Threading a path through the perils of a major technology investment is like stumbling though a bad dream involving one of those nested Russian doll sets.
Like the dolls, which pop up to disclose duplicate, slightly smaller dolls inside, IT projects reveal fresh levels of risk at every turn. Evade the hazards associated with bolting new hardware or software into place and the next problem swims into view: how to convince business managers to make use of it. Solve that issue and there is still the shadow of the productivity paradox to deal with. The high churn rate of CIOs worldwide attests to the complexity of the process.
Famous fiascos are a dime a dozen: from the historic cockup of Westpac's ill-starred CS90 initiative, to the $200 million capsize of a Bank One mega-project. Current efforts to implement enterprise business systems are busily harvesting their own crop of disasters. Why are corporate information tsars and tsarinas still having trouble transforming technology investments into bottom line business benefits?
Making vendor nonperformance into an all-purpose whipping boy isn't the answer. True, vendor-customer partnerships often take on the overtones of a bad marriage, as mutual dissatisfaction mounts over issues ranging from hardware interoperability and the cost of customisation, to middleware that isn't. But open systems and stiffer competition have smoothed the project installation phase compared with previous eras. And smart IT managers have learned to compensate for vendor promises exceeding vendor performance by anything up to an order of magnitude.
Unfortunately, successfully locking the technology tools in place simply causes another problem doll to pop up. It guarantees life won't become any easier for the project's owners in today's world of technology investments, which are expected to deliver intangibles such as quality of service and customer convenience in addition to the more traditional cost-cutting benefits. Huge second efforts are required to modify corporate cultures and business processes so organisations can leverage the new toolsets. This is the new killing ground on which CIOs find themselves exposed, according to consultants like Steve Moir, KPMG's associate director of world-class IT.
There is a widespread lack of organisational follow-through to crystallise the benefits of technology projects, Moir believes. "One of our clients spent several million dollars building a system to provide better service for their clients. But the system that came out at the end of the project didn't meet their functional needs and was never put into practice," he says. "A post-mortem showed the demand side -- that is, the business users -- didn't take the organisational responsibility needed to ensure that at every step along the way during the life of the project the decisions were made that would deliver the benefits needed.
"Over the past 12 months, we have seen half-a-dozen large organisations struggling to get their business people and IT people to talk together and share a common view of their roles and what they are trying to achieve for the organisation," Moir says. "We believe this is just a lack of experience and knowledge on how to run projects, in particular IT projects, and the level of communications that is required.
"What we call the demand side has to learn to express itself more clearly across the organisation."
To further complicate the issue, there is the running sore known as the productivity paradox, which refuses to completely disappear. It first surfaced a decade ago when Steven Roach, chief economist for Morgan Stanley, started alarm bells ringing with a paper entitled "American Technology Dilemma". Roach unearthed a wild mismatch between huge increases in computer power per white collar worker and that worker's near-zero productivity gains. Despite a $US862 billion investment in IT hardware during the 1970s and 80s, average productivity in the US service sector grew by only 0.7 per cent, he pointed out.
Roach's findings lit a bonfire of debate that is still burning brightly today. Many have jumped in with contrary views. They point out the tools used to measure service sector productivity were forged in the machine age, not the information age. So traditional yardsticks don't stretch to measure intangible but undoubted benefits such as quality of service and customer convenience. Others argue the data is outdated or that more time is needed for IT investment to reach the critical mass necessary before large-scale statistical studies can detect its impact.
Despite that, information productivity guru and veteran paradox proponent Paul Strassman, former CIO with General Foods, Kraft and Xerox, insists no absolute connection can be established between corporate IT investments and corporate profits. "Since 1982, I have tried to find correlations with IT spending, including every conceivable variable, such as revenue, assets, stock market prices and shareholder equity," Strassman wrote recently.
"So far, I can't report success, and nobody else (to my best knowledge) has found a positive relationship. Thus, the paradox remains a phenomenon; though it is also true that computers enable people to work faster, smarter and in ways that were never feasible before. After looking at all the evidence, I have concluded that the computer paradox is here to stay."
