E-BI: Revolution, Not Evolution?

To adapt to zero-latency demands, BI must change radically, says Nquire's Larry Barbetta

Here in the post-bonanza internet age, it's hard to think of a software category more betwixt profit and peril than business intelligence (BI). On the one hand, the data-driven nature of Web-based customer interaction and business-to-business (B2B) integration cries out for solutions that help decision makers analyze the present and model the future. On the other, users' experiences with the Web's amazing array of content, right at their fingertips, have changed data access expectations forever. Unfortunately, these expectations are beyond what many current tools - and IT departments - are prepared to deliver.

Challenges open fissures in the market, through which new competitors burst. Despite their (in some cases) massive investments in BI, data integration, and data warehousing, IT leaders are coming to realize that the Internet means more than putting a simple enterprise information portal (EIP) in front of the data. Once the EIP is there, users want greater access and richer, more flexible analytical capabilities.

I talked recently with Larry Barbetta, CEO, president, and founder of NQuire Software, one of the more interesting new arrivals on the BI scene. Founded in 1997, the company emerged from its stealth phase to introduce NQuire Server Suite in late 1999. It has been gathering press kudos and charter customers based on its approach to widening and deepening the impact of Internet-based "information intelligence," as the company calls it. In September 1999, NQuire announced that Royal Bank of Canada would be licensing its software suite to improve dynamic access to the bank's array of data warehouses, operational systems, and data marts, all of which run on different database software. Royal Bank's goal is to provide a wider selection of employees the access and analytical capability to build a complete picture of customer activity across different products and accounts.

Barbetta and his principal colleagues at NQuire are veterans of the BI industry, having played key technical, sales, and marketing roles at companies like Cognos, Information Resources Inc., IBM, and Platinum Technology. His own origins go back to the beginning of the "decision- support" industry when he was part of the start-up team for Decision Support Services Inc. This company was acquired by Metaphor Computer Systems, which itself launched the careers of many of the leading lights of the current BI and data warehousing industry. In 1987, Barbetta founded Prodea, which was later acquired by Platinum Technology when that company went on its incredible shopping spree a few years before being acquired by Computer Associates. Barbetta served as Platinum's senior vice president of BI.

I asked Larry how he viewed the challenges the BI market is facing and what is motivating his company's foray into the next generation. "Historically, you've had this sweet spot in the BI market centered on data warehouses, data marts, and analytic tools driving strategic planning and analysis," he said. "This has helped people answer, for example, how do I allocate my budget? In most companies, a relatively small number of people do this activity, generally using aggregated information presented at a very macro level. Strategic planning and analysis is not incredibly time-sensitive; in other words, I don't have to know what happened yesterday or the day before - last month is good enough. You might call this CYA reporting.

"What we're seeing now is a real push to use information to support decision- making throughout the value chain and broader business network. To expand the reach of BI tools, we've seen the development of things like EIPs and the ability to create standardized reports that are delivered via an intranet or extranet. But this type of program is just the first attempt to respond to the new dynamics. Reach is just one axis; the other axis is richness."

Strategic to Operational

Barbetta has NQuire focused on bringing far greater richness to a wider spectrum of decision makers. "Extending the reach of information to new audiences fundamentally changes the implications for richness. The information needs shift from macro to micro, from aggregated to detailed, and from relatively time insensitive to absolutely time sensitive. If your data warehouse or data mart is updated to the end of the month - well, your account rep might have a lot of customer issues that he or she needs to know about before the end of the month.

"Richness is key if we are to move from strategic use of information to an ability to affect the operational responsiveness of many users. Let's say your sales force is getting ready to call on a customer. You want your team to have the information to support that interaction. The salesperson needs to know specific information about the customer, not some aggregation or average about all customers. With the current state of reporting, we've enabled the delivery of standardized information. People take the standardized reports and go through them with highlighters to identify what's going on in accounts in their region. We need to move to a more personalized level; an account representative selling to 22 accounts shouldn't have to go through some five-foot stack of data.

"Data warehousing initiatives are at the center of the new information intelligence economy," Barbetta said. "However, being completely and solely reliant on data in the warehouse is not going to enable companies to differentiate themselves. Reaching out and embracing information beyond the boundaries of a single warehouse is a got-to-happen thing. The demand is to provide detailed, micro-level information that is very time sensitive. We need to take data warehousing initiatives to the next level and bring analytics and intelligence to users who are not yet information-enabled - despite the millions spent on data warehousing. Rather than ask, how do we get the data warehouse out to everybody - we should be asking, what information do people need to do their jobs?

We are moving into a world of dynamic, B2B exchanges where perhaps thousands of suppliers will need up-to-the-minute information about inventory and pricing. 'There's no way you're going to get everyone to put their data into one huge data warehouse,' Barbetta said. 'Think about it, if we decided that the way to respond to the Internet was to put all the data out there into one enormous data warehouse, by the time you finished updating the system, the data would be wrong. That's a simple thought, but it has profound implications on BI technology. We apologize if people don't want to hear this, but it's true.'"

Virtual Integration

Moving from a strategic planning purpose for BI toward one that supports operational decision-making makes it imperative that users can get at data from a spectrum of sources. "We need to enable more complete gathering of information," Barbetta offered. "Right now, we have data marts that are great and useful. However, they are mostly functional or issue-oriented - they are becoming analytical silos."

However, most approaches to data integration and heterogeneous data access involve building a physical staging edifice, such as a data warehouse. To Barbetta, this approach is flawed. "You're building latency into your cycle. People have enough places to keep their freaking data! We felt that if we gave you yet another managed persistent data store - well, the world needs that like it needs more taxes or politicians.

"What we did is cast a logical umbrella over the physical locations of data, to keep users independent of how the data is designed or structured. We can break a request down into pieces and get at only the information the user needs - if you just need today's data, it can go into the production system and pull out just that data for one customer. We aim to be highly efficient in how we interact with the underlying systems." The company's virtual integration approach also keeps it out of head-to-head competition with companies that do provide physical data stores, such as Oracle, IBM, or Microsoft.

The company's "secret sauce" contains a patent-pending process for breaking down logical business requests and bringing back detailed data. The NQuire Server Suite also takes advantage of an Internet cache architecture to speed up response time across a range of users. It also appears to take advantage of modern component-based adapters and pipes to create native connections to data sources and XML-based data.

What NQuire's emergence tells us is that we're entering a new generation of BI tools that will help IT bust out of its traditional focus for this type of solution and turn its attention to providing wide and rich information access via the Internet. "History tells us that any time you have a 10x change, it creates discontinuity, and you need to rearchitect. In this market, we've got 10x more users times 10x more data. We're going to see radical transformation and change in the BI space."