingly, autocategorization, creative visualization, content analytics and taxonomy support technologies are being added to this category.

This Magic Quadrant addresses enterprise information access vendors. In the past year, as Google has continued to develop market share and Microsoft's strategy for search in SharePoint and its integration of Fast Search & Transfer has taken shape, many vendors have oriented their technology specifically toward addressing particular business problems (such as legal discovery) or industries (such as publishing). Information access appliances are increasingly popular as enterprises like their simplicity. Also, content analytics is emerging as a category of its own; it is not fully covered here, but is a critical aspect of our innovation evaluation criteria. Many vendors that provide content analytics functions are not included here because they don't offer search.

In 2008, software revenue (new licenses and maintenance revenue) in the enterprise search market totaled $1.1 billion worldwide. We forecast that it will rise to $1.9 billion by 2013, with a compound annual growth rate of 11.7%.

The enterprise information access market, like most technology markets (although not necessarily many of those have existed as long), has seen many acquisitions, though the long-term trend for consolidation is partly offset by the arrival of new vendors. (In the past, Autonomy acquired Verity, Microsoft acquired Fast Search & Transfer, and Business Objects acquired Inxight Software. This year, Sinequa is new to the Magic Quadrant, as was Fabasoft-Mindbreeze last year.) That trend is slowing somewhat as the market consolidates. The variety of choices from midsize vendors aiming at all-purpose products is shrinking, as the megavendors compete in that market with growing enthusiasm.

Today, IBM, Oracle and Microsoft all have credible search tools to which enterprises can turn for an easy selection of search products, and frequently with virtually no impact on the price of their current contracts. Search is increasingly perceived as "infrastructure," especially when the projects to be addressed with it are not clearly revenue- or cost-facing. These large vendors (as well as Google, which is seeking to penetrate the enterprise market more deeply, and has done so with exceptional success in enterprise search) all represent default choices for organizations, governmental and commercial, that are forcing independent vendors to specialize in particular business problems, most of which cross multiple verticals.

Information access technologies find, collect and condense information or map its native location, so that users may actively seek information, analyze it effectively and remain informed. Enterprise search, in which users' queries are matched against an index to return relevant documents and data, is an element of information access. Information access also includes more sophisticated capabilities, such as content analytics (including, for example, entity extraction). Several vendors pursue OEM business — Autonomy is the most successful in volume and received positive reviews from the large software companies we interviewed. Microsoft still supports the OEM customers that Fast signed but is not pursuing new business. Vivisimo and Exalead are particularly interested in seeking OEM relationships.

Significant changes have occurred in the market following Google's rapid improvement of its search appliance line in the past two years and Microsoft's earnest efforts to address it via SharePoint-related products, as well as the Fast Search & Transfer acquisition. These two very important vendors now essentially dominate the market for the simplest installations, where overall project cost is less than $75,000 for the first two years. The most expensive part of the market, in which project cost is greater than approximately $600,000 for the first several years of a project, is increasingly tipping to vendors with specific skills for particular business problems. The middle market is now largely dominated by incumbent ECM and business application vendors that present search as an infrastructure element not worth seeking separately. Consequently, it is very likely we will choose to introduce in 2010 a MarketScope for mid- and small-cost projects that would be specific to "enterprise search" alone. The highest-value projects are frequently now offered to specific subsets of the overall vendor list, and that market might not require a separate vendor evaluation. Many such project types — e-commerce, e-discovery, content monitoring and filtering — are covered in different vendor evaluations.

While the need for search remains strong, it is not likely to see substantial growth again in generalist products, sold separately from other applications. The time and effort devoted by many major vendors to addressing search across their product lines means enterprises will continue to acquire and incorporate search into their projects, but not that they will invest in stand-alone search products from midsize vendors. The market will shift to very large vendors selling platforms as aspects of their product lines, and significantly smaller vendors selling applications founded in search that will not be otherwise available.

We have included in this Magic Quadrant only those vendors that incorporate search as the foundational capability of their information access products. Many include other capabilities such as autocategorization, taxonomy functions and clustering, but we excluded those that offer only these capabilities, with no search.

