Payroll Debacle Ignores Staged Tests

Posted under: From the Blog 12th of July, 2010

The Auditor General’s report of the Queensland Health Payroll Debacle is public on QAO website¬†

The QH/PW response is silent on staged implementation as if they have missed that key point. Errors would have been caught by the safety net that staged implementation provides, if it had been done. It is incredible to me, as an experienced project manager and former senior IT exec in Queensland Health in the nineties, that it was not.

In the nineties, we always staged big systems successfully into at least two hospitals first before rolling them out across Queensland. We usually started with one large and one smaller hospital because we knew they had differences the systems would have to cope with. And we picked hospitals that were willing and able to be first.

I am surprised by the agencies’ lack of recognition of this fundamental error of project management, and worried they have not learnt this fundamental lesson. ¬†Staged implementation is essential for big systems.

No system of this scale or complexity should be implemented in one stage – regardless of how much testing has been done prior to operational use.

The QH/PW response includes the laundry list of testing that had been done – not unusual for pre-production testing of a system of this scale. But no amount of pre-production testing removes the need for staged rollout of such a big system into production.

There is no test like the real world, with all its diversity of local interpretations, statistically turbulent data in the wild and inevitable unrecognised process quirks. That is what staging is for, progressive exposure to the harsh tests that real operations provide.

Staged implementation exposes a new system to its harshest tests in a risk managed manner, with scaled back damage and hence control, and easier rollback to old system if necessary. Because far fewer people are hurt by the inevitable errors. Because early stage hospitals are chosen for capabilities like resilience and culture of innovation, they handle teething problems better. What would have been teething problems for a couple of capable hospitals to help sort out, became a debacle because they were foisted onto an entire health system.

If the system had been staged into a couple of hospitals first, the problems would have been obvious but manageable. And they could have been fixed before wider rollout. No project is perfect and staged implementation provides an essential opportunity to validate the workability of a new system as it moves from its artificial test and development pre-production life into ongoing real world use.

Staging implementation is difficult, inconvenient, and costs money of course. It’s as hard as improving highway systems while cars are driving on them – difficult and expensive, but necessary.

Agencies should not be allowed to cut such fundamental corners. Or pretend that the problems were unforeseeable, like an act of God. The implementation should have been staged.

Let’s see if the next large state government implementation is.

1 comment. Would you like to comment?

Add your comment