Read the Original Article at http://www.informationweek.com/news/showArticle.jhtml?articleID=18401672
The market for enterprise data warehouses and business intelligence has reached $4 billion and is expected to grow to $10 billion by 2008, according to Meta Group. Karen Parrish, VP of business intelligence solutions at IBM, talked with Business Intelligence Pipeline about what's driving BI demand, the challenges involved in BI deployments and how IBM plans to keep up with the market.
Business Intelligence Pipeline: Can you quantify the extent to which the demand for BI is growing, and discuss what's driving it?
Parrish: If you compare it to about 3 years ago, it's exponential growth. We're seeing things in the marketplace driving this growth. The first has everything to do with regulatory compliance. For example, HIPAA was a major compliance requirement for healthcare and it drove the acceptance of business intelligence. In the post-Enron era, we're finding that Sarbanes-Oxley is also driving the need for business intelligence solutions. In financial services, we're finding that compliance with risk management standards such as Basel II is a major driver for business intelligence solutions. Basel II is about the management of your cash portfolio and how you determine risk.
Business Intelligence Pipeline: Has RFID become a major driver? If deployed as it's envisioned, radio frequency identification will produce a lot of data for many companies.
Parrish: RFID is indeed going to be another driver of business intelligence. I agree with you that the mounds of data collected through RFID will need to be analyzed. We absolutely believe that there's a growth in demand around information management. Here I mean both data sitting in the warehouse and unstructured data, like that in content management applications. I think there will be a convergence of these data that will require deep analytics. RFID will certainly be one of the technologies that will drive this convergence.
Business Intelligence Pipeline: Do you see recurring pain points that come up for companies attempting business intelligence initiatives?
Parrish: Yes. There's one very specific one. The fundamental pressure point is that there must be a marriage between the line of business and the IT community. We have found that when the line of business deploys a business intelligence solution in the absence of IT, they usually don’t get their biggest bang for the buck. We've also found that when IT deploys in the absence of business, it becomes a sort of "build it and they will come" kind of scenario. Neither will work. IBM really does have success in bringing the line of business and IT communities together to talk about the problem. They can work together to design a solution that meets the needs of both sides of the business. When that pain point is removed, our customers have much greater success.
Another pain point I would mention deals with the idea of organic growth. Many of the enterprise data warehouses out there grow organically. It's very important that our customers plan for that growth. If you build the infrastructure the right way for an enterprise data warehouse, it will grow with you naturally. If you don’t, it could be very expensive to grow. So we spend a lot of time with our customers building the infrastructure in a way that we believe will grow with them. This means having an open environment that they can plug into their existing infrastructure. It's also about accessing data where the data sources reside, and it's about providing a middleware and hardware stack that can scale when the customer needs it to scale.
Business Intelligence Pipeline: Is it really necessary to build an enterprise data warehouse to carry out successful BI? What about smaller companies that lack a data warehouse?
Parrish: I do not believe that business intelligence is only for the high-end organizations with enterprise data warehouses. I believe that a data warehouse is usually the way to go for a big company because of the mounds of data they deal with. But I absolutely believe, and it's part of our strategy, that the smaller customers will have to be adaptable if they're going to survive in their markets. Business intelligence is going to help them do that. They need a single view of the customer just like the big guys do. They have the same requirements, just without the big IT staff. They also don’t always have all the data themselves. That is, sometimes they have to pull it from elsewhere. We have an edition of DB2 Data Warehouse for mid-tier customers who want the same functionality at a lower price, and want it to be more turn-key. We've already seen major growth in the SMB space for business intelligence.
Business Intelligence Pipeline: You’ve said before that BI and spreadsheets will coexist in the future. What role will spreadsheets play?
Parrish: They're going to play a huge role, especially in the SMB market, because I think spreadsheets are the way that users work today. Until people use new technologies that are better than spreadsheets, why would we replace them? Our goal is that no matter where the data sits, we want to be able to drill down and analyze and help you get the answer you want to get. We clearly understand that data is going to reside elsewhere in the organization besides the enterprise data warehouse, and we must provide ways to get to that data. It's our expectation that spreadsheets will continue to play a role for a long time.
Business Intelligence Pipeline: What kinds of enhancements will we see to DB2 to support BI?
Parrish: You're going to see that our strategy on DB2 is to move the analytical function deeper into the database. You're going to see investment in analytics around DB2. We will create the necessary links to the analytical application providers so that when they work with DB2, it's far more sophisticated and better-performing than the competitive databases. We want to make sure that when Siebel works with DB2, that relationship between Siebel and DB2 offers a better value proposition than when Siebel works with Oracle.
The second investment will be around autonomic capabilities. These are self-healing, self-configuring capabilities. It's critically important. Those customers using business intelligence applications are using mission-critical data. That data has to be accessible at all times. We want the database to be self-healing and self-optimizing in nature. For example, we want the database to be optimizable for certain queries that are placed against it by a BI system. You'll see heavy investment in Query Patroller, which is a capability of the database to optimize the queries for the best performance results.
You'll also see that we're going to deliver abilities that complement the database. There's DB2 Cube Views for instance. This is a bridging technology. In today's environment, people who model databases model for a particular reporting tool. If they're reporting through Cognos or Microstrategy, there's a correlation with the way they model their data for that tool. We've simplified that with a metadata bridge. You can build your model in your metadata bridge and it can report out to any tool. This saves time and increases the speed with which reports are produced. You model once, and you use it everywhere. Much of the investment you're going to see from us is going to be around DB2, because we believe that's a differentiator for IBM -- the way we grow the database and leverage the database.
Business Intelligence Pipeline: Right now, which industries are leading the business intelligence charge?
Parrish: Clearly financial services, which is, as I mentioned, driven by regulatory compliance issues. I also see government acceptance of business intelligence for varied reasons. Government is interested in things like crime prevention, understanding certain activity levels of certain individuals. Even Social Security wants to understand more about their constituency and how to offer services to them.
Also healthcare. Compliance is important for those companies, but we also see growth stemming from their desire to offer more efficient protocols of service to a patient by understanding how a patient reacts to certain treatments. For example, healthcare companies are interested in lowering costs and increasing health care. If I go to the hospital and complain of chest pain, there's a protocol that helps guide my treatment. Perhaps those protocols can be improved. Those types of things in healthcare are involved not just on the cost side, but in providing better service.
I'm also seeing growth in retail. RFID has been a technology that retailers have struggled with for a long time because they’ve wanted to use it, but it's been prohibitive for one reason or another. Retail is an industry that has always tried to get a single view of the customer because they want to know how to market to them.
And I would say across the board, if you look not at specific industries but at role-type solutions, there are drivers there too. If you look at business process management, it's not an industry-specific process. Those types of solutions are about making improving operational environments and lowering costs within those environments. Operational analytics has become a major growth area for analytics. How do I take the processes in my business and make them more efficient so I can reduce cycle times, or enhance the view of my product line? They help businesses become more on-demand in nature. So we're seeing major growth there. Everybody wants to be able to predict how customers might buy or how their processes can become more efficient.