Projects completed
Of the 174 major IT projects that were either current or initiated over the last five years, 168 had been completed or were still in progress. Three had been dropped and another three were on hold.

Costs in relation to budget
A number (14 percent) of completed projects came in over budget. The total over-expenditure, however, was only two percent ($3.8 million) of the total of the budgets for all completed projects ($190 million). Three projects went over budget by 20 percent or more.

There was under-expenditure on projects, which corresponded to four percent ($7.2 million) of the total of the budgets for all completed projects. Overall, therefore, there was a two percent under-expenditure on the total budget.

Definition of failure
Few IT projects end up exactly as planned. Statistics from the USA indicate that only nine percent of IT projects in the public and private sector are com-pleted on time and within budget; 31 percent are cancelled before completion; and of those completed, 53 percent cost, on average, 189 percent of their ori-ginal estimates . These original estimates⁴ are commonly based on sophisticated project ma-nage-ment, with precise goals for budgets, staffing, functionality and timelines. More than 80 percent of new system initiatives fail to meet one or more of these goals, but IT managers are constrained to assess these projects as failures.

In many cases, sophisticated set-ups, the increa-sing complexity of IT, and rapid changes in technology mean that projects are unlikely to meet their budget. At the same time, be-cause projects reflect key business processes that may be undergoing significant re-engi-neer-ing, new implementations of core systems may take much longer than planned because their delivery requires the interaction of many departments and layers of management.

Because large-scale projects involve complex technical and managerial planning, expectations of achieving project goals should be set within reasonable boundaries, and in the light of changing technological developments.

Gartner Group suggests that a definition for large-scale project failure is going over allotted time or budget by 30 percent or more. On the basis of traditional timelines for return on investment, this level of divergence will, in many cases, delay the expected revenue stream for new systems long enough to cancel their business value.

New Zealand Government experience
The stocktake shows that the NZ Public Service success rate is different to success rates overseas, as shown by available figures. With very few projects dropped or coming in 20 percent or more over budget, departments appear to have performed well in terms of completing projects on budget.

Links to strategic and business planning
The stocktake indicates that IT planning and IT projects are tied to departmental business and strategy in many, but not all, cases. 32 of the departments had an individual or group assigned to overall strategic business planning for the entire organisation. 33 had a steering committee that met on a regular or ad hoc basis to review projects and priorities. 134 of the 174 projects arose out of business strategy or business processes.

Good practice⁵ occurs where there are strong links between the corporate plan and all IT activities in the organisation. IT needs to be subject to managerial control, just like any other major corporate resource. This should ensure that an organisation uses technology only in order to meet its business objectives.

IT management structures
Even though the departments reported different degrees of geographical and managerial centralisation in their entire organisation and in their IT, 37 departments reported that they had a clear-cut organisational structure, with a clear under-standing of who was the ultimate decision-maker.

This indicates that there is, in most cases, a strong link between the organisational structure of a department and an organised, strategic approach to IT management.

IT standards
Good practice⁶ in IT management involves establishing and documenting standards for relevant items such as: software, hardware, communications and user interfaces, policies for user equipment support, data security, software piracy, virus control, induction and training of new staff, and work practices.

Respondents were asked to indicate whether they had certain standards in place and whether these standards were documented.

40 of the departments reported that they had desk top hardware standards, desktop software and peripheral standards and standards for shared equipment or software in place. 38 reported that they had standards in place for data and voice networks. However, not all the departments had these standards documented.

Good practice⁷ is to have a standard methodology that is used for all significant projects.

Good practice⁸ is for every project to have policies, plans and pro-ce-dures covering risks to either the department or Government. Of the 174 major projects listed, 165 were reported as being mission critical/important or process critical/important, and of such a nature that failure could affect the department's business or performance. However, a risk assessment was undertaken beforehand for only 116 of them.

This reflects a need for greater awareness of risk management practices.

4 Charting the Seas of Information Technology Chaos. The Standish Group International Inc, 1994

5 Review of Computing in the State Sector. PA Consulting Group, State Services Commission, 1991

6 Review of Computing in the State Sector. PA Consulting Group, State Services Commission, 1991

7 Review of Computing in the State Sector. PA Consulting Group, State Services Commission, 1991

8 Review of Computing in the State Sector. PA Consulting Group, State Services Commission, 1991