How do you decide that it's time to upgrade your company's e-mail system? If you're like a very few IT decision makers, you upgrade every time your vendor releases a new version. Or, conversely, you may hang onto an older version even after the vendor stops supporting it. But more likely, you prefer a balanced approach like that used by Eric Goldfarb, CIO at PRG-Schultz International a $US376 million recovery auditing firm.
Before giving the go-ahead to upgrade his company's e-mail system from Lotus Notes 5.0.3 to 6.5.1, Goldfarb compiled what he calls the "David Letterman Top 10" reasons to upgrade. "I wait until things pile up. Then I look at that pile and ask, What's it worth to us?," he says.
"There's a cost to move [Domino] applications, so I want to make sure I'm going to get a good return on investment. With an incremental upgrade, I'm not going to get a good ROI." That payback, which Goldfarb expects to achieve six months after the upgrade is rolled out, will come in three areas: features that increase productivity among the company's 3200 users, fixes for software bugs that had burdened his helpdesk, and administrative improvements in security, storage and archival capabilities.
Goldfarb's view is typical. Many IT execs see an e-mail upgrade as inevitable but not something to rush into. Among Notes users, 75 percent of organizations were still on some version of IBM's Domino/Notes 5.x as of late 2003, according to estimates from Ferris Research. And as for users of Microsoft products, in a Ferris survey of 45 organizations in January, 37 percent of the respondents said they were using Exchange Server 5.5, while 63 percent had Exchange 2000 and just 35 percent had Exchange 2003. (Multiple responses were allowed because large companies often have mixed environments.)
"For the last couple of upgrades of Microsoft and Lotus, there've been many customers who've said there's not enough features to justify an upgrade. So now people are moving from Exchange 5.5 to 2003, and a fair amount of Notes people going from 4 to 6," says Ferris analyst David Via. He estimates that a major upgrade, such as from Exchange 5.5 to 2000 or 2003, typically costs $100 to $175 per user, whereas minor upgrades, such as Exchange 2000 to 2003, cost $40 to $90 per user.
One factor that scared off some IT managers from an upgrade to Exchange 2000 and 2003, Via says, is the need to move to Active Directory, a significant architectural change. Such was the case for one large user which will be upgrading from Exchange 5.5 to 2003 over the next 18 to 24 months. The need to convert 50 departments and 15,000 e-mail users to Active Directory added months to the timeline, says chief technology officer Ruben Lopez.
Another factor that has delayed some upgrades is the economy, says Shruti Yadav, an analyst at Nucleus Research. "E-mail upgrades have had to compete with other IT projects," she says.
The well-travelled path
But waiting too long to upgrade can increase the cost and complexity of the project. "The longer you wait, the more work you wind up having to do. You may find that the messaging piece is out of date, and the network infrastructure is out of date," says Sara Radicati, CEO of market researcher, The Radicati Group.
Lee Lovig, who was chief IT architect at a 5000-user financial services company until last month, says his former company's upgrade to Lotus Notes 6.5 -- required because of a merger -- was difficult because more than half the end users were still on Notes 4.5. "We waited too long," says Lovig, who's now an independent consultant. "There's a lot more bells and whistles in 6.5. The client is tremendously different, and we [were] spending a lot of time and money on training."
On the other hand, you don't want to rush to implement the latest release and find yourself the unwitting beta tester of the vendor's newest technology. Ideally, you should wait until you can identify obvious benefits from making the switch. The benefits might be new features that increase user productivity, reduce IT administrative hassles or improve business operations such as customer sales and service in the field.
That desire to achieve ROI without taking unnecessary risks is why Goldfarb prefers to stay in the middle of an application's product cycle -- not ancient, but not bleeding-edge, either. "I ride the middle of the curve, not in the front of it," he says. Robert Ashby, a support manager, has the same sentiment: "It's not that I want to stay back, it's just I'd rather see everyone else upgrade first."
Alan Boehme, CTO at Best Software Inc, and formerly CTO at GE Power Systems, also advocates a middle-of-the-pack position. Boehme says Best is upgrading from Exchange 5.5 to 2003 but expects to stick with 2003 for at least three years, "until we see what happens with Longhorn (the codename of the next version of Windows)."
In his view, there's rarely a competitive need to maintain a leading edge in e-mail technology. "Companies tend to overspend on e-mail," Boehme says. "Instead, think of e-mail on a 1-5 scale, with 5 being the best e-mail system in the world. Then ask yourself if you really need to be a 5. I think most organizations can get by at 3."