Let’s be illogical for a moment. The reason you can tell that IT has a shaky record on delivering measurable returns on investment is that there are tens of thousands of management consultants whose job it is to explain why your company needs to shell out yet more money upgrading a system that did a perfectly good job the last time you looked. If it were a straightforward economic decision – “Hey, we spent £2.6m on the new network infrastructure, and just look at these revenue gains! Our share price is soaring!” – you wouldn’t need the consultants. True, there are some IT investments in some industries that can create wealth through opening up new market opportunities. But these tend to be few and far between, and in comparison to the vast sums business spends on IT these opportunities are meagre. Paul Strassmann, former chief information officer at Kraft, Xerox and then the Pentagon (IT budget: $230bn, give or take a few million dollars), claimed in his book The Squandered Computer that in 1995, US corporations spent $500bn on IT – that’s $175bn more than they made collectively in profits. And spending has increased since then. Are the soaring stockmarkets of the last two years anything to do with this spend? Intel and Microsoft stock has risen as a consequence, but there are plenty of other factors which have boosted company equity outside the technology sector. Has productivity soared? Up to a point, but there is inevitably a law of diminishing returns on IT-generated productivity. Early payroll computers meant you could sack roomfulls of clerks, and database analysis tools can cut swathes through your marketing spend, for example. Many digital transactions are now done almost instantaneously. Is there a tangible benefit to having them performed absolutely instantaneously? Can one measure the return on such an investment? And what happens after that? How much more do we have to spend to get a fraction better at things we already do? The problem of calculating return on investment in IT is in fact twofold. Before you even get to returns, you have to pin down what the investment is. The last two years have thrown this calculation into confusion, as total cost of ownership (TCO) considerations have forced most companies to junk straightforward purchase costs as a measure of IT spend and look at the intangible costs of their IT infrastructure instead. If it’s hard trying to measure the intangibles – support, the cost of downtime, additional charges on phone costs, even printer toner – then, when you add in the fact that most companies of a certain size simply don’t know what systems they have installed, the process becomes almost impossible. Here’s an example: right now, how many users of Word are in your company? And how many copies of it? And what version? And how many licences for it? And what do they cost? In many companies, the IT assets themselves are simply unknown. And how do you measure returns? You might make the execution of orders quicker or be able to reduce headcount, but what about the downside? If your Internet roll-out means half the staff spend an hour a day looking at porn or sports websites, the net returns start to look shaky. And ongoing support of IT can sometimes be more expensive than the wages of the clerks you fired in the first place. But let’s not get ahead of ourselves. Measuring the investment in IT is the first hurdle, and systems management companies are now offering tools to help businesses understand just how much they are spending on IT. One of the leaders in this field, Maincontrol, sells an asset management package which aims to tackle some of these problems. The idea is to audit all the hardware and software that’s in the enterprise and measure the total cost of ownership associated with corporate IT. From that figure, you can look for places where serious bottom-line savings can be made. As Alex Pinchev, president and CEO of the Virgina-based company, says, “There’s a lot of pressure from shareholders, in the US at least, to translate IT investment into corporate value.” Pinchev is keen to point out that distributed computing (a PC on every desk linked to a server which is connected to other servers) is the key problem here. “(Asset management) was easy on a mainframe: it was one item on the ledger,” he points out. Once the computing power started to spread out, measuring both investment in the infrastructure and the consequent returns became problematic. “If you don’t know what you have, you can’t even deploy new applications correctly and you can’t plan your corporate strategy,” he continues. “If your initial numbers are wrong, your tax, depreciation and insurance are all wrong as well.” Take, for example, an IT leasing contract up for expiry. If the contract has been badly handled because asset tracking has proved impractical, a company may end up spending huge sums having to buy obsolete hardware that it didn’t even know existed. Equally, a software audit could reveal that a company has purchased 1,000 user licences for a piece of software but 700 of the users have either deleted the application, never use it or have never installed it in the first place. These situations are clearly not delivering return on investment. In simple terms, Pinchev is arguing that whatever the return is, it can only get better by slashing waste and getting to grips with exactly what the investment to date has bought. According to consultancy IDC, something like 70% of all systems management implementations fail. Pinchev quotes this figure to illustrate that even recognising that there is a problem won’t give you a solution – and, of course, that rival systems management products are missing a trick. But another statistic which should worry FDs rather more is that in recent surveys, Price Waterhouse found no companies with a completely accurate accounting record for their IT investments. Maincontrol also recommends a