Several years ago, Todd S. Coombes proposed a project at a company where he was then CIO. "It was a legacy modernization of some of our older systems and data conversion to the newer platform," he recalls. "It was very important to IT."
But his proposal had to compete against hundreds of other uses of the company's capital, measured against time to payback. "A lot of companies like to do projects if you can get an 18- or 24-month payback," says Coombes, who recently became executive vice president and CIO at ITT Educational Services, a post-secondary educational company based in Carmel, Ind., with 140 campuses around the country. "A lot of these legacy data conversions will have about a seven-year payback, and that's not so great."
The first year that Coombes proposed the project, it was rejected in favor of others with better ROI. The next year, he proposed it again, and again it was rejected. But the third year, he got the go-ahead. What had changed? "In that second year, we had some major, business-impacting outages that were a result of not having done this project," Coombes says. "I explained that it was just the tip of the iceberg of what could happen if we continued not to address things that needed to be addressed just because they didn't match the hurdle rates we were trying to achieve. I think it was eye-opening for a lot of our business counterparts."
If you don't have and can't acquire that level of financial expertise, make sure someone in your department has it, says Peter Campbell, senior vice president of IT at Sprint-Nextel. "I have MBAs who work in my organization," he says. "I'm not a DBA; I don't know how to fix issues in Oracle, but I have DBAs who can do that. In the same way, I have financial people who can analyze ROI."
Coombes says the best solution is to establish a project management office charged with not only managing projects, but also doing cost-benefit analyses and other ROI calculations. "The PMO has at least initially worked in partnership with finance to create templates for calculating ROI," he says.
Make Friends With Folks in Finance
For a "refresh" project such as the legacy migration that Coombes undertook, similar formulas can estimate the cost of not doing the project, both in terms of future maintenance needs and the increased risk of disruption. At Sprint-Nextel, IT asks business units to help create a business risk assessment to measure the potential effect of a failure in a given system. "It's used as a concept for a lot of things, including design requirements," Campbell says. "A high-business-impact system is designed with a lot of redundancies; a low-impact system may not have as many."
Weis, on the other hand, is skeptical of any method of attaching a dollar amount to the risk of a system failure. "If we say it's a million-dollar problem to have an outage in this system, what are the odds of an outage? It might be 0.03%, but you never actually get that percentage. You either get zero or a huge number." So, instead of trying to calculate risk, Weis includes information about such things as the business risk of an outage in his narrative discussion of a project's potential benefits.
"You have to be genuine about it," he adds. At Matson, he says, any threat of business disruption from a failure of an essential system is "as far as the conversation needs to go." But, he cautions, it's important not to raise the specter of a dramatic outage unless it's a genuine possibility. "You have to be fair and reasonable when you make that claim," he says. "And you can't play that card too often."