We have included only vendors that actively market their products in at least two of the following markets: Asia/Pacific; North and South America; Europe, the Middle East and Africa. We also required vendors to have grossed more than $10 million in revenue (for any kind of sale) in 2008. We considered publicly traded vendors' filings as sufficient evidence of this. We asked privately held vendors to provide documentation supporting such figures, and we consulted third-party sources.

We also required that vendors actively market their products in more than two major vertical markets, and in more than one horizontal application category (such as e-commerce and customer self-service).

Open-source search engines do exist, but none is significant enough to threaten the commercial market imminently. Lucene is the most prominent, but we have excluded it because it is not a commercial entity and is supported by a noncommercial community. It is growing in influence, as numerous vendors are turning to it to manage indexing. Lucid Imagination, a small systems company, seeks to establish itself as a commercial provider of Lucene technology.

In addition, we required that vendors are able to sell their information access products separately from other products (such as ECM, portals, infrastructure, records management and CRM offerings).

Sinequa is new to the Magic Quadrant. This European vendor has traction in France and has started to try to penetrate the U.S. market.

We have removed Open Text from the Magic Quadrant because it no longer sells its search technology separately from its various content management platforms and other products.

Product/service includes a product's versatility in providing connectors to content repositories. Additionally, the operating systems supported are a minor component of this score. Proof of a product's ability to scale to large volumes of content or queries and its ability to address desktop search have also been transferred from Completeness of Vision in this year's evaluation.

Overall viability of a vendor is determined by a number of financial indicators that are elements of the Gartner Vendor Rating Financial Model. We also considered whether a vendor is publicly traded, as this affects enterprises' perception of its strength.

Sales execution/pricing reflects the appeal of the vendor's pricing models and the vendor's success in attracting numerous and effective system integration partners. Companies prefer pricing models that are transparent, predictable, and sustainable for their budgets; our evaluation scores reflect that preference.

Customer experience reflects what we know of a vendor's performance from references and impressions of its position and performance gained through our inquiry and outreach services.

Operations assesses a vendor's geographical breadth in sales and support.

Market responsiveness and track record and marketing execution are not covered. We address responsiveness to market needs in customer experience, and responsiveness to overall market trends is reflected in revenue success. Marketing execution is difficult to measure and compare in a market with extremely different go-to-market strategies.

Market understanding is a combination of the vision that a vendor's management has for how to thrive in the extremely challenging market for information access technology and of how that vision is manifested in practice.

Marketing strategy assesses how a vendor actually brings its products to market. We rate it as standard value because it is difficult to compare the marketing strategies of vendors with such disparate targets.

Sales strategy addresses vendors' selected target markets and their means of succeeding in them. The variety of target markets and strategies to reach them makes it a difficult criterion on which to compare vendors; consequently, we give it a standard value.

Offering (product) strategy reflects a vendor's abilities to: provide innovative and broad relevancy models using document content and other information (such as user behavior); analyze user queries; federate, provide valuable administration tools; establish user profiles and make search relevancy transparent.

Business model assesses a vendor's product and service delivery models; greater breadth in such models — particularly those with the greatest potential for growth — confers a better score. Delivery models, for example, increasingly include cloud-type delivery, as manifested in models based on application service providers, hosting and business process outsourcing. Such models are increasing in popularity and are a factor in determining Completeness of Vision. This is a comparatively isolated element of vision, and enterprises do not treat it as a significant decision factor.

Vertical/industry strategy looks at visionary installations that vendors describe. We do not value it more highly to avoid inflating the impact of a few marquee capabilities, without ignoring the illustrative value of particularly effective installations.

Innovation is assessed by looking at several different factors. These include: the vendor's abilities to address nontextual documents and objects (including audio and video); perform content analytics; aid users in disambiguating results through conversational interfaces; use elements not resident in documents to establish relevancy (such as citation analysis); and deliver "social search."

Geographic strategy is covered not here but under operations (see the Ability to Execute criteria). In search, the availability of sales staff and support is an effective issue in execution. In inquiries, Gartner enterprise clients identify the availability of sales or support as an element in deciding whether to consider a vendor — not as an element of its vision and direction.

All the Leaders demonstrate significant architectural flexibility. They also have strong, innovative and broad means of determining the relevancy of results returned to users, and of providing developers with the tools and the flexibility to tune relevancy settings. They have the finances to weather hard times and sufficient resources to invest in both organic and inorganic technology and business growth.

