How To Build a Well-Run I.T. Department

At Arch Chemicals, we’ve demonstrated that a well-run information-technology organization under our roof can be less expensive and more agile than one run by an outsourcer. But it took several years before I could assert that.

I joined Arch Chemical’s predecessor, the Olin Corp., in 1997 as CIO. The $2.4 billion manufacturer of chemicals, metals and ammunition had been without an I.T. leader for nearly two years. There was no cohesion within the department. It was not unusual to have the company’s e-mail fail a couple of times each week. People just accepted that.

Upon arrival, I studied the company and worked with Gartner consultants, getting to know Olin’s people and processes. Some new CIOs will act immediately. That approach can create a terrible mess and distrust. I wanted to make sure I understood how we could put things together in a positive way.

First, we needed to get I.T. managers working as a team and not as a series of individuals. We started by jointly writing a technology strategy based on the company’s business priorities. Then, we built a roadmap with six major objectives: Improve strategic linkage between I.T. and the business; focus on operational excellence; improve project execution; build a high-performance organization; improve customer satisfaction; and make sure I.T. spending stayed within budget and compared favorably with industry benchmarks. (We have since added two objectives: security and acquisition/divestiture projects.)

I told the I.T. organization that I thought we had about three years to make improvements, or else our work would likely be outsourced. It was up to us to earn the right to keep our work in-house.

From our technology strategy, we developed detailed written objectives with measurable outcomes for every I.T. employee, starting with my position. Our human-resources staff helped us devise a system that quantified performance against measurable objectives and offered rewards; a key component was a variable pay plan for every person in I.T. With this new system, mutual dependencies were recognized and rewarded, and people could see in the money they took home how their work contributed to the company’s success.

To improve project management, I introduced the Software Engineering Institute’s Capability Maturity Model, a framework for measuring software development on a scale of 1 to 5. Moving from level 1 to 2 got us a common approach to I.T. project management; moving from 2 to 3 helped us optimize the project management process as well as focus on cycle time reduction and quality improvement. We are now at about 3.5.

Early on, we hired Hewlett-Packard to help us develop a high-availability technology architecture supported by cutting-edge processes. We rebuilt our systems and networks so that a single point of failure would not interrupt service.

We also adopted the I.T. Infrastructure Library, an approach to I.T. service management. We focused initially on incident and configuration management. Our network availability went from about 98% to 99.99% today.

In 1999, Olin split up its businesses. Its specialty chemicals businesses became Arch Chemicals, where I stayed on as CIO. In 2001—at the four-year mark of our roadmap—my boss asked for a study to determine the cost of outsourcing the majority of our I.T. operations. I contacted four companies, and their bids were stunning: If we went to an outsourcing firm, our I.T. costs would go up by at least 50%.

Why? Most important, we were a well-run operation. In an outsourcing contract, we would have had to pay for overhead, plus an outsourcer has to build in a profit margin.

And while an outsourcer brings in strong processes and skills, we had already installed first-rate operational systems and processes.

Nine years after we outlined our vision in a roadmap, we’re still on track. Instead of being viewed as a cost that’s tolerated, I.T. is now considered a fundamental enabler of business. —Written with Anna Maria Virzi

Al Schmidt has been CIO of Arch Chemicals and, before that, Olin CORP. since 1997. Earlier in his career, he worked in R&D and then product development and design at AT&T’s Bell Labs.