The race against time

Soaring costs and staff shortages are making it hard to defuse the 2000 computer time-bomb, says William Pavia

When in 1990, the millennium bug first reared its ugly head, it was greeted with incredulity. It was apparent that the majority of computer operating systems and programs in use were developed using six-digit date fields - YY/MM/DD. Come midnight, 99/12/31, the first two digits of this code would roll over to 00 and computers would be thrown into confusion, recognising the date as 1900 or simply ceasing to function. But surely the industry whose genius has allowed men to walk on the moon could not be susceptible to its own calendar?

Incredulity remains. SBC Warburg carried out a survey earlier this year into how far businesses had progressed towards 2000 compliance. The reply from the London office of a prominent French bank was that they were leaving it to their parent in France, on the grounds that they would hit the problem an hour earlier.

Incredulity among senior management is what still causes 2000 compliance programmes to be delegated to companies' information technology (IT) departments. However, escalating costs, inadequate resources and a revised deadline have made this a problem that has to be dealt with by senior management. It is about prioritising, managing scarce resources and close co-operation across sectors between firms that normally compete.

That costs are rising, and will escalate further, is well demonstrated by the financial services sector. The imperative there is obvious. The very basis of its main activities - the time value of money - will be disrupted unless 2000 compliance is achieved. This time last year the British Banking Association (BBA) estimated that it would cost its members (who constitute 95 per cent of banking business in the UK) pounds 50m. Twelve months on and the sum cost of just three major clearers is estimated at pounds 300m - Lloyds TSB, National Westminster and Barclays have indicated that their programmes will work out at around pounds 100m each.

Such figures remain entirely manageable - a spokesman for NatWest said: "It's a big programme, but it only represents 4-5 per cent of our total spend" - but the possible implications of these developments have not been missed. If anticipated expenses have risen this much in 12 months, what might they look like in another 12 months' time?

"Anyone who says they can put a figure on it is misleading you," said a spokeswoman for the Royal Bank of Scotland. "It's only when we get to the testing phase that the costs will become clear."

Most experts believe testing will comprise at least 50-60 per cent of any 2000 compliance programme. Transactions across the year 2000 must be simulated, both looking forward into the next century and back from the twenty-first century into this century, before what the BBA refers to as "cross-organisational testing" can begin. This is essential, because however punctilious individual banks may be, they will still suffer "knock- on" damage should even a small proportion of the banking sector fail to comply.

This shatters the delusion that sooner or later someone in the IT sector will create a "silver bullet" solution, and the millennium bug crisis will resolve itself into an enormous marketing opportunity for the lucky software vendor. Fixing the systems is only 20 per cent of the task; tools will, at best, only cover 12 per cent of the job. After testing, at least 20 per cent of a programme will be devoted to finding resources.

Indeed, finding the resources will be the key issue over the next two years, causing costs to rise further. "There already is a shortage," said Rob Wirszycz, director-general of the Computer Services and Software Association. "Never before have we had everybody sharing the same problem, and there just isn't the capacity. IT supply normally caters for about 25 per cent of British industry, not 100 per cent as is now the need."

The issue of legal responsibility has, until recently, caused suppliers to hold back from declaring the chronic shortages publicly. Then, two weeks ago, it was revealed that IBM was divided at senior management level on its 2000 strategy. The reason? The world's largest IT supplier is having to be selective in coming to the aid of its users, targeting only its most important customers. When the dominant computer company is floundering, one hardly holds out much hope for the rest of the field.

Such fears were given further confirmation last week when Cap Gemini, the largest European computer services company, publicly aired its concern over resources of software and staffing. "If you factor government estimates of a pounds 31bn cost to British industry, by IT industrial capacity, both currently in IT departments and available external resources, then you're talking about five or six years' supply of resources," explained chairman Geoff Unwin.

The financial sector only has 17 months to act. It has to be ready for January 1999, from which time many transactions will look forward into the following year. In any case, 1999 is full of critical dates: on 9 September, date fields will read 9/9/99, which many computers will recognise as infinity, or at least a good time to shut down. That the first day of 2000 will be a Saturday, whereas the first day of the year 1900 was a Monday, will cause computers to confuse Mondays and Saturdays. In fact, anyone who wants to beat the millennium bug needs to have completed testing by 1 January 1999, leaving the rest of the year free for contingency planning should any test results prove negative.

"Firms that have not started by now are running a tremendous risk," warns Robin Guenier, executive director of Task Force 2000. This is so even before one takes account of the staff shortage. "If you take the staff numbers on a competent programme like Abbey National's, and cross-reference them with those of BT, then you're left with a figure that can be applied pro-rata across the board. Making percentage allowances for smaller companies, you're left with a manpower requirement of 26,000," says Mr Guenier. The total currently employed is about 30,000. However, there are other pre- occupations - EMU, Internet banking, the new electricity market - and computer programmers would much rather be playing with the latest Java systems than tinkering with old mainframes.

So the scramble for IT staff has already begun, and it can only intensify as companies who have not yet got their act together try to catch up. "The odds are you're going to lose good people," says Richard Coleman, an analyst at Merrill Lynch. "They'll be headhunted in the coming months. The question is, can you pay enough to keep them?"

Addressing this question last week, the Cheltenham & Gloucester building society announced its intention to reward IT staff who stay until the 31 March 2000 with pounds 20,000 loyalty bonuses. The Royal Bank of Scotland's pay review is expected to include a 100 per cent pay rise for their IT staff. BT is "monitoring the situation on a case by case basis".

"The market rate [for staff] is rising and will go further," says Dr David Walton of CDT Enterprise Ltd. "With it will rise the cost of achieving 2000 compliance, but throwing money at the problem will not solve it." It has to be taken on at the level of senior management, not delegated to IT departments. For a start, it has long been argued anecdotally that leaving a project with an important deadline to the IT sector is a recipe for disaster. The Institute of Work Psychology at Sheffield University published a study last year which concluded that 80 per cent of computer system projects are not completed on time, and 40 per cent fail altogether.

The fact is that there are not the resources to do a proper job. To expand year-dates to four fields is now impossible, so programmers are working around the problem, rather than solving it. On some systems, the year dates 70-99 are going to be treated as twentieth century, and dates 00-69 will be recognised as twenty-first century. This still leaves pension schemes started before 1970 in a spot of bother, and discrepancies are likely to develop across sectors as companies adopt different cut-off points.

The way forward must start with wholesale co-operation between industries and software suppliers, which has hitherto not existed. Next week the Department of Trade and Industry, in conjunction with the Computer Services and Software Association and Taskforce 2000, will host the "Millennium IT Skills Shortage Summit". Senior managers of IT companies will mingle with senior managers of blue-chip industries. The intention is to galvanise industry into action.

Once this happens, the fruits of shared experience and several as yet only theorised steps could meet the growing shortages of time and resources. One proposed solution is that, as the work is fairly simple, an army of the unemployed could be recruited and set to work.

It has also been suggested that retired programmers be encouraged back to work. They would bring with them a wealth of experience on the older systems which are, after all, the ones in need of attention. With salaries set to rocket in the next few years it would certainly be worth their while.

Whatever means are adopted, it seems something must start happening now. For some time there has been a note of desperation in the voice of the 2000 lobby as it warns of the end of civilisation as we know it. So far the IT sector is putting its faith in increased productivity. Mr Guenier says: "It's amazing what you can do in an emergency." Maybe, but even this may not be enough.