There are few IT projects more fraught with danger than upgrading core systems. If you do things wrong, you've crippled the workhorse that carries the business.
On the other hand, the benefits of such a project can be enormous, says W Douglas Lewis, executive vice president and CIO at Six Continents Hotels, an international hotel management company. Lewis is in the midst of installing a revamped central reservation system and local operations management systems for 3200 hotels.
To make the project succeed, Six Continents partnered with end users from the start, says Lewis.
"Involving the people who leave fingerprints on the keyboards is critical, because they know what the hotels need," says Lewis. "If we approached this from an IT point of view, it's likely we would have built the wrong solution."
Upgrades of core systems are among the top five projects that IT leaders said they'll tackle this year. Getting things right is important because core systems upgrades aren't cheap. Lewis and others share strategies for improving the odds of succeeding at these large, critical projects.
Test Your Powers Of Persuasion
Although Steve Scott, vice president of information systems division at Rancho Cordova, California-based Vision Service Plan (VSP), says he expects that upgrading the claims system will lower the nonprofit's cost structure, the more compelling rationale for his project is increased functionality.
Scott says he envisions the network's 20,000 optical businesses being able to do more self-service tasks, such as checking patient eligibility online. The new system also opens up the possibility of VSP's 36 million consumers being able to use the Internet rather than the telephone to do things like check on a claim's status, he says.
This type of business case is common, says Allison Bacon, an analyst at AMR Research in Boston. It's hard to pin down a return on investment for a core system upgrade, because you're spending money on something that already works. Selling the project to upper management often requires making a compelling argument in favour of hard-to-measure benefits and new features.
Rent the Necessary Brains
If it's not a core competency, then outsource it, say IT leaders. Six Continents hired subcontractors to deploy the new systems in the hotels and conduct training. "It doesn't make sense to build that capability internally for a one-time activity," says Lewis. But he relies on his staff to do the application programming because they know the business better than an outsider would.
Henry Volkman, director of IT and CIO at Del Taco in Lake Forest, California, says his company will make use of outside help to upgrade its 254 point-of-sale systems. The fast-food chain is using an outsourcer to install the hardware in its restaurants, rather than stressing its busy internal staff with that chore. It's also working with Microsoft to design a SQL Server inventory database.
The vendor of an application always knows its ins and outs better than you do, says Volkman, so it only makes sense to marry its expertise with your business knowledge.
Bacon says that having a mix of internal and external development teams gets you the most skills at the lowest cost. Using your own people means spending less and moving faster. But you need to lean judiciously on experienced vendors and outsourcers in order to gain the greatest benefits out of something that is new to your team, she says.
Make Sure Your CFO Says No
"My CFO can make a vendor cry and a nickel scream," says Volkman, who meets with his company's top financial executive weekly. He's the reason the upgrade project is on budget, Volkman says. If the project runs into a snag, IT has to find a way to fix it with what money is on hand. "You'll never get him to agree it was right to spend more money than budgeted," he says.
Volkman adds that it pays to hire only the best project managers. "You can take a good manager with a mediocre crew and run rings around a poor manager with a high-quality crew," he says.
Lewis also lays the credit for keeping his project on track at the feet of talented managers. "I can't brag about anything unique we did," he says. "We just managed the heck out of it."
Lewis adds, "I have a premise that big-bang projects blow up CIOs." So he stops every few steps to evaluate the results of the latest "little bang." For example, he says, putting the new system in a few hotels for beta testing revealed a problem with the communication systems. It was easier to change course at that point than after several hundred hotels were involved, he says.
Scott echoes that strategy. Interim checkpoints are a must, he says, and their associated performance metrics must be put into the plan from the very beginning. But it's just as wrong to be a slave to metrics as it is to ignore them, he says. Sometimes you have to do a reality check and recognise that a metric isn't working — and then you find another one, Scott says.