Jaa


Why WiX: Modeling the centralized setup business

I've invested my career in setup for longer than most and I've seen setup for a few product cycles for different businesses. From this collective experience, I've derived a model of the centralized setup business cycle. Business types tell me there are big wins to taking cyclical business and flattening it out. Before we talk about possible solutions to flatten the cycles, let's talk about the problem so we know what we know what we have to fix.

Introduction to the Model

The y axis here is volume of work and the x axis is the stage of the Microsoft product cycle.

The "bathtub curve" that runs through the graphic is a generalization of the workload I've observed for a centralized setup team. This bathtub curve forms the essence of the centralized setup business cycle.

There are two dashed lines that run horizontally through the graphic represent the two resource-ing approaches I've seen to funding setup. The higher line represents the resource level that would allow a centralized team to absorb the load of the product at it's peak: the beginning and end of a product cycle. The lower line represents the resource level that would allow the centralized team to keep continually busy even during the slowest times of the cycle: the middle of the product cycle.

If you're the manager, which dashed line do you choose? In my experience, most managers choose to resource the centralized setup team at the lower line. This is certainly a justifiable choice from a business and shareholder point of view. If you're the manager that selected the bottom dashed line, how do you account for work that occurs at either end of the bathtub?

The historic choices I've seen to fill-in for the difference between peak and slack loads include:

  • New Hires: slot new hires that are hired during the product cycle into these tasks.
  • Shift Load: take disciplines that are supposed to be slack at this time of the cycle and have them do the work.
  • Shift Talent: take talent from other parts of the product and shift them to focus in these less attractive parts of the product (carrot: better rewards for those with wider scope; stick: just gotta do it to ship).
  • Hire Contractors: pickup slack with folks from temporary agencies

Each of these solutions eventually get the work done and the product shipped (yeah! party time!) but the long term impacts of each of these choices weaken the setup domain (in my experience). The impacts to the folks that choose to invest in this domain are also captured in the colored zones in the back of the graphic. That part of the problem will have to wait until tomorrow.

That'll have to do for tonight. More tomorrow…