Two research papers released by the independent research consultancy IBRS shows that the more effective information technology departments are those that spend less money on their systems but take more care about how they spend it. However, the papers also note that getting rid of legacy systems, in favour of loosely coupled modular systems is always a good idea, and the sooner the legacy systems are superseded the better.
The first paper, by analyst Rob Mackinnon, is a study which follows up a McKinsey & Co survey, released early last year of the spending patterns of banks. The McKinsey survey found that the banks judged as delivering the greatest value on their investment in IT were also the ones spending the least.
In his discussion of that result, Mackinnon notes that one mistake in free-spending IT departments is to pay too much attention to administrative systems such as those in finance and human resources.
"It is rare that investing resources in such systems will help the organisation achieve its strategic goals. The exception is when these systems are not performing in accordance with business expectations, in which case remedial action is warranted," Mackinnon says.
Spurred by the McKinsey survey, the IBRS study identifies 10 practices that denote effective chief information officers, including practices such as focusing on core businesses and IT issues, understanding the business at the early part of the planning cycle, taking a long-term view, and measuring performance.
The Mackinnon paper examines three of the 10 practices: focusing on the core business, developing strong execution skills, and understanding the business.
In discussing strong execution skills, the paper notes that the keys to these skills include "a project management methodology that is easily comprehended by users, project governance arrangements that engage senior business representatives, and ensuring that deliverables can be spaced at regular intervals through the course of the project".
But there will always be projects or project components that fall behind, despite the best corporate governance. The good chief information officer has to be ready to "euthanase" those projects, to free up resources for the others, the IBRS paper says.
As well as good management of projects, a good CIO becomes involved in the early stages of organisational planning.
"In most organisations, strategy development is an ongoing process. Though organisational goals and aspirations may be clearly articulated, strategies tend to become refined over time. Variations may be introduced for a variety of reasons including market forces, competitive behaviour, climate, legislative changes and overseas influences. Unless the IT executive is part of the strategy setting inner circle, subtle variations in emphasis may be missed when delivering upon the perceived strategy," the paper says.
One area where the CIO may want to spend a little more is in getting rid of legacy systems - systems that may have been cutting-edge 20 years ago, but which technology has long overtaken. In addition, the system software may be written in a programming language no longer taught.
In fact a computer system may reach legacy status and be ripe for replacement, after just a few years.
In another paper produced for IBRS, analyst Jorn Bettin says: "Large software applications can reach a stage - within a few years - where no one has a reliable grasp of all the design patterns used in the implementation, and of all the dependencies between the components that make up the application."
Once a project reaches that stage, Bettin says, it is unlikely to ever be completely replaced by a new system. In fact, any new system will often still rely on parts of the older system that it is supposed to replace.
"Over the course of two or three decades this leads to archeological layers of software stacked on top of each other that are very costly to operate and nearly impossible to understand within an acceptable time frame," the paper says.
Organisations with large systems may rely on advice from vendors but that also has problems in that the vendors want to lock the client into particular technologies and usually have less experience in large project software design than the client.
Bettin says a better solution is to build the skills of the IT team to incrementally switch to modular, loosely coupled systems. The IT team can also be properly incentivised to move to modular systems, as well as renovate existing code, the paper says.
©Fairfax Business Media
Join the CIO New Zealand group on LinkedIn. The group is open to CIOs, IT Directors, COOs, CTOs and senior IT managers.