In March this year when the Prudential pulled out of the general insurance market, the decision was announced to its UK brokers by fax. Electronic information distribution was the easiest way for the Pru to inform all its business partners simultaneously of the news.
What the company may not have realised was that within hours of the fax being sent out, it had been distributed electronically to all the Pru's rival account executives at Sun Alliance. No sooner had a copy been acquired by Roy Barton, national sales manager at Sun Alliance UK's broker division, than he scanned it in to his computer and sent it down the line to all of his 60 account executives and managers. Each was then in a position to decide whether the Pru's withdrawal would create a sales opportunity.
It was not the first time Barton had thanked his decision in 1990 to go for an electronic Big Bang. During the past 12 months, his new computer system has saved around £1 million and enabled sales executives to increase their average time spent with brokers from less than two hours a day, to nearly five. "The more time they spend face to face with brokers, the more business they generate through building up better relationships," says Barton.
All his salesmen now have portable computers which can be linked in to host computers on a local area network at the division's headquarters in Oldham. By re-orientating their lives around the mobile machines, salesmen have been able to abandon their offices for a life on the road travelling in the opposite direction to the rush-hour traffic. When they get home in the evenings they transmit details of the day's business to head office via the phone, and download any fresh information concerning their own accounts. Salesforce confidence is riding high thanks to the new-found keyboard skills. Meanwhile, brokers have responded enthusiastically to requests for data about their businesses.
The notebook computers in question were bought off-the-shelf. Such is the pace of technological development that Barton knew they would be out of date as soon as he got them. "Sure enough they were." But he isn't really complaining. He's had a year's usage already.
Besides, the real weapon is the software. It creates individual profiles of each broker, including details of how their business is set up and directions in which they intend to develop. Salesmen can review this before each call, plan objectives for the meeting and subsequently evaluate their success. The information is also transmitted back to base so that sales managers can monitor the progress of individual salesmen.
Through its link to the host computer at head office, the system provides management information statistics such as an individual broker's profitability, which products he is selling well, how much cash he has generated in premiums, and whether his account is paid. This then enables the salesman to help the broker spot new opportunities such as selling health policies to motor insurance holders.
Barton spent a long time looking around for a shrink-wrapped package that would do the job. Nothing was sufficiently flexible to accommodate all his requirements. So he opted for Lotus Notes, a programming toolkit, with which to build his own applications lego-style in partnership with the IT team. "It is a superb applications development tool," he says. "In half an hour we can now devise programs that would have taken weeks on the mainframe." This is not just faster, it also puts the power to specify, create and develop each application in the end-users' hands.
Barton's approach highlights the trend towards enabling users to develop their own programs, rather than leaving it to rarified boffins in the data-processing department. This produces far more satisfactory results and enables computer users strapped for cash to cut down on the soaring costs of software.
Hardware vendors, who once gave away programs with their machines, are increasingly turning to software to make up for their lost revenues. Last year, European companies spent £5 billion on applications programs, 10 times what they paid out a decade earlier, according to Peter Lines, managing director of Input Europe, an international IT consultancy which specialises in software. During the same time-span, the proportion of IT budgets spent on applications software doubled from less than 2% to 4%. Input forecasts that it will increase similarly by 1996, to 8%.
The bulk of expenditure is not on new programs, but on the upkeep of old software which is difficult to maintain but which companies cannot afford to throw away. "Around £3.5 billion a year, two-thirds of UK spending on IT operations, goes on the cost of maintaining old programs," Lines says. "Some is 25 years old and still critical to business." There is also a vast log-jam of applications awaiting development.
While the performance of computer hardware is doubling every 18 months, software has failed to keep up, says Lines. The first business programs, which were introduced in the mid-1970s for administrative tasks such as payroll and accounts, achieved instant productivity gains. Further efficiency improvements were made from 1975 to 1985, when software was applied to operational tasks such as warehousing, point-of-sale and factory planning systems. The problem became apparent with the move towards office systems. "Staff are staring at desktop PCs which are far too difficult to operate," says Lines. There is no evidence that office computers have improved productivity.
This view is shared by Malcolm Padina, managing director of Informix, a database supplier. Padina believes that the main problem comes from computer vendors who concentrate purely on selling powerful boxes, without providing software to exploit them fully. Some 90% of the PCs in most organisations are used between only 5% to 10% of their functionality, he argues.
A user who complains to his hardware supplier that his computers are not giving value for money, will be told the answer is to buy a bigger and better machine," says Padina. He cites the large number of users who were happy with dumb terminals yet were persuaded to buy PCs. "After that they made the move to Windows, and soon they will probably be sold workstations running Unix. At each step they only realised a fraction of the performance advantages of the machine."
