Hunting high and low

With ever growing reserves of company data that need to be accessed quickly and accurately, an efficient enterprise search engine is a must. IDC’s Sue Feldman talks to Exalead, Google Enterprise and IBM Information Management Software.

SF. Define what you mean by "search". Is a search engine all that you need for today's enterprise search?  What are the pieces that are requirements for an enterprise search application?
AB. Search is the way by which information-seeking users, be they an organization’s knowledge workers, executives, or customers, find the insight they need to make decisions and take action. It is the unifying metaphor for quickly discovering accurate information, answers, and business insight on demand and in context, regardless of where the information is stored.

While there are many search engines available with varying capabilities, a truly effective search solution goes far beyond the engine itself. Effective enterprise search solutions address business problems, integrating search into the context of a business activity or process. Whether it’s accelerating intra-organization collaboration by helping knowledge workers find and reach subject matter experts, driving real-time efficiencies by providing critical customer information in a contact center, reducing risk by enabling proactive compliance monitoring and eDiscovery, or even giving organizations a view of what their customers are saying about them, search is a fundamental part of modern information-centric business activity. And so it’s rarely just a search engine – a true solution extends the engine with semantic processing, deep linguistic understanding, extensive and secure information connectivity, and rich, targeted user experiences.

FB. The etymology of the word “search’ traces from the Anglo-French cercher, searcher, which means to travel about, investigate.  It has lots of synonyms.  However, for the purposes of this interview, I see the term as something that is harder than it needs to be. Accessing information is not a satisfying experience for most. That was the impetus for the creation of Exalead.

Those who think a ‘search engine’ adequately handles the needs of an enterprise perform do a double disservice for their organisations. We call this the consumerisation of enterprise search: the simplistic approach people use for Web search applied to the enterprise environment. Enterprise search deals with a data environment that is not homogenous.   Instead of primarily HTML content, enterprise search must deal with 295 different file formats, plus access controls and security protocols.

As evidenced by exalead one:search technology, we believe there needs to be a blend of the old school of search, typified by an ability to handle the complexities of information access, and the new school, which is exemplified by a clean user interface, similar to what most use for their Web searches. But with a twist: the ability to search and navigate the way you think, by serendipity.

In addition to the user interface, there are other critical elements for an enterprise search application. It should be easy to deploy and integrate with existing applications, easy to use and manage, secure and scalable.

DG. Search is the oxygen of the information age. The search box has become the veritable command line interface to the world's information. Users being ‘trained’ by the consumer world on Google.com expect to be able to enter a few keywords into a search box, click a button and find exactly what they're looking for in a fraction of a second, with access controls applied in real time. Enterprise search systems need to be relevant (giving the user the right answers), comprehensive (providing a single interface across all content sources in the enterprise), secure (showing users only the content they have access to see), and fast (giving users the answer in a fraction of a second).

SF. The search and discovery market is growing rapidly. What is driving this growth? Why are customers investing in search and discovery applications?
AB. Initially, the growth of the search and discovery market has been driven by organizations turning to search as a way to tame the massive explosion of enterprise information while meeting demand from their users for easy, Internet-like access to corporate content.

Increasingly, however, enterprises are recognizing the hidden value in their information assets, and today’s investment in search is being motivated by the need to create competitive value out of business information, enabling organizations to free information from silos so it can be better used to improve efficiency and drive new business innovation. We see this desire to fully leverage information as the key driver behind the continuing growth of not only basic enterprise search and related technologies like content classification, but also emerging high-growth adjacent spaces like text analytics solutions.

FB. There are two main drivers of growth. Firstly, there is the shift from structured to unstructured data. In the last few years, the scales have tipped, dramatically, as it relates to the type of data created in an enterprise and needing to be searched.  Computing systems once revolved around structured data fields.  Today, with email, documents, presentations and other content forms, this unstructured data represents 80 percent of the content in an enterprise.

Secondly, we have the ever-growing challenge of finding the relevant information. It is very much a reality of our work to utilize previously produced material. The more efficient and effective we can be in searching for, locating and accessing just what’s needed is an imperative.

Companies are investing in search and discovery applications to be in compliance with governmental and industry regulations and to maximize the value of the intellectual capital that they have produced.

DG. There are two main forces driving growth in this market. First is the overall consumerisation of enterprise IT. Users expectations are being set by the technology they interact with in their personal lives and search has become a defining technology on the internet. In the sea of content on the Internet that grows daily, search is the way people are able to find and leverage information that is useful to them. Users are looking for the same capability in the workplace, and those heightened expectations are being underserved by the traditional search vendors. Second, legal, compliance and regulatory issues are forcing companies to look at their approach to information management and oversight. These two trends along with companies' desires to better leverage their collective wisdom and company information are fuelling this market.

