Industry Perspective: Vijay Gurbaxani on Utility Computing and ROI

In Part 2 of csc.com’s interview with University of California, Irvine, IT professor Vijay Gurbaxani, the main topics are utility computing, return on investment and BPO/BPM. Also read Part 1 of this interview.

csc.com: What’s your take on utility computing?

Gurbaxani: Utility computing is a good idea in theory, but I think we are far away from it in practice. Think about what the term “utility” really means. If you think of a utility like electricity where you can take a shaver and plug it in just about anywhere in the world — a true commodity — I just don’t see IT in the same way. If you look at the question in terms of "are we even exploiting the economies of scale for IT within a single corporation?" even then the answer is usually "no."

Too often you find, for example, applications that run on a single, unique set of application servers. In many cases, we don’t even share workload between applications. This is mainly because we want to guarantee availability and the cost of hardware is low enough. We don’t really want the economies of scale. We’d rather have the availability and we need to protect the service level of that particular application.

Think about a true utility, where we buy the same application as others are buying. There’s a supplier who provides us with application services and is providing them to hundreds of other clients as well. That’s a long way off. We don’t have the technologies, and I’m not completely convinced that most companies will buy into models like that yet. Such an offering is more likely with services such as networking and communications systems, which are the closest we get to utilities in the IT business.

We don’t even have the billing systems to pay on a usage basis, so can we really charge per megabyte in the case of, say, storage systems? It’s very, very hard to pull off. I’m not convinced that customers will buy into it.

The ASP (application server provider) model, which is another example of utility computing or on-demand computing, hasn’t really taken off. Companies want control over their applications.

There will be some traction around business process outsourcing. But that presents a huge integration problem. As we keep outsourcing business processes — and these are owned by different owners — HR, for instance, may go with company A, accounting with company B and so on. We have spent the last 10 years building integrated enterprises and now we are parceling processes out to different specialists and providers.

We really ought to be thinking about what the optimal business model is for our corporation in the future. Then think about outsourcing within that context rather than making an outsourcing decision tactically and getting stuck with a problem we have to fix later on.

			True utility computing is a long way off. We’ll see baby steps along the way.

True utility computing is a long way off. I think we’ll see baby steps along the way.

csc.com: Do you see a trend of large companies focusing more on return on IT investment, as opposed to focusing on developing new technologies?

Gurbaxani: Absolutely. And I think it’s long overdue. One of the things that I’ve never figured out is why companies respond as much as they do to hype. Clearly, there was some over-investment in technology. The barriers to achieving high returns, by the way, often are not technological — the barriers are organizational.

Size can be a barrier because larger companies tend to be a little more rigid. But I also think industry structure matters — how competitive an industry is. When an industry is very competitive, it motivates companies to work that much harder to derive a return from their investment in all assets, and certainly IT.

csc.com: Give us a better idea of the importance of measuring the return on IT investment?

Gurbaxani: The importance of measuring return on IT investment is immense. The scale of spending on IT has become so large that this demand for justifying investments is long overdue. I’m surprised that it took as long as it did before CEOs and CFOs started asking the hard questions.

One of the issues is that the question is being asked of CIOs most often, and that is the wrong person to ask. We really need to ask the question of business unit heads. Business users need to own their investment in IT.

The big challenge, of course, is that there are tangible and measurable returns on IT investment and intangible returns on IT investment. Moreover, there are significant complementary investments to be made in management practices and in organizational change every time you make a large IT investment.

So holding the IT executive accountable is not particularly useful because it is rare that he or she is an end user of the system. Moreover, focusing on the IT costs underestimates what the true cost is. The overall cost is often higher than we anticipate because of all the organizational changes that we need to make.

Therefore, the first thing companies need to do is choose a methodology. It doesn’t matter what methodology is used, as long as it is reasonable, and a company uses it consistently. The theoretically correct approach is to use net present value or discounted cash flow analysis, which any business textbook will tell you.

The real question is, how do you estimate the payoff? Two things happen. For the tangible payoffs, companies often don’t exploit the value they could potentially derive. Second, how do you value intangibles such as improved customer service? We often hear the claim that we don’t know how to value intangibles; this needs to be examined. Most managers know when something’s working for them and when it isn’t.

You solve the problem by allocating the decision responsibility and the cost in the hands of the same executive, one whose incentives are affected by the outcome of the decision. Clearly, you’re never going to get a precise answer to what the return on investment is, but if somebody says, "I could spend this money on advertising," or, "I could spend it on a better IT system," let that person whose bonus check depends on performance make the call.

One approach that needs to be brought into wider use is a financial methodology called real options, which focuses on capital assets. This approach says, I am building the capability because I have options in the future to make follow-on investments, or to actually reverse my investment or stop investing. If you use the real options approach, you can sometimes get very different results than if you use what is sometimes called the naive net present value approach. This approach needs to enter the vocabulary more in IT management circles.

csc.com: Can you discuss how IT and personal computers have improved employee productivity?

Gurbaxani: The research is showing, without question, that PCs make employees much more productive. It’s not just saying, OK, we threw more capital at a person, which with computers makes a heck of a lot of sense because computers get cheaper every year. What’s much more important is, can we change a given production process so that structurally we are that much more productive?

			We have found that there are large variations in the payoffs from IT by industry.

With computers, the answer to that question is yes — in some industries, but not all industries. We have found that there are large variations in the payoffs from IT by industry. Companies in more competitive industries tend to realize higher payoffs. High tech manufacturing and financial services are two sectors that have seen tremendous productivity gains, and there are others.

csc.com: What is your take on BPO and BPM?

Gurbaxani: These represent one of the biggest trends because it’s a move toward the corporations of the future that will be focused on their core competencies. Moreover, as we move toward real-time enterprises, we clearly want to move to "on-demand" or utility computing, because you want to use as much of a variable cost structure as you possibly can. So you buy IT services on an as-needed basis without having to make investments in the infrastructure. This enables you to be much more flexible and adaptive.

BPM is more of an advanced form of BPO where you don’t want somebody to take over your process, but you want them to optimize it. You’re seeing a shift in the vendor offerings here, too. Initially, vendors would just come in and take over your process, but now they are increasingly saying, "Here’s what you can buy."

Related Information:

Contact Us and Let Our Experience Help You Produce Results.

Read Part 1 of this interview, in which Professor Gurbaxani disusses important recent trends in IT, including value networks and the timely topic of offshore outsourcing.

CSC’s Outsourcing offerings include Application Outsourcing, Business Process Outsourcing and IT Infrastructure Outsourcing.