Sophisticated extras such as graphical screens and the mouse pointing-device have done much to increase the user-appeal of computers, but they can actually have the effect of slowing users down, Padina maintains. "Telephone operators in the US, who were given systems with a mouse, became far less productive than they had been on old keyboard systems where they had to type entries into boxes on the screen."
Snazzy displays and on-screen icons are also bad news for program development. The inconvenience that they remove from the end-user's environment is transferred directly onto the software developer, points out John Markel, president of SmartStar, a US-based supplier of programming tools. "Managers think graphical interfaces are great because users can be trained much more quickly and seem happier using the systems. Unfortunately, to create such systems requires a new and very difficult style of programming."
The result, Markel maintains, is that in future software is going to take more, rather than less, time to develop. "Instead of average development times plunging from two months to two weeks, as managers have come to expect, they could easily double to four months," Markel says. "It is a huge problem that very few businesses have begun to realise."
For companies keen to make better use of their existing machines, the easiest solution is to buy a ready-made software package. This is the cheapest and quickest approach for the 90% of activities that are not key to your competitive edge. As Lines at Input Europe puts it: "There is no point in reinventing the wheel. For tasks such as payroll and accounts there is almost always an application on the shelf."
Even if it means changing your procedures a little, such is the cost of software development that it is almost always worth adapting to fit the package. Far better to have 90% of an application computerised now than 100% in perhaps a year's time.
But for the 10% of the business that gives you your leading edge, perhaps management information, engineering design or customer services, you may feel, like Barton at Sun Alliance, that it is worth developing independent software.
For development staff with technical expertise, there are a number of interesting tools to be considered. Fourth generation languages (4GLs) make it easier for system buildings to specify their requirements, and computer-aided-software-engineering (CASE) tools help them develop software on time and under control.
The technique currently in the limelight is the one called object-oriented programming. Though still very much in its infancy, object-orientation should do much to speed up the development of complex applications, reckons David Flint, managing associate at IT consultancy CSC Index.
It works by breaking up the classic one-piece program into independent components such as communications, generating visuals, or mathematical processes. The parts can then be pieced together to build a total application.
It differs from conventional modular programming in that the components are very small," says Flint, who has seen the method speed up projects, and improve error rates, 14-fold. A strong set of international standards is being devised so that software from one developer should interlock with modules from other suppliers to make instantly working programs.
"Whereas 4GLs are good at relatively straightforward problems, object-orientation excels at complexity - the more convoluted the problem, the more advantage it provides," says Flint. So far, the method has been largely confined to technical applications. But in future, more general business applications are likely, such as warehousing, ledgers or customer accounts. The jury is still out on object-orientation, reckons Lines. "Software typically takes 15 years to move out of the lab and develop the robustness needed for widespread use. There is enormous inertia on the part of users to increasing their investment and add layers of icing and marzipan on the cake." For example, so-called relational databases have offered far more power than conventional systems for quite a number of years, yet they still account for only 7% of databases.
Your database may seem like a vast unintelligible bank of data-fields, yet it is one of your greatest assets. With the right piece of software on the front it could deliver vital clues as to where your business if failing or how to spot new customers. This is the aim of the Executive Information System (EIS).
You might want to know what products are selling at what price and where, your average response time to customer queries, or a breakdown of staff salaries and length of service. A good EIS can slice and dice the database however you like. It gives a clear and simple window on the company through which to focus on whatever areas is of interest.
Originally, aimed at only a handful of senior executives, the EIS has proved to be so immensely popular that there are now a number of organisations with close to 1,000 users, according to Ian MacNaught-Davies, managing director of Comshare, one of the market-leaders. "If your boss has one then you need one too," MacNaught-Davies says. "The empowered workforces and flatter management of the 1990s cannot function if the guy at the top sits on all the information."
According to the conference, publishing and research company, Business Intelligence, an analysis of organisations currently using EIS shows that proprietary systems now account for the vast majority of installed systems.
But software is not a panacea. Real gains in productivity and profitability require business re-engineering, maybe switching to electronic teams that share data as at Sun Alliance. "It is no use just throwing software at existing structures and processes," says Lines. "You might give someone a spreadsheet which enables them to accomplish their work in an hour's less time. What do you do then? Pay them less? It may be that they get on the phone and talk to their friends for an hour, in which case your costs increase."
The most important trend in software development is the involvement of end-users, says Lines. "Computers are no longer kept in glass rooms and tended by white-coated high-priests. That set-up has been blown apart." Now they are on desktops, and users can program them to do whatever they want.
Lines confidently expects that there will be a quantum leap in software during the next decade, though he is not very clear about the direction it will com from. In 10 years time he believes we will look back at today's software and laugh. It will no doubt seem "as old-fashioned, cumbersome and unreliable as vintage cars do now - like part of a bygone age when cars were modelled on horse-carriages."
For reprints of this article, contact Anne Oakley (071) 413 4336.