Instead of simply weighing the costs of the implementation from an IT perspective, Howell's group worked with marketing to determine what recruitment efforts would be most effective with the target customers. "Historically, would we have put that much weight on marketing? Maybe," Howell says. But only subjectively and only at senior management's urging if they thought it might be a driver in the project's success, he says.

Toward the end of the project-valuation effort, Citibank called on a startup consulting firm to verify the predicted returns. IValue looks at shareholder value to assess ROI, using the same formulas that Wall Street analysts rely on to value companies. It builds economic simulations of IT projects that quantify things such as the demands they will place on a company's IT systems and architecture, other costs, and soft measures of success such as customer loyalty and adoption rates. Those are put into a cause-and-effect formula that traces the project's impact all the way out to its effect on the company's stock price.

The simulation can be tested by changing the cost components or demand parameters to determine what aspects will influence the project's economic performance, says Chris Gardner, who co-founded 6-month-old iValue. "Then we can make better decisions, guide development efforts, and establish priorities," he says.

IValue's simulations agreed with most of the results from Citibank's in-house valuation model. "We had to change some things in regards to how we were going to amortize the software costs," Howell says. Citibank plans to use the valuation model to evaluate future IT projects, and it hopes to cut the time needed to analyze a project to about seven weeks from the four months Howell's team spent on the portal. IValue will start selling its applications next month for businesses to create their own valuation models.

There aren't always hard numbers with which to quantify a lot of the intangible outcomes that companies look for when they implement an IT initiative, so a speculative model based on historic data is the answer, Meta's Rubin says. For instance, retention rates are a good measure of customer loyalty. A company looking to purchase loyalty-enhancing tools should build a model that correlates percentage change in customer loyalty with business goals. The models for developing these calculations need to be in place for all projects, Rubin says, so there's consistency in the decision-making process across the company.

Schlumberger Ltd., a services provider to the oil industry and an IT consultancy that spends about $500 million a year on information technology, has a new ROI-assessment process that considers an expanded set of variables. Instead of considering one estimate for the completion time of a project, the new methodology has three: a worst-case scenario, a likely case, and a best case, says Jane Walton, an IT portfolio manager.

One recent project was expected to take six months to complete and cost $6 million using Schlumberger's traditional ROI methodologies. It ended up running 20 months and $9 million. When the new ROI model was applied after the fact to the same project, it accurately predicted a worst-case scenario of $9 million. "If we had determined the range of possible outcomes and assessed the variability and risk associated with the project, our ROI analysis would have come out completely different," Walton says. "The business case wouldn't have been there."

	Shareholder benefit is what counts, Buckley says.

Companies with the most rigorous approaches to valuing IT projects use standards that link projects with the overall business strategy. At mutual-funds firm The Vanguard Group Inc., CIO Tim Buckley charts out where his company is in terms of the business goals at issue, then overlays that information with the direction a technology initiative is likely to take the firm and what the final benefit will be to its mutual-fund holders.

Vanguard has made low operating costs a centerpiece of its strategy, so spending gets high-level scrutiny. The executive team, usually including the CEO and CFO, meets in "sunlight" sessions where members debate a project's projected benefits. "We have to challenge each and every assumption," Buckley says. "You can easily get enamored with technology, and you think everyone is going to adopt it. That's where you have to have experience to give you insight into reality."