Why do so many of today's systems replacement projects end in disappointment? On the face of it, the prospects should be favourable. By definition, these companies are experienced users of at least one other system, so they should have a shrewd idea of what they want from the replacement, and be aware of the pitfalls to avoid.
The problem is that, typically, the amount that smaller companies feel able to spend on a new system is barely enough to replicate the functionality of their existing one.
A limited implementation budget is quickly eaten up by mundane, unavoidable stuff: the monthly and statutory reporting, the basic management information. After all the pain of a six- or eight-month systems replacement project a company can find itself with no more functionality than its original system was giving them.
Replacement is all very well for a large corporation with the resources to go on developing the new system beyond the initial implementation phase. But few small or medium-sized enterprises have this luxury. By the time they have got back to where they left off with their old system, they have blown this year's IT budget. So whatever wonderful possibilities the new system was supposed to open up, such as Web-based access, integrated management reporting, or online customer information, in practice many organisations are unable to achieve these.
But was the software really the problem? The irony is that the system probably did not need to be replaced. The majority of companies already have a package that can do most of what they want, but they don't fully understand its capabilities.
If a system was bought from a reputable source at any time in the past 10 years it almost certainly has all the business information that the company needs somewhere inside it. The company can usually get the extra functionality it is looking for by adjusting the way it uses the system, or by adding some of the highly sophisticated software products and tools that have been written by the current generation of developers, such as business intelligence or knowledge management software.
It is not surprising that companies are over-eager to throw out their old systems and begin again. Human nature dictates that when you are trying to run a company you blame anything that goes wrong on the system - it can not talk back, after all. Often, too, the use of a system has degenerated over time. The first users may have only understood 80% of its capabilities, and by the time they have trained their successors the knowledge that survives is likely to be only 80% of that 80%, and so on. When a system has been in use for several years there is unlikely to be anyone using it who still understands anything like 80% of its abilities.
Not surprisingly, when the new man or woman at the top takes control and inherits a system there is a strong urge to start with a clean sheet and software suppliers are sometimes guilty of encouraging companies to yield to those irrational urges. These are the suppliers that are only too happy to sell a licence and do the minimum required to get the system up and running, before they rush off to sell to the next client.
The alternative to "knee-jerk" systems replacement is to take a closer look at what you already have, preferably with some expert help. Companies are often surprised when they discover what is possible with their existing systems. A modest investment in consultancy can reveal forgotten capabilities, or suggest simple enhancements that will deliver big returns and restore confidence in an older system.
This may allow the current IT budget to be re-deployed to address other areas not currently integrated or even automated, such as customer relationship management or service management.
Of course, there are times when an old system genuinely does need to be replaced. If it is older than 10 years it is likely to be costing you more in maintenance than it makes sense to pay. If a system does not use standard "open" technology, such as relational databases, there will be a limit to its ability to work with other software, such as reporting tools. Or, if your business has changed so that, for example, you need multi-currency capability that is not catered for in your system, then replacement may be inevitable.
Even so, you may find that you can get what you need by moving to a more up-to-date version of your current package. An upgrade usually costs a fraction of what you would pay to license and install a completely new application. And with an upgrade you are not losing all the effort you have already put into the system.
It is important to be aware of what new technology has to offer, but to make best use of a limited systems budget, financial directors and IT directors should temporarily forget about the march of technology and ask themselves instead what aspects of their business have altered since the original system was implemented. There is a strong probability that the requirement for core business systems is the same as it ever was, although peripheral needs may have evolved.
Why spend a limited IT budget on replacing core systems, if all you are going to end up with is prettier screens? How much more worthwhile to build on what you have and invest your money in meeting the needs that actually have changed. That is the way to get identifiable and tangible return on investment.
Having analysed the gap between what the business needs and what it already has, the company can formulate a rational strategy for systems enhancement. The missing or inadequate elements should be addressed according to the priorities of the business, one component at a time.
With today's open architecture, there is a lot of flexibility as to the order in which this can be done. If the "point of pain" is unsatisfactory reporting, for example, it should be possible to get a good suite of reports now, then replace the underlying business systems later with minimal disruption to the reporting.
For less than the cost of implementing that flashy new package at the most rudimentary level you can probably increase the usefulness and sophistication of your current system beyond recognition. It will cause considerably less disruption to the business, and you can pretty much guarantee that you will end up getting what you wanted.
Gary Waylett is managing director at Eclipse Computing