Peter Weill, director of the Centre for Management of Information Technology at Melbourne School of Business, has taken the paradox debate a step forward. Weill and GartnerGroup program director Marianne Broadbent laid out a new way of looking at IT investments in their book Leveraging the Infrastructure: How Market Leaders Capitalise on Information Technology, published last year by Harvard Business School Press.
They found IT investments can be divided into four main categories: transactional (cost-cutting), infrastructure (a base set of IT services, such as networks, shared across an organisation), informational (improving information flow, quality of service and customer convenience), and strategic (gaining competitive advantage). They suggest successful corporations are those who manage the four categories like asset classes in a financial portfolio. Each one, like different financial products, is associated with a different level of risk and can be managed by applying time-honoured investment portfolio theory.
The cost-cutting transactional IT investment might be likened to blue chip stocks and is the category with the most measurable bottom line performance, Weill says. Infrastructure investments correspond to financial options and should represent about 55 per cent of an organisation's IT portfolio, he says.
"Having more IT infrastructure gives an organisation more flexibility, faster time to market and higher revenue growth. The downside is lower profits in the first few years," Weill says. "So there is an upside and a downside, but senior managers are comfortable making that kind of balancing investment decision."
Strategic IT investments are the equivalent of investing in Latin American funds and represent the high-risk, high-return portion of the portfolio. "About 50 per cent of projects in this category fail, 40 per cent break even, and 10 per cent do remarkably well by making very large contributions to overall profits and giving a two to three-year market lead," he says.
According to Weill, the final category, informational IT investment, is the most difficult to assess in terms of measurable benefits. "We didn't find any evidence that [companies which] spent more on providing more information did better on a bottom line basis. We did find they had higher quality and faster time to market; but it was not strong enough to reach the bottom line, because of the dilution from all the other efforts going on in the [company]."
By their nature, informational investments are tricky to measure, because they generate intangibles such as quality, convenience, reliability and timeliness rather than countable widgets. The correlation is even more difficult to establish, because some organisations are more adept than others at capitalising on improved information flows. Some are disciplined at using the information, according to Weill, while others are information cowboys, who pass on a small percentage of useful information internally.
CIOs working the coalface are comfortable with the IT investment paradigms that Weill and Broadbent advance. "They are quite relevant and valid and certainly are the types we have focused on," says Rod Walden, group business systems manager of giant food wholesaler Pauls Ltd. Walden agrees measuring the impact of informational IT investments -- data warehousing projects, for example -- poses a difficult problem. "It is a question of making IT tools available to managers and then asking how well have they used those tools to grow the business."
So running a yardstick over the results would involve gauging the productivity of managers in some way. But Walden does not think current measuring sticks are subtle enough to tackle the job of managerial productivity. "At Pauls we haven't taken the step of measuring managers' productivity, because mid-managers perform a role that is fluid and constantly changing," he says. "How would you measure it? By comparing how many crises a manager solved last month compared with this month? It would be very difficult."
As to the CIO's role in the process, Walden admits he has difficulty with the term CIO, because "you can no longer be focused just on information technology systems; they are business systems, and you are integral part of the business just as much as the personnel manager. You can't measure the productivity of IT without measuring the productivity of the whole business."
The act of installing an IT system and making it run efficiently is clearly a CIO's responsibility, Walden notes. However, the second half of the equation, relating to how wisely an organisation uses the system to add value to its bottom line, is something a CIO can't be held totally responsible for. "The results will be spasmodic and piecemeal, unless you have total buy-in at the very top of an organisation."
Walden's sentiments are seconded by Ian Mackay, a Sydney consultant whose background includes years of experience as a senior IT executive with multinationals. "The key element of business process change is the commitment of senior managers to what IT infrastructure is capable of delivering," Mackay says. "A project may be capable of delivering productivity improvements, but if the business manager doesn't choose to exploit those capabilities, they will be wasted."
