A crash waiting to happen

THE WIDELY REPORTED difficulties experienced in implementing eNatis – the Transport Department’s new system for licensing vehicles and drivers, plus tracking accidents and traffic contraventions – begs the question of what lessons could be learnt to help avoid such problems with future major Government projects.

The State is expected to spend a significant amount on new IT projects over the next few years, including the biggest: its estimated R11bn integrated financial management system. We asked experts at companies that stand to benefit from future Government spending on IT projects what might be done differently to avoid possible problems similar to those experienced with eNatis.

Derek Wilcocks, strategy executive for Dimension Data South Africa, says although the company wasn’t involved in eNatis in any way, there are probably lessons that could be applied to any large IT project. Those include spending enough time analysing the business requirements, including the system, but also aspects of change management, such as usability requirements and risk. Those involved in that analysis should include actual users of the systems, not only the implementation experts.

Wilcocks says service providers must spend time translating the business requirement analysis into functional and architectural specifications, as it’s often tempting to cut corners in that regard to ensure a speedy implementation despite it causing problems later.

A substantial portion of any large project – anything from 25% to 100% of the development time – should be spent on testing the system with staff and users, Wilcocks says. Up to 25% of time must also be spent on change managing, including formal and hands-on training, skilling super users to support their colleagues and preparing manuals.

Business Connexion (BCX) group executive for outsourcing Mike Sewell says both the client and the service provider must often take the blame for the failure of large-scale IT projects due to immature relationships between them. Clients want fixed price contracts, with specific timeframes. However, a set price contract is a mistake, particularly with high-risk projects.

Sewell says the risk of a project must be clearly articulated up front so that realistic expectations can be mapped out. Even organisations such as US space agency Nasa, with its massive funding and intellectual capital, overshoot budgets and deadlines on large projects, he says.

In fact, high-risk implementations can take double the amount of time and cost twice the price originally estimated, he says. Cost and having the right credentials often wrongly seem to take precedence over a company’s ability to do the job and its track record with previous implementations, particularly if the procurement department awards the project.

Irnest Kaplan, an independent technology analyst who recently authored a report on Government spending for T-Sec, says major projects should be awarded to players according to their specific strengths, as happened with the recent SA Revenue Service contracts.

BCX designed the original Natis but not the eNatis system, which is entirely new. And contrary to reports that it had been called in to bail the project out there had been no such formal invitation from either the service provider (the Tasima Consortium) or the Transport Department, says BCX group executive for professional services Marthinus Wissing.

Sewell says where a "sudden death" approach is necessary – switching the old system off and the new one on immediately afterwards – because one design is so radically different from another, thorough testing must be conducted to ensure it will work and that the equipment it’s running on is correctly sized. It should also be phased in gradually.

Bytes Technology Group CEO David Redshaw says that although he’s not familiar with eNatis, anybody who thinks you can implement a significant software contract without problems is "living in cyberspace".

Redshaw says there will be problems with a project of this size, no matter who implements it. However, there are added risks when giving such a project to a consortium that hasn’t worked together before. Clients and service providers need to carefully consider how best to implement such a project. Training of users is crucial.

Kaplan says many other major project implementations also face significant hiccoughs and need to be phased in gradually, with a specific focus on the "softer" issues, such as training. eNatis was a particularly high profile case but should serve as a warning to other Government departments not to repeat the mistake.

Comments

Finweek

Latest news

Share this article
A crash waiting to happen