The complexity factor

Why IT is getting more complicated all the time—and what your company can do about it

If you've ever spent long, frustrating hours troubleshooting a problem on your home computer, consider what it must be like for J. P. Savage. As Scotiabank's senior vice-president of systems operations and technical services, he has, among many other responsibilities, oversight of the PCs used by some 20,000 employees at 1,000 branches across Canada. Because they link back to central databases, all those computers need to run pretty much in the same way. It's not an easy task. “When we make a change out there,” Savage says from his suburban Toronto office, “it has to be managed in a very delicate manner.” Without factoring in the costs of purchasing and installing new technology, the logistics of a rollout to all those far-flung branches typically costs the bank about $4 million.

It is much the same story at any of the big banks–or, to some degree, at any corporation. Technology that is supposed to improve how a company functions becomes a drain on resources. Savage cites well-known statistics about how the purchase price of a PC represents just 15%-20% of its total cost. “The real complexity,” he says, “is in managing the technology.”

In fact, as a business problem, technological complication seems to be growing worse. Every new software application, every new function, adds a layer of complexity. The pace of change is relentless: e-mail begets password maintenance (and employee training), then a flood of spam, which requires constant virus monitoring, and now off-site access–sometimes to multiple devices. It can drive an IT department to distraction.

There was a time when computers were used to crunch numbers; now they're used to exchange information. That requires technologies to work together like never before. “There is more integration between applications,” says David MacMillan, manager of office network systems at Shoppers Drug Mart, “so there are more interdependencies between those apps. That definitely tends to increase the level of complexity.” The more parts, the more difficult it is to make them work together–and the more likely it is that something will toss a spanner in the works.

Technology opens up new ways of doing business, but it can come at a high cost. Anthony Picardi, a senior vice-president at research firm IDC, says polls of companies 15 years ago showed that 75% of IT budgets went to buying software; today, 70%-80% of IT spending goes to maintaining and fixing existing systems. Getting IT complexity under control has become a pervasive concern of executives, who regard it as an ideal focus of efforts to cut costs. In a 2003 survey by consulting firm Booz Allen Hamilton, nearly 90% of CFOs trying to reduce operating costs cited “managing a patchwork of different systems” as their main IT challenge. Another study, by London, Ont.-based Info-Tech Research Group, found 95% of IT departments aren't delivering some projects on time or to the full satisfaction of execs.

For companies that sell corporate technology, complexity is both a threat and an opportunity. They face a marketplace full of cynical customers who have learned hard lessons about the real costs of buying technology. As a result, many top firms have tailored sales pitches around the theme of reducing IT complexity–which, however efficient the IT department, has become an unavoidable reality of modern technology. Sometimes, the only place to turn for answers is new technology.

The PC problem

In 2001, Savage faced an expensive decision: whether to update Scotiabank's aging collection of PCs. In the 1990s, the company, like many financial institutions, had used the client-server model, where desktop PCs (“clients,” in tech-speak) were connected to a central branch server. As the 21st century dawned, the bank used the extra fibre capacity to deliver high-speed networking to the branches, and started routing data on a secure intranet. The PCs, however, were still the same, and under the load of the latest software applications, like web browsers and Microsoft Office, they were grinding to a halt. The infamous three-finger salute–keying the Control+Alt+Delete sequence to restart a computer–was becoming all too common. Customer reps trying to close a loan application had to make idle chit-chat for 10 minutes as their PCs booted. “As a moment of truth with that client, that's not a very good experience,” says Savage.

One solution was to just buy PCs, but Savage didn't want to get trapped in a Wintel churn cycle (shorthand for the Windows-Intel alliance under which the two companies design more power-hungry software and faster processors in tandem, which forces companies every three or four years to buy the biggest PCs they can afford). Beyond the cost, managing all those desktops would be onerous: even simple software upgrades would take seven or eight months. Instead, Savage put the branch PCs on a diet, making them so-called thin clients–stripped of nearly all applications. What was left was a kind of browser employees use to access applications residing on centralized high-security server farms. The PCs don't process or save data in any way: all they do is provide a window into the applications running on the servers. Even the branches' old Pentium 1 computers could handle that.