SF. One of the trends that IDC has been following is the demand for unified access to information. That is, the requirement that both structured data and unstructured content in any format be accessed from a single interface, and that it return merged results from all repositories. Is this a practical goal, and do you see this demand from your own customers and prospects?
AB. We do believe this is a practical goal and IBM is striving to make it a reality, leveraging our historical strengths in both structured and unstructured information management. Our customers have made it clear that they are very focused on leveraging their existing investments in both unstructured information as well as structured sources like databases and master data repositories. We see particular interest in combining structured and unstructured information in rich exploratory analysis environments that allow integrated navigation, correlation, and visualization of both types of data. Indeed, we’ve already seen these integrated analytics solutions go beyond traditional BI to deliver breakthrough business insight to customers in a variety of industries, from telecom to manufacturing to banking to healthcare.

FB. For us, this is not a goal. Exalead is delivering that today. Our technology is the only form of search software that affords a truly unified access to information, whether it resides on your PC, using our powerful desktop search solution (exalead one:desktop), on the corporate intranet (exalead one:enterprise), or on the Web (exalead one:websearch); all through a unified user interface based on our patented Search By Serendipity technology.

Further, it can search both structured and unstructured data in a process we call “hybrid vertical search”. Users don’t care about the data form. They need information. We let them find it quickly, easily and with certainty.

DG. Not only is it practical, but it is necessary for search to ‘work’ in the enterprise. Most applications today have a search function. However, there are so many applications and so many sources of content, that when findability is dependent on the user knowing where to look, the search often results in failure. Not only must a search system provide universal search across all enterprise content sources, but it must also span the user's desktop as well as search over public information (internet). Google's Enterprise Search solution was the first to deliver this comprehensive enterprise search across all types of content including unstructured, structured and real-time business applications as well as the desktop (via Google Desktop for Enterprise) and the Internet (via Google.com).

SF. Today, the market is heavily weighted towards US customers. In your own customer base, do you think that you will see a shift toward other regions? In what order is this likely to happen, and do you have some evidence that this is the case?
AB. At IBM, we have always had a worldwide customer base for search, analytics, and related services, and today our customer base is evenly distributed among Europe, Asia-Pacific and the US. Perhaps our distribution more accurately reflects demand as IBM is truly a global company, unlike many of the smaller players that have historically defined the search market. We continue to see healthy worldwide demand for search and analytics services, including strong growth in Japan, Australia and India. In Europe, we see a more varied pattern: in some countries, like Germany, there is a strong trend toward open source for generic search infrastructure on the one hand, and a trend toward advanced analytics for applications such as eDiscovery and legal compliance on the other. In the UK, we see a strong demand for advanced solutions as well. Looking forward, we anticipate further growth in Europe driven by the promise of innovation built around advanced text analytics.

FB. The market is weighted toward US customers because there are more of them than in other markets. That’s not an expression of bias, simply an observation on the lay of the land. 

Organisations in Europe are stepping up for the latest forms of enterprise search already.  The only question is how fast the rate of adoption will be. My belief is it will be quick, given the number of proposals we’re fielding.

One posture that continues to be held by companies, both European and American is that the software needs to be ‘American’ to be good. My concern is not borne out of the fact that Exalead has its headquarters in Paris. Rather, it seems incomprehensible that in the shrinking world such a view can still exist. It is especially perplexing, not only for Exalead, but for search providers from Canada, Norway and the UK. The country of the software’s origin should be meaningless. How it can perform must be the primary criterion.

DG. We see growing demand for search worldwide. Although requirements differ somewhat due to regional systems differences, mobility proliferation and so on, we are seeing strong and growing demand from all geographies.

SF. Search applications are complex, even if the search box looks simple. What do you feel is the greatest technical challenge that your company faces as it develops the next generation of information access application?
AB. The greatest challenge in any information access application is to deliver the most relevant, synthesized information via the most effective possible user experience. As we move to the next generation of information access, all types of information must be brought together and integrated in a way that enables the everyday business user to cull the key insights they need to be effective. The technical challenges in doing this are immense and range from information integration to security to entity extraction to rapid interactive analysis to relevancy calculation to Web 2.0-driven visualization and interaction metaphors. But most of all, it’s solving these problems together in a way that enables a business’s core processes and activities – a level of innovation that IBM has been hard at work delivering for decades.

FB. Most search applications available today are buckets of APIs that require lots of time and money to be installed. 

Exalead’s enterprise search software is very much the next generation application.  It can be installed with little to no professional services. It can be up and running in a matter of days, not weeks or months. Further, it handles all the complexities of the most involved enterprise infrastructure in a straightforward manner. And the user experience is such that usage rates of Exalead-equipped search are multiples higher than the prior methods.

