Q&A: Jonathan Koomey Sees Integration of People as a Key to Transformation
Global surveys conducted for the Transform to Better Perform initiative have found that people are often obstacles to true technological innovation. In one, we saw business leaders complain about the lack of innovation from their IT teams. More recently, IT team members warned that business leaders leave them out of technology planning sessions.
Jonathan Koomey has been studying business, technology and change for many years. He’s currently a Research Fellow for the Steyer-Taylor Center for Energy Policy and Finance at Stanford, and has held visiting professorships at Stanford, UC Berkeley and Yale. For more than two decades, he conducted research at the Lawrence Berkeley National Laboratory. He is currently teaching an online class about modernizing enterprise datacenters.
A good software designer needs to understand the physical and the other limitations of their infrastructure. The fact that IT is not being evaluated on business performance allows software folks who don't understand that stuff to prosper.
Koomey even has a “law” named after him. Koomey’s Law is an observation about power efficiency that parallels Moore’s Law, but that started in the vacuum tube era. It says the energy efficiency of computers doubled every 18 months from the mid 1940s until the early part of the 21st century.
We asked Dr. Koomey to share his insights on some of the obstacles that have surfaced in the ongoing technology transformation. As you’ll see, he focuses on integration – not the integration of front-end software with back-end systems, but the integration of business managers and IT workers.
Transform: Our survey found many business execs feel their IT teams aren’t delivering transformative innovations. But some of the tech managers we’ve spoken with feel they’re brought into planning process too late. What’s your take on this?
Koomey: IT people are often not treated as part of the team. They're brought in after the fact. And the idea that business people aren't including them in the conversation in the modern era is really unfortunate. Because if you're generating revenue and reducing costs because of IT, you really need to understand what the limitations are, and what the potentials are, for deploying IT in the most efficient way. And the only people who can actually speak to that in the business conversation are the IT people.
At the same time, the IT people need to understand and speak the language of business, and that requires training and learning that isn’t normally included in computer science or IT operations training. That also needs to change if IT is to be brought more into the business planning process.
So it’s a general problem of lack of integration between the business side of the house and the IT side of the house. Business folks treat IT as a separate thing, but it shouldn’t be. Number one: it's part of the business. Number two: it should be judged just like any other part of the business. That's why you need to have metrics in place, so that people can figure out where the best return on investment is for deployment of IT or any other business investment for the company.
Transform: It's interesting, isn’t it, that the business people in our survey wanted to apply metrics to measure IT performance, but at the same time they may not be integrating IT into planning.
Koomey: They're not treating the IT department as a true business partner, and that's a problem. IT is treated as this thing that we have to have. It's just this black box. We have to pay for it. It's expensive, but we just buy it because that's what we have to do. Instead of thinking of it as an opportunity, management just thinks of it as a cost center. And that’s a mistake, because IT people have information that is valuable for the business. They can help design deployments of new business units and other activities in a way that will be more effective than they would be without their input.
Transform: Vendors have been aggressively going after business managers the last few years. Are we making a mistake to have the business people talking to the vendors in the first place?
Koomey: That's the same fundamental problem that we just talked about, just a different manifestation of it. Business people are making decisions about IT deployments without fully consulting their IT department. That's pretty clearly a problem. Now internal IT departments may not have sufficient expertise to evaluate all aspects of the business problems facing their company, but they have knowledge that is relevant, because they are the ones in their business now deploying IT. So, they will know the constraints, the choke points, and the opportunities, in a way that no one else will. The people on the business side will not be able to properly evaluate those issues alone.
Transform: Another complaint from IT people, particularly on the network side, is they fear that recent software hires aren’t well trained to understand the subtleties of the data center. Who's going to evaluate these new software application builders?
Koomey: A good software designer needs to understand the physical and the other limitations of their infrastructure. The fact that IT is not being evaluated on business performance allows software folks who don't understand that stuff to prosper. If people are evaluated based on business performance, and the performance is measured properly, then the message will come to those folks: ‘You need to learn about X, Y, and Z, because we will waste resources otherwise.’
But again, it's this lack of integration, this lack of attention to treating the IT side – all aspects of it – as part of the business. I think once you start integrating IT much more closely into the business, and then evaluating IT based on business performance and return on investment, then people will realize that they need to train the software folks in the infrastructure details. It's often a lot easier to change things in software than it is to fix them in hardware, but that requires software folks to pay attention to hardware.
Of course, you could also cause big problems if you're not careful, if you're not doing things in a mindful, careful way. I can't speak to the question of what's happening with the training of software engineers. But I can say that if the incentives are set up right, and the metrics are set up right, then people in software will get the message that uptime is critical, we need to improve performance, we can't have bloated code, we can't have code that's going to cause the network to go down. That all is going to come through, as long as people are being measured on their contribution to business performance and return on investment.
Transform: Are we discounting that this transformation will require an evolution in the way businesses operate? How long do you think that kind of evolution might take?
Koomey: Well, let's look back at history. Electric motors started to be widely used at the end of the 1800s. They allowed the factory owners to put motive power almost anywhere. Before that, they had to design buildings around the water wheels or steam engines, which were very concentrated sources of power. Then they had to figure out a way, using gears and pulleys and other crazy contraptions, to get the motive power to where they needed it in the factory. And so electric motors came in, but for a while things didn't change very much. People didn't see much improvement in productivity, because they were just putting electric motors in the same buildings, in the same institutional structures, where they had the old power sources. It was only when people realized that, ‘Oh, I could actually put this electric motor in other places around the building, and optimize my production process around that’ that you started to see the productivity improvements that investors expected. It wasn't just that the buildings themselves took a different shape, it was that the organizations took a different shape as well.
It was the same thing with personal computers in the ’80s. Companies bought a lot of computers, and people thought that this was going to revolutionize business, and they were right. But for a long time – 10 or 15 years – nobody saw the expected productivity enhancements in the statistics. The reason was that companies bought PCs and plunked them down on people's desks, but didn't change the organizations in which these PCs existed. They didn't change the information flows, they didn't change the authority structures.
Every time there is a new big innovation, it takes 10, 20 years before the institutions figure out how to take full advantage of it. There's always some institutional innovation that's needed, that is generated by these new technologies, and it takes a decade or two typically, before we can take full advantage. Technology changes at a very rapid pace but people need a little more time to figure out how to make best use of it inside their organizations. That's what we're seeing. That's why you're seeing this lag between what we could do and what we're able to do in practice.
Transform: Any closing thoughts?
Koomey: The most important step is to map all of your businesses processes on to the IT infrastructure, so that you can measure a return on investment for any investment in IT. That means analyzing both the revenues and the costs, because IT is not and should not be just a cost center, it should be a cost-reducing profit center.
But in order to take that step, you need integration between IT and the business. If you can achieve that integration and develop metrics to reward people for delivering better business performance, then you will have actually achieved the ultimate goal of using IT to full advantage for business.
Point of view storyQ&A: George Rockett, DatacenterDynamics
Srini KoushikChief Technology Officer
Point of view storyThe Traditional Model of IT is Dead - Meet Exponential IT
Zahl LimbuwalaCEO and Co-Founder