When FedEx Corp. CIO Rob Carter told Wall Street analysts earlier this year that just one of his company’s Web-based applications — Web tracking of shipments — was delivering $25 million (U.S.) in savings a month, CIOs everywhere must have dropped their jaws and turned green with envy. $25 million

a month?

In these tight-money times, being able to quantify the benefits from IT projects so decisively is pure gold. About the only way to get project funding in many companies is to present an airtight business case showing a significant hard-dollar return on investment (ROI). And you’d better hope you can prove after the fact that you really did reduce costs or increase revenues.

Given this kind of pressure, how far should CEOs, CFOs, board members and shareholders trust big ROI numbers for IT projects — either calculated before or after the fact?

ROI calculations done to support a funding request are probably suspect at best, especially so if based on some software vendor’s calculation tool. “”Don’t trust them,”” consultant Ron Babin says of the numbers generated by vendor tools. For one thing, they’ll be laughed out of the boardroom in companies serious about fiscal responsibility.

The real ROI picture is heavily dependent on a company’s particular circumstances, says Babin, an associate partner with Accenture in Toronto. “”Finding the money”” — figuring out which technology projects will deliver a true ROI — is a rigorous process that should take weeks or even months. And that’s the way most “”professionally run”” firms do it, he insists. Well, maybe.

Post mortem ROI calculations — though harder to fudge — may not always be what they appear either, as we’ll see. But maybe the problem isn’t so much that IT managers cook ROI figures as that in companies with a one-track ROI mentality they have to. More and more analysts today are arguing that ROI is not the sine qua non when justifying IT investments. Shareholder or business value is far more important.

“”Simple ROI calculations done on a project basis more often than not ignore the potential broader impact an investment may have on the company’s balance sheet,”” says consultant Adriaan Davidse, senior manager in the strategy group at Deloitte Consulting’s Toronto office.

The FedEx Web tracking application is a good illustration of the sometimes deceptive nature of dramatic ROI numbers and the potential flaws in a strict ROI approach. It’s not that the numbers Carter presented are fudged. It’s just that, under scrutiny, it becomes clear they’re not really ROI numbers at all. And they need some qualifying.

The company’s Canadian CIO, Nancy Reynolds notes, for example, that FedEx first rolled out the Web tracking application in 1994, long before the Net was a business staple. It’s not as if the company saw significant savings from day one — or indeed for many years.

“”It’s just in the last few years that (use of the application) has really skyrocketed,”” says Reynolds. “”The growth rate in Web tracking is now in the triple digits year over year.””

Furthermore, despite the “”savings,”” costs at FedEx have not necessarily gone down. The savings are really costs avoided. If Web tracking were not available, the number of calls to customer service requesting tracking would have continued to grow and FedEx would have had to expand its call centres.

“”The customer service department is still there, of course,”” Reynolds says, “”but we assume there are a lot fewer agents than there would have been and that we’re paying less for telecommunications services, etc.””

The operative word here is assume. FedEx knows that it now receives 2.3 million Web tracking requests a day worldwide. It knows it receives 500,000 customer service calls a day worldwide, of which about 18 to 20 per cent typically are tracking related.

“”But how many extra calls we would have received (at call centres if we hadn’t implemented Web tracking), we really don’t know,”” Reynolds concedes. Of course, they don’t. How could they?

Carter presented two other tantalizing numbers. It costs FedEx $2.14 to track a package through customer service, he said. The cost to track a package on the Web: four cents.

Sounds impressive. But Reynolds reveals that while the $2.14 figure factors in all operational costs including agents’ salaries, facilities and telecom, the four cents for Web tracking only includes operational and “”ongoing”” development costs, not the original costs to build the application.

Also, the four cents is based on the number of Web tracking requests FedEx receives today. But clearly, customers now track shipments far more often than they ever did in the past when they had to call customer service. Web tracking is easier and less time consuming for customers too, so they do it more.

None of this is to say that FedEx shouldn’t have built the Web tracking application, or that it’s really ROI negative. There is nothing at all to suggest this. Cost avoidance is a valid objective. And there are clearly solid qualitative benefits — improved customer service if nothing else. But they’re apt to get lost in the rush for quantitative gold.

FedEx, Reynolds says, is generally rigorous about having detailed business cases for IT projects, and they get “”picked apart pretty well”” by the company’s bean counters. But it likely did not work up an ROI case to justify investing in Web tracking — though since it was so long ago, no one is quite sure at this point.

“”It was probably one of those applications developed on the basis of innovation — to get it out there, to establish our presence (on the Web),”” Reynolds says. “”I don’t know if a business case was done, and I’m not sure how meaningful it would have been if it had been done.””

Exactly right, Davidse would say.

ROI approaches simply don’t work for many kinds of IT projects, he argues. Deloitte now urges clients to adopt a more “”holistic”” approach to prioritizing investments. It recommends that they divide IT projects into four distinct categories: experimental, transformational, renewal and process improvement. Only the last category lends itself well to an ROI approach.

FedEx’s Web tracking application would clearly have fit in the innovation category in 1994. “”You don’t really know what returns are going to be,”” Davidse says. “”Typically you start by making small investments to reveal what the ROI could be.”” This is precisely what FedEx did.

Transformational investments, which might include major ERP implementations and network infrastructure upgrades, focus on increasing the size or scale of the business — growing revenues without changing the underlying business model. “”The risk of not doing them is often higher than the risk of doing it,”” Davidse notes.

Renewal or maintenance projects, such as upgrading desktops, also do not lend themselves to an ROI approach because they typically cross business unit lines and it’s often more a question of “”(not being able to) do the job properly without upgrading.””

Process improvement or productivity projects might include shifting customers to lower-cost channels, streamlining processes, automating data capture. This is where the FedEx Web tracking application might fit today. “”They’re not strategic, they’re operational,”” Davidse says. “”They’re very focused within business lines and ROI works very well here.””

Companies should allocate funds separately for each category based on its strategic objectives. Transformational projects should not have to compete with process improvement projects for funding — which is too often the case in organizations that adhere to a strict ROI approach, Davidse says.

There is more to the holistic approach Deloitte and others are advocating. The complex nut of it is learning how to analyse projects with a view to how they will contribute to overall strategic objectives. What is clear, though, is that a knee-jerk demand for short- or medium-term ROI too often results in projects not being done that should be done. And vice versa.

Share on LinkedIn Share with Google+