The enterprise software market has consolidated and changed dramatically over the past decade. Steve Mills, senior VP and group executive, IBM Software Group, says the company is sticking to its middleware-centric growth strategy. IBM's software revenue has doubled over the last decade to $22 billion, fueled by acquisitions of more than 65 companies. But is the company now ruing a 12-year-old decision to stay out of enterprise applications like ERP and CRM? Mills talked with Intelligent Enterprise about customer needs, the competition, and the opportunity to help companies take advantage of business analytics.

Between the launch last year of the Business Analytics & Optimization practice and the acquisition of SPSS, IBM has made a huge bet on analytics. There's a compelling vision, but how far would you say we have to go before analytics can become a truly mainstream practice among your customers?

If you apply a broad definition to analytics, which would span everything from basic report writing and what you might characterize as fairly simple manipulation of data all the way to the most sophisticated mathematical analyses, the market is measured in many billions of dollars. It's not a brand new market, but it's a market that has now reached an inflection point because the technologies have advanced far enough and price points have come down.

There's also a sense of imperative on the part of customers. They have obtained a lot of value from the investments they've made in the past, but they still don't feel as if they have enough visibility into changing conditions within their company and within the markets they serve. They believe that they can use computing to optimize the way their business operates, to optimize the way they interact with their customers and to find new opportunities that they couldn't see before. Companies always talk about cross-sell and up-sell, and they wonder whether they really know what their customers are doing and what they want. With better insight, they could sell more products and services. This is propelling businesses around the world to not just continue to invest in analytic-based solutions but to increase their investment.

IBM has made dozens of software company acquisitions over the past decade. Are there new areas where the company is planning to invest or where there are gaps you still need to fill?

After each acquisition you still have to make significant organic investments. We continue to invest, and we've built an ever-larger development organization. Of course, every year we're on a path to do anywhere from eight to a dozen new software acquisitions. I don't see that pattern changing. I'm obviously not going to declare what it is we're going to buy, but the past is prologue to the future. Our acquisition patterns have been very consistent over the last 15 years. The companies we have been acquiring have all fit within an overall architecture and strategy. Our business analytics investments are a logical extension of a many-decades-old information-management and data-management strategy that goes back to the founding of the company. There's a very logical progression of technologies, particularly around things we refer to as middleware, where we have built our software business and where we will continue to build.

The Software Group recently went through a reorganization . Does the creation of the Software Solutions group reflect a change in the market in terms of customer interest in integrating software and getting “time to value?”

No. What Mike Rhodin [senior VP of Software Solutions] is managing is things that we have been working on for a long time. Time to value has always been an issue. The Software Group reorganization didn't really reflect a change. The change took place back in the 1990s. The vision of the portfolio, what it could be and where it is today has been a long-term journey that began a long time ago. I've been building out the software business for over 20 years. The pathway is one that I've seen for a long time. The process of getting where we are today obviously took time. The push toward line-of-business relevancy and increased interaction began back in the 1990s. We expect to get larger still, so it needs nurturing.

As the resident gardener, eventually what happens is that your garden plot becomes so large, you start having a hard time getting all the flowers and vegetables to do what you want them to do because you can't tend the whole garden fast enough. No human being scales beyond a certain point regardless of how willful they are. You have to bring in others to help you. Mike Rhodin and Robert LeBlanc [senior VP of Software Middleware] are two of the more experienced software executives we have, and it was an opportunity for them to move up in the IBM Company. The reorganization we announced in the beginning of the year was a logical next step that we had been preparing for many years. There was a day when IBM executives would say, "We're not an applications company." But a number of acquisitions that have since put you into that space -- FileNet, Cognos. How would compare and contrast solutions and applications?

The message we've been trying to deliver around what we do and what we don't do has been one of trying to position our intentions with our independent software vendor (ISV) partner community. There are areas where we compete and areas where we have chosen to partner. Back in 1998, we tried to bring more clarity to vertical application providers as well as key horizontal application providers. That has given us the ability to create very effective partnerships around things like CRM, ERP, supply chain and dozens of verticals. Our independent software vendor partner community today is measured in the tens of thousands.

Some of the things that we now deliver, people would say, "Gee, those are really applications." Lotus Notes is an application to many users. Products like FileNet, which follows on to more than 30 years of image processing, and MRO, which does physical asset management and life cycle control, are perceived as applications. Some people view Cognos from an applications perspective, though historically, that market was defined by analysts as infrastructure.

We work with companies that are self-integrators. They love to buy the pieces and put them together, and that's fine. But that model is transitioning toward outcome-based decision-making and purchasing... Our technology delivers important elements of the solution, but there are often third-part application companies that add to that solution. No one vendor delivers everything required. The average large business, if you went into their compute centers around the world, runs 50,000 to 60,000 programs that are part of 2,000 to 4,000 unique applications. Certainly they will have at least hundreds of applications. It's enormously diverse, and there are many elements that make up systems. If we take ERP as an example, if you look inside, it's not just the branded ERP package, but also many other packages and offerings.

You mean it's a suite?