Michael Coomer, CIO of National Australia Bank, feels there have been recent advances in the ability to measure returns in both the informational and strategic categories of IT investment. "There are some measurements which Peter Weill was not aware of when he wrote the book. Some of the information being provided by the Federal Reserve in the US, particularly on strategic investments, suggests there are robust and tangible ways of measuring the impact on productivity." As supporting evidence, Coomer cites US economic tsar Alan Greenspan's documentation of the profound impact technology has had on the US economy over the past five to 10 years.
As far as informational investments are concerned, Coomer points to Internet developments, which wring value from customer databases, and suggests they open up possibilities of providing tangible measures of return on investment. "Whether it has to do with customer retention or growth, there are reasonable and reliable business cases that we can use," he says. "So off the top of my head, I'd say there are measurable outcomes for all four of Weill's categories."
Like other CIOs, however, Coomer concedes these outcomes aren't dependent on IT alone, but must be accompanied by business process changes.
Take any ERP project and for every dollar invested in actual computer systems, the cost of associated business process changes can tack on another three-to-five dollars. "The sorts of numbers we see for the nuts and bolts hardware and software component are in the 20 to 30 per cent range," Andersen Consulting partner Graham Henry says. "The rest is in reprocessing, reskilling and re-engineering the organisation; and that is the area where people don't make the investments they need to."
Pauls Ltd is fresh from a significant ERP exercise, in which it implemented a SAP R/3 suite as part of a complete replacement of its core systems. The project was successful, according to Walden, partly because Pauls consciously avoided biting off more than it could chew. Carrying out an ERP implementation and a business process re-engineering exercise in one large gulp is a huge mistake, he says. "To undertake a huge, risky business process re-engineering exercise at the same time you are installing an ERP is inappropriate. The focus of ERP implementers (who suggest the all-at-once approach) is quite wrong and misleading for the market. Most of the CIOs I mix with have come to the same conclusion."
The way to proceed is to complete the ERP implementation before launching progressive changes in the BPR space, Walden says. "An ERP implementation that is done properly in an organisation is an enabler for continuous staged improvements [in business processes] for the next 10 years. It is better to implement an ERP with a few bad processes but get it in and running first, and then look for improvements."
Finding the right balance is not a trivial exercise, says Ian Goch, general manager of IT development for Coles Myer. "Every project is bigger and harder than you thought it would be. There is always the temptation to break it into phases, because you don't want to disappear into a project for two years and pop up to discover the world has moved elsewhere," he says. The drawback with that approach is that maintaining each new phase as it is completed draws resources away from the next phase, so forward momentum falters.
Andersen's Henry defends ERP projects against accusations their returns don't justify the risks they represent. "Many enterprise business systems don't deliver their promises, but it is a big step from that to where you say the investment is not worth having at all," Henry says. "There absolutely have been failures . . . and one lesson is that if you can possibly avoid it, you never want to have the all-embracing, all-encompassing project.
"You are much better advised to do it a piece at a time in such a way that you minimise your risk. The mega big bang approach is in the high risk category from the start." But Henry notes that many IT investments are not optional for organisations. "A lot of investment has gone into building systems without which organisations couldn't conduct their business. Airlines couldn't run without reservation systems and banks couldn't run without having their customers online.
"I am sure that on a project-to-project basis, some of those developments would have been successful when subjected to a business case analysis and some wouldn't have. But the reality was that if those companies wanted to be in business, they had no option but to do the project."
Henry also points out that many projects fail because the business cases on which they were made are made obsolete by events before the project can be completed. "Those sorts of projects often don't live up to the business case they were developed on because initial assumptions and knowledge levels change. If you were to do the business case armed with your new level of knowledge, you might find you shouldn't have done it." By then, the organisation has already invested x million dollars so the argument now switches to questions of marginal returns.
Fear And Faith
In the absence of quantitative methods of predicting returns from informational and strategic IT investments, it can be argued they are happening on the basis of either fear or faith. That is, fear some competitor will gain an advantage if the investment isn't made, or faith the investment is taking the company in the right direction.
Whatever the ultimate answer to the productivity paradox, it seems clear the key to unlocking corporate benefits from IT investments involves costly, risky and time-consuming organisational changes. Over many of these changes, CIOs have little or no control. Russian dolls anyone?