Back in January of 2013, the City of Seattle first contemplated a new $43 million customer care and billing information technology system for its two utilities, Seattle City Lights and Seattle Public Utilities. By the project’s launch a year later, the scope had expanded and the cost had grown to $66 million.
Currently, the final budget for the New Customer Information System (NCIS) is $109 million. The time to complete also stretched from the originaly estimated 21 month schedule in January 2014 to 32 months in the real world.
The Seattle City Council turned to the Seattle Office of City Auditor to find out why project cost and schedule grew and why council members weren’t informed along the way.
The NCIS audit, which was released in April, is a great primer for any government official who is involved in managing an IT project. Many of its findings are familiar to us, as they recur in reports and audits about project management.
In our view, the most universally applicable points made in this audit can be summed up in these four statements:
The initial schedule was unrealistic. The audit details the drive to establish an aggressive timeline “to save costs.” This meant reducing the projected schedule by three months from the 2013 estimate, even though the project’s scope had risen significantly by its launch in January 2014. Nearly all the cost overrun came from additional labor costs when the project timeline subsequently grew by 11 more months.
Elected leaders were not thoroughly informed about the frailty of original cost estimates. In a complex IT project, there is always shakiness in initial project estimates, this “cone of uncertainty” grows even bigger when two separate departments and multiple applications are involved, as was the case here. There were also limited procedures in place to inform the Council when the budget and timeline started to grow.
Important risk factors were identified, but not addressed in a timely way. The audit describes the work of a well-respected Quality Assurance expert, who was part of the project team, and identified 10 high risks for the project, including a workload that was too high for existing staff and resources. According to the audit, meeting minutes for the project’s Executive Steering Committee show that the workload issue was brought up in August of 2014 and in 13 subsequent meeting reports. But actions never sufficiently addressed that problem. As the QA expert wrote in a May 2015 report, “this issue has been the root cause for slippage in many other areas of the project and is the primary driver for the decision to re-plan the Go-Live date.”
Initial training of project staff was insufficient. As the audit says, “some City team leads had little or no project management experience or experience with implementing a new IT system, let alone one of this scale. While City IT staff were provided with some training, it focused primarily on an overview of new software applications.”