An Analysis of the Canadian Government’s Phoenix Payroll Disaster, Part 2

In the August issue of The Word in Workforce Management, you read my blog where I introduced you to one of the largest workforce management project disasters ever: the Canadian Government’s Phoenix implementation.

To recap, in 2009 the Canadian government began a pay transformation initiative for their 290,000 employees in 101 departments and agencies. The initiative had two main goals: replace their 40-year old payroll system; and centralize their regional pay offices who process pay for approximately 70% of their employees.

The system continues to experience major issues and still makes news headlines. As of June 2018 (the latest data available), there were 187,000 employees who still had outstanding pay complaints. The anticipated annual savings of $70m have evaporated and been replaced with an expected cost of $2.2b to fix the system (although there are doubts it will ever work correctly). Employees report being scared of receiving their pay cheques, never knowing whether it will be accurate or not, and the largest labour union has asked for compensation from the government. The government is also in a very weak position now that it is time to renegotiate labour contracts.

In my last blog post, I summarized the problems that led to this disaster. This time I provide insight into how those problems should have been handled. Keep these tips in mind when you are planning for any type of workforce management transformation project.

Problem # 1: Attempting to transform technology, people and process all at the same time, with too much focus on technology and not enough on people and process.

  • Don’t under-estimate the undocumented knowledge your payroll department has built up over decades of service. These people have hundreds of processes in their head that they execute pay period after pay period. The more complicated the labour agreements, and the more independent the different departments are within your company, the less standardization of pay practices exists. Ideally, all these processes should be documented before any WFM project begins, including any exceptions or one-offs that can occur. These then lead to accurate and realistic requirements for the project.
  • When moving to a new system you must assume a ramp up time for skills and efficiencies of employees, managers and payroll staff. This must be included in your ROI calculations. This ramp up time correlates to the amount of change in technology and processes you are implementing. If part of your anticipated savings includes being able to reduce payroll staff, these changes should not be made until system is stable and proven. All staff, whether new or old should receive enough training in the new system as well as the new processes.

Problem # 2: Cutting corners without understanding the impact

  • Reducing scope due to time and/or budget constraints must be done very carefully and only non-critical functions should be considered. A full impact analysis is required to determine whether it is feasible to cut material items from your overall project scope. If the alternative plan is to replace them with manual processes, the time taken to perform these processes should not be under-estimated – a true accounting of the real cost is always necessary.
  • With large, complex systems, it’s worth considering independent verification of testing to ensure sufficient coverage and proper validation of the results. Independent verification of project go-live status is also worthwhile, as it removes the potential lack of objectivity from the project team as to the true status of any open issues.
  • When it comes to rolling out the solution in large companies, nothing beats a pilot project for reducing project risk. A pilot allows you to find all the issues in the system without the pressure of affecting all your employees at once. If the pilot is successful, then rollout should proceed in waves, with one wave being fully stable before moving on to the next wave.

Problem # 3: Lack of truly independent oversight

  • Independent oversight is critical and it must be truly independent. It should not report up through the project management structure – rather, it should exist in parallel to it. The purpose of independent oversight is to ultimately deliver bad news if needed. If it is not tied to the success of the project, the advice provided will be much more objective and useful.
  • While this oversight may have costs associated with it, those will pale in comparison to the costs of project delays and budget increases due to failed oversight. When project risks arise, they should not be sugar-coated or downplayed, especially when it involves employee pay. It is important to create a project structure and culture where issues can be raised without hesitation in order to address them immediately. Bad news never gets better with age.

Problem # 4: Lack of contingency planning

  • You must have a robust contingency plan for any transformation project. Especially when you’re dealing with employees’ pay, mistakes can be costly. And when those mistakes occur, your employees will be vocal about it – both internally and externally. So, be prepared and have a plan for when things go wrong – and communicate that plan quickly and clearly to your employees.
  • Make sure you provide enough support to your locations and ensure proper communication channels for all types of communications. Bad news is hard to receive, but no news is even worse. It’s important to keep communications open (and honest) regardless of the state of your project.
  • While no project is risk free, a properly run project will yield a good understanding of the risks and the proper risk controls in place to ensure it is successful. Look to your internal experts to help with your plan. Where you don’t have all the right resources you need, bring in external help so you minimize your points of failure and give your project the best chance of success.