Challengers possess sufficient resources to penetrate the information access technology market effectively. However, they typically lack the vision to address all information access opportunities.

Visionaries demonstrate imaginative and insightful approaches to the information access technology market, but currently lack the resources to prove their leadership and guarantee future strength. They all possess architectural flexibility and creative means of establishing relevancy. Greater financial resources and more market traction would improve their position. No Visionaries appear in this year's Magic Quadrant, as changes to what we value as Completeness of Vision versus Ability to Execute caused significant movement.

Niche Players possess the attributes necessary to fulfill the needs of certain types of information access project, but they lack the depth and breadth to satisfy a wide variety of projects. In some cases, they lack the proven financial resources of Leaders and Challengers. Nor can they demonstrate the depth of vision that indicates they are leading the market. On the other hand, they are typically right for a particular set of needs, and they offer attractive pricing, special capabilities and vertical-market or application knowledge.

Autonomy acquired Interwoven in 2009 and is seeking to expand into law firms and Web content management, including online marketing. It has introduced a customer care program to improve customer relations and service.

• Autonomy has extremely broad sales and support presence in many regions. In Europe, it can sometimes position itself as the best global software vendor with local presence.
• Autonomy provides for deep user profiling, including users' search histories, explicit roles, behavior in day-to-day authoring, and interactions with external and internal information sources.
• Autonomy has particular facility with nontextual multimedia such as video, having invested in video before any other vendor in this year's Magic Quadrant. Content analytics is another strength.

• Autonomy's opaque pricing model can irritate prospective customers, as they cannot predict what price the vendor will offer. Also, they often find its proposed price the highest of any they receive.
• Sales processes are insistent, and sometimes aggressive.
• Gartner clients report they find it a challenge to get sales support and attention for upgrades and improvements.

Endeca's use cases include an element in which structured and unstructured data are viewed and explored together; such convergence is difficult to accomplish and Endeca has been an early leader.

• Endeca's connector family gives it a particularly broad capability to index from a variety of content sources. These include manufacturing applications to which few, if any, other vendors connect. Additionally, Endeca has a unique search federation connector strategy that extends its reach through enterprises in an innovative fashion.
• Endeca records user searches, navigational behavior and the intersection of the two, providing a unique perspective on users' past choices and unusually rich user profiles. This information can be combined with explicit profile data from enterprise-established permissions and privilege lists.
• Endeca has a particular facility with action-oriented content analytics, such as analysis models covering a broad spectrum of user behaviors and textual analytics (including concept extraction). Its recommendation and document summarization engines are especially effective.

• Endeca lacks the same level of ability to natively address nontextual data such as audio and video that some other vendors have, but can partner as necessary.
• Challenging market conditions make Endeca a uniquely attractive acquisition target because of its size and position.
• Endeca would benefit from a more transparent pricing model.

Exalead is well regarded for managing search installations with extremely large scale.

• For a comparatively young vendor, Exalead offers rich content analytics.
• Exalead has demonstrated the ability to handle extremely heavy search traffic, including 800-query-per-second "spikes" on a popular site with 15 million documents.
• Exalead has successfully developed a strong revenue profile in Europe.

• Exalead does not invest significantly in the ability to exploit users' historical behavior or explicit status in an organization as a means of determining the relevancy of results.
• Although Exalead has improved its position in North America, it still relies largely on European business.
• Exalead must expand its sales efforts in the U.S. to fully address its opportunity.

Expert System is seeking to expand into the U.S.

• This vendor is extremely good at analyzing and searching external data sets for competitive and market intelligence.
• Expert System is highly extensible and allows for significant customization.
• Enterprises with the resources to devote the necessary time and effort to Expert System customization are extremely pleased with its depth and power.

• Expert System is still Europe-centered.
• Expert System targets small-scale installations measured in gigabytes.
• Expert System does not invest significantly in the ability to exploit users' historical behavior or explicit status in an organization as a means of determining the relevancy of results.

Fabasoft-Mindbreeze, an Austrian vendor with a rich customer base in Central and Northern Europe, is not well known in the U.S., but it is taking steps to establish itself there.

