It isn’t just schoolteachers who disagree about how best to teach and benefit from the “three R’s.” But the three R’s of business intelligence (BI) are not Reading, ’Riting and ’Rithmetic; they’re the Right information, at the Right time, in the Right person’s hands. Those three R’s, particularly the last one, have always been BI’s holy grail, but the producers and consumers of BI technologies have generally focused on an exclusive set of end users. International Data Corporation’s Dan Vesset says that the main BI market drivers are performance management and compliance — two obvious executive-level concerns.
He notes, “Executives and managers are looking for more data and information to support their decision-making processes.” The source of the disagreement is often over exactly how BI’s three R’s will be defined and delivered. The debate notwithstanding, IDC’s March 2006 sizing and forecast for the BI tools market (which includes query, reporting, analysis, data mining and statistical tools) was $5.8 billion worldwide and is expected to grow at about a 10 percent compound annual growth rate (CAGR) over the next five years.
Traditionally, executives and managers would be “fed” their information by a team of BI analysts in standardized and custom-built reports, executive briefing books and information systems, and the like. Intelligent dashboards have streamlined the flow of information to the traditional constituency, but BI professionals still have to perform the underlying analysis behind what is presented, and IT professionals have generally had to help build the dashboards or reports that the executives consumed. That’s beginning to change now, especially as BI and enterprise search technology start to combine, promising greater transparency into how businesses operate for a much broader audience inside — and perhaps outside — companies around the world.
BI for Everyone
Matt Glotzbach, head of products for Mountain View, Calif.-based Google Enterprise, says, “The real focal point from our point of view and, we believe, from the perspective of our BI partners is the realization that BI systems contain much valuable information about a company’s process, workflow, etc., but that information is untapped by the average employee. If you look at the traditional BI implementation, a very small number of users actually access the BI system and use its data, so the question becomes how to make this extremely valuable information accessible to the broader user base.” Google recently announced (April 2005) an initiative into BI around its Google Onebox for Enterprise technology. The initiative includes partnerships with a number of BI and enterprise application and professional services providers, including Cognos, Information Builders, Salesforce.com, SAS, and others.
IDC’s Vesset observes that “one of the big emerging BI trends is providing access to more users within the organization, not just to the analysts who have been the traditional users of BI. As the user population grows, new interfaces will be required to help out those ’less analytic’ users to gain access to BI data.” Vesset thinks that the initial impact of the combination of BI and search will be on information access, but over time, the BI/search combination “will eventually impact analytics, when more organizations will look to combine both structured and unstructured (e.g., text) data in their analysis of corporate and external data.”
Still, there are some who look at the combination of search and BI as little more than hype. One dividing issue seems to be whether search functionality is going to be capable enough to help users refine questions (as they would when using the BI toolset or BI-enabled application by itself) or whether they’ll just be searching a catalog or some other “store” for reports that have been generated previously. In the latter case, search may just be a way for vendors who have not yet integrated their products to achieve some limited integration from a user interface point of view.
Search can, in fact be seen as a lightweight means of integration/federation across the enterprise, not just for BI but for all types of systems. “As companies struggle with giving users a unified view, search can play a powerful (role) in providing that,” claims Google’s Glotzbach.
Gartner Group’s Kurt Schlegel observes that there is real value in providing broader access to existing reports from a variety of systems: “Companies have hundreds, thousands of reports. Search will make it easier for the average user to find them.” But then he gets at one of the other issues dividing the BI community. “I it isn’t just reports. There are also metrics, hierarchies, and other objects that are stored in BI metadata that search should make it easier to find.” Schlegel notes that “for the BI and search combination to work well, it needs to be built on rock-solid BI metadata.” Vendors that take advantage of search most effectively probably already have a solid metadata framework, he thinks.
Metadata, and the master data management initiatives underlying a rigorous approach to BI, are clearly important issues. Without the rigor they bring to defining key terms, there may be more than one version of the information users are looking for. These versions may be differentiated on the basis of the time the information was generated — or it may be that the information is different because the underlying definitions of terms used to perform a search are themselves different. In either case, there is potential for problems. ”What we’ve heard time and time again from our customers is that without a single, accurate version of the truth, search is essentially useless,” says Tobin Gilman, senior director of product marketing for Hyperion Solutions Corp., a business performance management software company in Santa Clara, Calif.
The Promise of Search
But even skeptics agree that the combination of search and BI has some promise. “Search holds potential to be one of the interfaces, but there’s the element of translating a simple search phrase to a really meaningful BI search,” notes Dave Menninger, VP of worldwide marketing and product management for Westborough, Mass.-based performance management software company Applix. “It is a perfectly reasonable approach to help search through a catalog of reports. If you named a report correctly, search will help find it.”
In fact, that’s one of the key benefits of the BI-and-search combination. From a user access point of view, the combination means that the average user won’t have to go through the large amount of training that an expert BI user would have to go through. “It gives them a fast front door deeper into the BI system,” says Google’s Glotzbach. Christina McKeon, manager of business intelligence marketing for Cary, N.C.-based SAS Institute, points out that one of the overriding factors is that most of the potential users of BI in a company are going to be casual users. “If they’re the majority, you have to come up with easier ways to make BI accessible,” she says. “This is about the business user, not the tech user. Everything is aimed at accomplishing the business goal.”