Today’s search market is growing rapidly, but it is also getting more specialized. Entry-level search solutions will be sold by well-established players in volume. High-end solutions remain both highly complex and very expensive but a new market is emerging, the market of hybrid vertical search engines, with a requirement for mission critical solutions that are not necessarily as complex or expensive as today’s high-end solutions. This is the ‘long tail’ of the high-end market. It is simpler and packaged, yet full-featured with high-end solutions that can address this market. We believe Exalead is best positioned to serve this promising emerging market.

Information to help businesses make decisions does not reside solely within their “walls.”  Much needs to be searched and evaluated from the outside. Further, much of what an individual needs sits on their own hard drive. For search to deliver real value all possible sources need to be searched, easily, at once, with findings provided in a single set of results. Further, those results need to be sorted through by narrowing or widening the discovery path, not by scrolling through page after page of lists.

DG. The single biggest blocker to broader deployment of enterprise search is the lack of standardisation of content authorisation systems. A key requirement in enterprise search is respecting the access controls and policies of content systems and business applications. Unfortunately, these systems were not designed with search in mind. Therefore delivering a search system that provides an easy to deploy and highly performant search experience can be difficult. Many vendors sacrifice performance or require complex deployment scenarios and redundant access control systems to solve the problem.  At Google, we continue to innovate and partner with key players in the security and content systems space as well as pushing the standards forward to deliver what users and IT administrators want: relevant, comprehensive, secure and fast search that is easy to deploy and easy to maintain.

Aaron Brown, Ph.D. is Program Director, Content Discovery for IBM Information Management Software, where he is responsible for strategy and planning for IBM’s Content Discovery software including enterprise search, information discovery and content analytics offerings. His primary area of focus is on the introduction of new Content Discovery products and offerings, and evolving the Content Discovery portfolio to leverage key market growth opportunities. Previously, Dr. Brown was a senior program manager responsible for strategic alliance development within IBM Software Group, where he created cross-brand partnership strategy and developed and managed IBM's strategic relationships with key technology partners. Prior to his work in Software Group, Dr. Brown was a researcher at IBM’s Thomas J. Watson Research Center, where he created and led foundational work in the areas of system manageability and system complexity reduction, contributing to ongoing improvements in IBM software products and strategic outsourcing services.

Dr. Brown joined IBM in 2003 and holds an A.B. in Computer Science from Harvard University and a Ph.D. in Computer Science from the University of California, Berkeley.

François Bourdoncle is a pioneer of the search engine software market and President and CEO of Exalead. In 2000, he co-founded company with the goal of revolutionizing the search engine software market by providing users with a unified technology platform to access information in the enterprise.

With a deep background in engineering and R&D, Bourdoncle is truly passionate about technology. He can be credited with the development of Jazz, a programming language designed for synchronous digital circuits, as well as Exascript, an object-oriented XML programming language based on JavaTM technology.

Prior to Exalead, Bourdoncle was a researcher at both Digital Paris Research Laboratory and Digital Systems Research Center in Palo Alto, Calif. While at Ecole des Mines de Paris, he collaborated with Patrice Bertin, an Exalead co-founder and the company's chief technology officer, on the LiveTopics project for AltaVista. Bourdoncle holds a PhD. in computer science from Paris-based École Polytechnique. He is the author of several various technical papers and publications, and is a frequent speaker at industry events.

Dave Girouard, Vice President and General Manager, Google Enterprise. He is responsible for all aspects of Google's enterprise business including sales, marketing, product development and customer support. Prior to coming to Google, Girouard was senior vice president of marketing and business development at Virage, a provider of multimedia search and content management software. In this role, Girouard oversaw product management, product marketing, business partnerships, lead generation, sales force readiness, and press and analyst relations. Girouard also founded, developed and managed Virage's application services business, which was launched in June 1999 and led to the company's initial public offering in June of 2000. Previously Girouard held positions at Apple, Booz Allen & Hamilton and Accenture. He holds undergraduate degrees in Computer Engineering from Dartmouth College and an MBA with highest distinction from the University of Michigan.

Sue Feldman is Research Vice President of IDC’s Content Technologies Group, which specializes in research on search and discovery software and digital marketplace technologies and dynamics. Feldman won the 2003 James Peacock Research award at IDC for her work on modeling and forecasting the search and retrieval technology markets and her current work includes creating an interactive model for the digital marketplace. Prior to joining IDC in 2000 Feldman was President of Datasearch, where she consulted on new retrieval technologies such as natural language processing, search engines, usability of online systems, and digital libraries.