• Fabasoft-Mindbreeze has a simple pricing model based on named and unnamed user count. It provides a predictable initial cost base.
• A significant number of index connectors allows for a variety of content sources.
• For a comparatively new vendor, Fabasoft-Mindbreeze has strong federation capabilities.

• Fabasoft-Mindbreeze remains Europe-centered.
• Fabasoft-Mindbreeze offers limited user profiling.
• Fabasoft-Mindbreeze must attract more system integrators and expand its recruitment if it is to grow swiftly in North America.

Customer interviews indicate Google's products are extremely easy to use. It is now also layering on greater sophistication and the ability to handle significantly larger-scale installations. Google's most recent revision shows significant promise in achieving larger-scale installations and has added substantial conventional improvements, such as more robust load-balancing. It continues to lag behind the most sophisticated vendors in richness of functionality but leads in simplicity and ease of use.

• Extremely simple pricing and delivery models include application service provision and well-defined tiers of search appliances.
• Google has demonstrated an ability to handle very large-scale installations.
• Google is extremely well regarded for the simplicity of its installation and administration.

• Google does not invest significantly in the ability to exploit users' historical behavior as a means of determining the relevancy of results. (It has recently added role-based profiling.)
• Google lacks relevancy tuning capabilities to compete with those from other vendors, but it has added new dimensions to such features.
• Google does not provide the ability to expose how much of its relevancy score is calculated, which limits administrators in their ability to tune the results obtained.

IBM has advised us that it still pursues general-purpose enterprise search, but that increasingly its strategy is to target specific business problems that bring together structured and unstructured information. Examples include e-discovery, healthcare process streamlining and fraud reduction. We have detected this in our inquiry trend as well. This strategic shift, as well as our transference of some elements from Completeness of Vision to Ability to Execute, is at the heart of its return to being a Challenger, after having been a Leader in 2008.

• IBM is investing significantly in the Lucene open-source search kernel as a foundation for its products.
• IBM has particular facility with nontextual multimedia and content analytics.
• IBM has demonstrated in its briefings that its capabilities in Web 2.0, such as social search, are growing.

• IBM has multiple information access products, including a free product developed with Yahoo, a basic enterprise edition, an edition with significantly advanced capabilities for strategic deployments in particular vertical sectors, and an offering (IBM Content Analyzer) intended to fuse search and business intelligence. Although IBM's ability to address lucrative horizontal sectors such as e-discovery bodes well for the future, and its broad range of offerings caters for deals of different sizes, we believe these factors also present IBM with a challenge when it comes to rationalizing its branded products.
• For many of its products, we have not seen evidence that IBM is investing significantly in the ability to exploit users' historical behavior or explicit status in an organization as a means of determining the relevancy of results.
• IBM must rationalize its own use of some competing search technologies elsewhere in its organization to increase its reach and credibility.

Isys' search products constitute a simple platform suitable for a broad spectrum of search projects.

• Isys Search Software has a significant sales and support presence in many regions. For a small vendor founded in Australia, it has a particularly strong presence in Asia.
• A simple pricing model based on seats and servers makes it easy for prospective customers to predict their costs.
• For a small vendor, Isys has particular strength in content analytics.

• Isys does not invest significantly in the ability to exploit users' historical behavior or explicit status in an organization as a means of determining the relevancy of results.
• Isys lacks the highly flexible architectural facility of the most visionary vendors.
• Isys must continue to align itself to particular specialties or be squeezed by megavendors.

Note: At the time of publishing, EMC announced that it has signed a definitive agreement to acquire Kazeon. Kazeon has emerged as a Challenger because of significant financial success and an effective, transparent pricing model that has aided it in developing a good sales channel.

• Kazeon has an extremely simple pricing and delivery model for its appliances.
• Kazeon focuses closely on the compliance and e-discovery sectors.
• Kazeon has very strong customer references.

• Kazeon does not invest significantly in the ability to exploit users' historical behavior as a means of determining the relevancy of results.
• Kazeon lacks the same level of ability to natively address nontextual data such as audio and video that some other vendors have, but can partner as necessary.
• Kazeon does not have a strong direct presence throughout the globe, but does have strong partner, reseller and OEM channels.