But the first search may produce the best result. “With BI, most times you know the data sources you are searching for and are using the BI tools to filter/refine your views on those data sources to discover trends or gems of information. With search, you usually are unsure where the data is stored or even whom to ask to find the answer to your question. And in search, the search itself is not important but rather, the results you get back. You want actionable results that you can quickly drill into, refine and reuse as you see fit,” explains Francois Ajenstat, Microsoft’s SQL Server group product manager.
The BI and search combination also promises to help build a bridge between unstructured data and the structured world of BI and business applications. “In a company today, just on the BI side, you build up lots of databases you can report on, and hundreds, if not thousands, of reports based on those databases. As those proliferate, you may not know where they exist,” says Kevin Quinn, VP for WebFOCUS product marketing for New York-based Information Builders, Inc. (IBI).
Quinn gives this example of how it might work: A searcher has heard some complaints about a product sold by a consumer electronics company, but he’s not sure exactly where to look to find out what the complaints were, how they were resolved, or if they were part of some larger issue. In looking through search results, he might come up with a particular product ID. He might then want to find out what’s in inventory and what’s in the sales forecast for that product. Then he can start to refine his searches with results coming from the BI system and related business applications and the unstructured world of comment fields in a CRM database or user comments on a customer feedback web page, for example.
Is this combination likely to replace intelligent dashboards and critical performance indicators? IBI’s Quinn doesn’t think so. “Initially it’s an adjunct, but there’s a good chance that a large portion of what’s in the BI environment may be just found by search.”
Sanju K. Bansal, vice chairman of the board, executive vice president, chief operating officer and secretary for BI software firm MicroStrategy in McLean, Va., sees the picture differently. “What I can see as clear as day, looking out five years from now, is that every Fortune 500 corporation will have hundreds, if not thousands, of BI applications in production. Wherever there is a database in a corporation, people are thinking about what kind of reporting/analysis to apply to get at that data.” His own company has several hundred database instances, and a large financial services company might have thousands, or even tens of thousands, he asserts: “Each one of them represents a reporting or analytic opportunity.” While BI is great for accessing structured data, BI is generally not good for accessing unstructured data — even if you can attach it[Question: “it” OK here?], you can’t get to it, Bansal says.
For MicroStrategy, it seems that the question is a matter of the firm’s and its customers’ business priorities. “We think that the unstructured search area is ripe for R&D, but there’s much more value in structured search. We believe that there’s another five to 10 years of incredible activity in the structured BI world that will be extremely high value use cases.”
Definition Is Key
Cognos’ Don Campbell, VP of platform strategy and technology, has yet another perspective. His Burlington, Mass. firm has partnered not only with Google, but also with IBM, Fast Search & Transfer, and Autonomy. “Enterprise search with keyword searching gets you only so far,” he says. “The next level of value comes in when you understand that a report has been described in its metadata form, and yet it doesn’t have to exist with those data items pre-run. You can run that report right now, but fill in details of what you’ve searched for.”
The key thing is that the underlying metadata artifacts have to be well defined and well described. “When you push that information out to more people, there has to be a realization that the information has to appear in a form that’s consumable by those consumers,” Campbell says, “so indeed you have to take care when describing those artifacts that will be used by the broader community.” He also notes that the issue isn’t particular to search. “As any technology’s audience grows, you would see the same effect.”
Despite its promise to help expand the community of BI users out into the more general user population, search may also have some value to the expert BI analyst, particularly with respect to the question of how to improve results relevance. For the expert BI user class, search may help with navigation around the BI system. “You may already be an expert, but search reduces the time to find what’s there,” Campbell says. “It also helps, for example, in an analysis session when you come to a number worthy of more investigation.” Cognos’ approach takes the context of where the user was in the search and returns other results that are based on that context.
When more casual users come into the BI realm via search, the information presented to them can appear in a couple of different forms. You might see a results list like the one you’d see with a basic web search, or you might see a chart, a graph, or an interactive report. Ideally, the system will figure out what your underlying question is.
Matt Eischner, VP of strategic development at Cambridge, Mass.-based Endeca Solutions, an enterprise search vendor with a number of customers who have built BI applications, believes that the combination of BI and search ought to allow user exploration of critical business data in real time — without involving IT. It combines interactivity and relevance. “You can start your analysis in any combination of business intelligence and search; they all access the same information,” he explains. Furthermore, analysis can be integrated with the search process, which means that users can perform “unanticipated investigation.”
Regardless of how it will be approached, the ultimate goal of the BI and search combination is better decisions and better business performance. Rob Lancaster, VP of channels for the U.S. arm of Norway-based Fast Search & Transfer, another enterprise search provider, emphasizes that “preparation for convergence is not about data management, file management, or rich media management, but about the common goal of getting the right information to the right person at the right time. Search provides one approach to providing broader accessibility to critical information.”
In a sense, search and BI combined refines the basic concept of querying and reporting. “It introduces the concept of drilling sideways versus traditional drill down — in order to get answer to question,” Lancaster says. The result from the BI environment might well be the basis for a lateral search of attached data or other data sources outside the BI environment. “The combination can bring content together that helps give you a more complete picture,” he adds. “The transparency of content with respect to source and destination — search really does that well.”
Ultimately, that transparency may be what matters most to both the professional and the casual business user of BI and search technology.
Damian Rinaldi, a former Wall Street software sector analyst, IT market researcher and trade publication editor, is now an independent writer and consultant. He can be reached at dvrinaldi@yahoo.com.