No. It's more than the vendor's suite as well. Nobody runs SAP stand-alone, and nobody runs any of those Oracle applications stand-alone. They run them in concert with other applications, which are often not delivered by the vendor that produced the core application. The rhetoric one often hears is that customers are somehow making singular decisions. [Oracle CEO] Larry Ellison loves the story that if you just bought Oracle applications, you wouldn't have to buy anything else. That's something that Larry says that's just so untrue as to not even make any sense.

Back in 1998, saying "we're not an applications company" put vendors like PeopleSoft, JD Edwards, Siebel and many others at ease, but is there a different position today in the wake of so much consolidation?

No. And the reason is because the software requirements surrounding these applications remain quite significant. I'm not talking about midsize and small businesses that can focus on very few packages. But in the enterprise, for every dollar invested in ERP, there will be five dollars of investment made around that ERP package to get it fully implemented, integrated, scaled and running effectively. Those five dollars go into hardware, services and software. That deployment needs a database. It needs systems management. It needs development tools. It needs integration software. It needs messaging infrastructure. Application-to-application communications have to be enabled through the middleware infrastructure below the application layer.

All of that is not native to any one application. Furthermore, there are so many different connection points, we're the only ones that have a portfolio that can deal with all that. That's the value of this very-large integration portfolio and the many decades of experience that we have around managing data, managing transactions, managing message flows and so on. Do businesses view all that as plumbing? Is there a customer perception -- particularly with ERP -- that applications are the mission-critical piece?

The things that we deliver are foundational in nature. You have to have a good foundation, or you can't scale. We're the ones that provide horizontal integration, deep levels of transactional integrity, sophisticated optimization around data. Our perspective of the customer's problem spans not just the individual processes but the aggregate of all processes. We're the company that has the widest and deepest view of what the company is trying to accomplish with all the IT assets that they have. We know their legacy environments. We know all their hardware platforms. We know Unix, we know Windows, we know Linux and we know the mainframe world. We know storage and telecommunications. There isn't anything they do with IT that we don't know. And we have a perspective on how to optimize all of those things....

Do you see any threat from Oracle's new single-stack strategy with the acquisition of Sun? Or to ask it another way, what's really driving customer technology selections these days? Is it hardware, middleware software or applications?

It's money. That's the No. 1 motivator. And money is not a single-dimensional factor because there's short-term money, long-term money and money described in broader value terms versus the cost of a product. The surrounding costs are far in excess of products. Every month, customers convert from Oracle to DB2. Why do they do that? Well, Oracle is expensive. Oracle tries to use pricing power to capture a customer and then get the customer to keep on paying. Oracle raises its prices constantly. Oracle does not provide a strong support infrastructure. There are many customers who have decided to move away from Oracle across a variety of products because of those characteristics.

Oracle is not a very strong technology company, but they are a very high-testosterone company. They love to beat their chest and talk about how they always win and never lose. They're very declarative in that sense because their perspective is that lies that go unchallenged become the truth. If you scream loud enough, it's hard for anybody to get a word in edgewise.

If you focus just on the highest-value pieces for large enterprises -- the data integration and management, the BI, the decision support, analytics -- aren't enterprise application providers in a position to influence all sorts of important software selections?

If we try to distill down what companies do with IT, automation of business process is just one dimension. That is the core of what ERP companies do. They are an integral part of process automation. Companies today think of not just applications in isolation but process from end to end, and they are striving toward business process integration. In financial services, you hear the term straight-through processing. If you talk to any large business, they will tell you they have a large number of application assets that support that process dimension.

Analytics is another dimension. It can be related back to processes in that you can collect data from, generate reports on and analyze processes. But beyond processes, companies want to understand other aspects of the very nature of the business that they are in. If I pull data together that allows me to analyze my customer base, for example, can I understand patterns of consumption? Can I get a better sense of demographics and buying behavior? Can I formulate superior up-sell and cross-sell strategies that improve the value of each customer? That's just one example among many in the world of analytics.

Another example might be an energy company that generates and transmits electricity. They want to optimize the operation of their infrastructure. They know that if they are monitoring it effectively, they can collect a lot of data around both generation and transmission. In doing so, they can look at various aspects not just of load but of wear. And if they can identify wear patterns and do predictive analytics on substations and local transformers, they can do preventative maintenance and avoids outages. They can then reduce the amount of capital and operational expenditures in sustaining that infrastructure. They avoid emergency service, so they can have fewer trucks and avoid excess crew overtime.

That all started with instrumenting a transmission grid. You then have to assemble the data and do time-based correlations of events across that data to analyze what's happening on the grid. Finally, by applying historical pattern information together with sophisticated analysis, you can do a better job of managing that environment, and there's a huge economic payback to you as a power company.

In that entire scenario, I didn't say a word about ERP. You could not have said, "SAP is doing that," because they are not doing that. There are thousands of domains out there that are touched by a whole variety of companies and businesses that don't relate to ERP. It's not that you couldn't create an interesting connection between some aspects of ERP environments. But it's a big, big world, and the amount of money that will be spent on analytics related to physical world instrumentation, data collection and optimization will dwarf the traditional back-office markets for business analytics.