Microsoft has completed its acquisition of Fast Search & Transfer and is seeking to embed it into SharePoint while maintaining a version for the most visionary installations on company websites.

• Microsoft's technology has demonstrated the ability to handle extremely large data sets and heavy traffic, including installations with spikes of more than 1,000 queries per second and multiterabyte corpora.
• Fast's connector family gives it a particularly broad capability to index from a variety of content sources. These include records management applications that relatively few, if any, other vendors connect with.
• Microsoft has particular facility with nontextual multimedia, especially video.

• Microsoft's multiple products — following the acquisition of Fast — cater for a broad range of needs, but can confuse prospective customers.
• Microsoft must continue to assimilate Fast effectively.
• There is potential for Microsoft to become distracted as it strives to provide effective search facilities in multiple products and divisions.

Oracle's recent efforts have focused on integrating its search capabilities across its own product line. This intentional curbing of its ambition is a factor in it receiving a lower score for Completeness of Vision in this Magic Quadrant than the previous one, along with our revaluing of some elements in our scoring system. However, Oracle's efforts have extended the Secure Enterprise Search product's influence throughout Oracle dramatically, which is wise.

• Oracle's search is broadly adopted among the vendors' many applications, and serves as an effective means of bridging the information and interfaces within them.
• A simple pricing model based on the number of CPUs and users (or workers) makes it easy for prospective customers to predict their costs.
• Oracle's foundation technology, Oracle Text, is stable and well regarded.

• Oracle does not invest significantly in the ability to exploit users' historical behavior or explicit status in an organization as a means of determining the relevancy of results.
• Oracle provides limited transparency of its relevancy calculation model.
• Oracle lacks the same level of ability to natively address nontextual data such as audio and video that some other vendors have, but can partner as necessary.

Recommind has expanded its efforts into other application categories targeting the legal industry and enterprises' legal and regulatory compliance concerns, including e-mail archiving and e-discovery.

• Recommind has invested extremely effectively in providing collaboration and process support for professional services and litigation-oriented markets. Examples include effective support for expertise and project location, and automated document review and analysis.
• Recommind is extending its European presence with sales and support offices.
• This vendor is unusually accomplished at using statistical analysis to extract meaning from documents in an automated fashion.

• Recommind must continue executing its vertical-market strategy if it is to fully penetrate other lucrative markets effectively.
• Recommind lacks the same level of ability to natively address nontextual data such as audio and video that some other vendors have, but can partner as necessary.

Sinequa is prominent in France but has not yet become well known in the U.S.

• For a comparatively new vendor, Sinequa has strong content analytics and relevancy capabilities.
• Sinequa is extremely well regarded for the simplicity of its installation and administration.
• Sinequa has very strong customer references that describe particularly effective engagements.

• Sinequa remains Europe-centered.
• Sinequa lacks the same level of ability to natively address nontextual data such as audio and video that some other vendors have, but can partner as necessary.
• Sinequa lacks significant capabilities for federated search.

For a strong platform, Vivisimo is relatively easy to use but will need to emphasize its particular suitability for some application categories.

• A simple pricing model based on corpus volume or document count makes it easy for prospective customers to predict their costs.
• A significant number of index connectors allows for a variety of content sources, including external sources used by legal departments, for example.
• Vivisimo's next-generation capabilities, such as social search, are growing.

• Compared with the other Leaders, Vivisimo's financial resources are fairly limited.
• Vivisimo must continue to align itself to particular specialties or be squeezed by megavendors.
• Vivisimo lacks the spectrum of delivery models of some of its rivals.

ZyLAB's historical focus on e-discovery and investigative tools has positioned it well for the currently strong market in these capabilities.

• For a comparatively small vendor with less extensive financial resources than other Leaders, ZyLAB has an unusually broad sales and support presence in Europe and Asia.
• ZyLAB addresses images, video and text with unusual effectiveness.
• ZyLAB has a strong strategy for addressing lucrative vertical markets related to litigation and investigation.

• ZyLAB supports only Microsoft operating systems.
• ZyLAB does not invest significantly in the ability to exploit users' historical behavior or explicit status in an organization as a means of determining the relevancy of results.
• ZyLAB's focus on its chosen horizontal markets has guided it away from addressing such broad-spectrum innovations as social search.