total lifetime system for monitoring IT assets. “If you have accurate depreciation information, you might find that it’s cheaper to dump the machine rather than repair it,” points out Pinchev. And even dumping it can yield a return. Companies should have a well-planned retirement programme for obsolete IT kit. The modular nature of PCs makes them perfect candidates for cannibalisation; some machines could be sold off, and the software licences transferred internally to keep ongoing costs down. Even giving the machines away could generate psychological returns, whether it’s to a charity or to employees for home use. To some extent, the industry’s obsession with cost of ownership is making it easier to hold down IT outgoings. And new technologies such as thin-client computing, if they deliver the savings and increased asset management promised by the likes of Oracle’s Larry Ellison or Sun Microsystem’s Scott McNealy, will make life even easier for the FD. But the payoff, the ‘return’ part of the ROI calculation, is much harder. It’s just not as simple as looking at increased profitability or reduced overhead in the ‘user’ parts of the business. In The Squandered Computer, Strassmann illustrates the point with a comparison between IT spending per employee and return on equity in 468 US corporations in 1994. The resulting scattergram – which shows absolutely no correlation between spending on computers and profitability – makes an apt cover for the book. Strassmann covers a wide range of issues, not least the politics of computer administration in organisations. By divorcing IT from the mainstream of the business – simply referring to it as a cost centre – companies are missing potential returns. So IT spend should be rigorously examined and, if there’s no tangible business benefit, junked. Annoyingly for IT managers and FDs, there are simply no ‘best practice’ templates for a successful, high return IT project. Strassmann notes that 31% of computer projects in the US are cancelled, and 53% will overrun their budget. True, by applying sophisticated metrics like economic value-added, IT projects can be assessed on a case-by-case basis. But failed and over-budget IT investments can never make the returns they promised, even if there was an accurate forecast of additional revenues or savings related to them in the first place. And according to Strassmann, for 55% of US firms, the IT budget actually exceeds the company’s economic value-added anyway. IT, quite literally, is eating away at the true value of the shareholders’ investment. He stresses the need for companies to cut spending whatever happens, since without demonstrable returns at project inception, IT investments simply suck money out of the bank. The bottom line is that IT spending is starting to run out of control. The counter to this argument – one beloved of the IT vendors, of course – is that ROI itself is the chimera. In many industries, investment in IT is not actually supposed to generate identifiable returns. It’s just the ante to get into the game. In high-tech industries, such as Internet commerce or financial services, for example, this may be true. It’s also inconceivable that any bank would do without, say, cash machines, clearly a useful technology which has generated returns over the years. But is there any benefit to continual upgrades of ATMs? Once NatWest has upped the ante and got holes in the wall, how much additional return does it generate from replacing them all with a newer version that uses, for example, retina matching rather than a PIN number to confirm the customer identity? It might save a fraction on fraudulent transactions, but there may be greater customer and shareholder value to be had in simply not spending money on the upgrades in the first place. In effect, this divides up IT projects into those which generate measurable returns by cutting costs, those which create business opportunity and the ones which keep a company competitive. This division is also supported by Paul McNabb, president and founder of Cambridge Information Network (www.cin.ctp.com), a forum for senior IT and financial decision-makers to discuss best practices and IT strategy. McNabb sees distinct differences between Europe and the US in assessing ROI, based exactly on this model. “In Europe, (the IT function) is much more likely to report to an FD, so IT is a cost to be controlled,” he says. “In the US, IT is seen as a source of business advantage. The US has much longer time scales for returns on investment. In Europe it’s one to two years, whereas in the US companies it tends to be three or four years. “And the decision-making process is different in the US, where the IT department takes the lead; in Europe, it tends to be done much more by the business units.” But the US model may not always be the best solution. Indeed, one of the reasons Strassmann is so vociferous about cutting IT spend is that the technologists are more likely to champion investments without necessarily measuring potential returns. But like the rest of the industry, McNabb is still keen to emphasise the importance of measuring both intangible costs and benefits. And so the wheel turns full circle. You need to be able to measure the computing assets in the organisation. From this, you can start to understand what the investments actually are, both for ongoing operations and new IT projects. And then the returns become easier to quantify, and you can plan more effectively the next round of investment. At the very least, even if ROI is hard to calculate, companies can get closer to understanding the true costs of IT, and may even start to think about where the returns might come from. And even if the task is impossible, it’s still worthwhile making the assessment to see whether, if returns can’t go up, investment in IT can come down and the capital redeployed to parts of the business where value is easier to find.