The thin-client technology, designed by Citrix Systems Inc., based in Fort Lauderdale, Fla., is just one example of systems built on a so-called served-based computing model. The approach is becoming more popular in part because it puts control into the hands of central IT administrators. When Savage upgraded the operating system and browser used by the branches, he did it over four weekends, all at one central site, by simply upgrading the server farms. “And I didn't spend anywhere near $4 million,” he says.

The bottom line? Scotia's thin-client branch system will save the bank $40 million over five years. And that doesn't take into account improved productivity, now that employees no longer need to give the three-finger-salute several times a day.

The misery of spam

In the past 10 years, e-mail has become an essential form of communication, and much of its success arises from its relative simplicity. Yet few pieces of IT infrastructure cause more anxiety. So why the unease? “Managing e-mail isn't complex,” says Ronald Knol, IT supervisor for Rainmaker LP, a special-effects and post-production house in Vancouver with 175 employees. “It's the spam that's the problem.”

Rainmaker does most of its IT in-house, but spam got the better of its tech staff. Five years ago, Knol wrote his own spam filters for the e-mail server; three years ago, Rainmaker installed third-party software to catch the junk. Even then, Knol was getting as many as 20 spam messages a day. And the hundreds of thousands of junk mailings Rainmaker received every month was gumming up its own servers. Sifting through the cached spam took 20 hours a week for his IT team of four. For a while, Knol chalked up the ongoing battle as just a fact of life. But when the third-party spam filtering company wanted more money, Knol began to research alternatives.

He found a B.C. company, Electric Mail, that provided an outsourced security service called PerimeterProtect. Now, all e-mail destined for Rainmaker first goes through Electric Mail's servers, which clean out spam and viruses. “At first, it feels kind of weird to let somebody else handle your mail,” says Knol. “But then we thought, all e-mail passes through other servers anyway before it hits you. So who do you trust more?”

In May, 600,000 messages came into Electric Mail for Rainmaker–but only 26,000, or 4.3% of them, were legit. That's a lot less to bog down Rainmaker's own servers. The only IT resources it spends on managing e-mail now are the 30 minutes a week Knol devotes to making adjustments for new clients' addresses.

In Knol's view, it's not technology that's complex–in fact, it gets more advanced and easier to use all the time. “Trying to make it work in our workflow, in our building, is the hard part,” says Knol. “That is most often where the complexity comes in.”

Network games

In June 2003, all five vice-presidents in charge of technology at the subsidiaries of the Co-operators Group met at a Sheraton hotel near Toronto's airport for their regular two-day quarterly meeting. But this gathering carried a little more importance, because Co-operators Group CIO Vivien Fong and her vice-president of information technology, Kevin Hutchinson, were proposing a dramatic change: consolidating the group's 40-odd contracts for telecom and data services into just one outsourced service based on an IP network.

As a co-operative, the group runs democratically. All the vice-presidents needed to weigh the risks and benefits of the change both for the good of the group and for their individual companies. The discussion took up much of the day, and continued through dinner. “There was debate,” says Fong, “because it's always difficult to ask people to give up something they have responsibility for.”

But the proposal made sense. Managing networks is not the core competency of even technology vice-presidents. By contracting out the job to a telecom services provider, some 600 office locations would get higher bandwidth, reliable service–and costs for the subsidiaries would not increase. The IP network also served a strategic purpose, future-proofing the group for new applications and services.

Co-operators signed a six-year, $66-million contract with Telus in May 2004. The IP network rollout took most of 2004 and overflowed into early this year. But apart from Telus having to juggle installation schedules around striking telecom workers, the project went smoothly. There was one unforeseen problem: some older communications connections at office locations were housed in secure steel boxes–so secure that installers couldn't find a way of wiring them with the new cables. Extra construction work was required to replace them with standard network closets. “You think of all the different things that could go wrong,” says Hutchinson, “and that was the big thing–a physical, low-tech problem.”

Sometimes, even the simplest things can end up being complex. The challenges posed by complexity are unique for each company, because they so closely reflect how businesses are organized and function. There is no silver bullet, and many CIOs are resigned to fighting an endless battle. New technology will solve old problems, but also create fresh ones. As technology changes, so too will the nature of its complexity.