How DevOps changes the delivery of IT functions
Source – cio.com
Labor arbitrage and shared services companies have had a perfect marriage over the last 20 years. Then along came the Digital Revolution with new business models and a new construct for services. One component of the digital model construct is DevOps. It makes a significant impact on business services, but it’s important to understand how it changes the picture for labor arbitrage and shared services.
Shared service companies are structured on a functional basis. One way to think about them is they are a stack of functional expertise. In the case of IT, the stack includes such functions as infrastructure, security, application development and maintenance, and compliance. There is a multiple stack hierarchy, with each functional layer having shared service champions responsible for delivering that function cost-effectively at a high level of quality. Labor arbitrage fits perfectly into this equation in that each functional layer uses people, and the work can often be done more cost-effectively offshore than onshore.
Into this mix comes the Digital Revolution. It has a new business construct, a component of which is DevOps. DevOps uses cross-functional teams aligned by business service. The teams are responsible for the entire stack that was delivered in shared services.
Let’s use infrastructure, apps and testing as an example of the difference in a digital stack. In DevOps, we move those workloads out of the legacy environment and functional structure into a software-defined environment. In a software-defined environment, we effectively collapse the stack. The cross-functional team has responsibility for both development and operations. Rather than aligning the work by functional expertise, it’s aligned by business services.
DevOps is disruptive
This digital construct around DevOps threatens the shared services model and, by extension, the labor arbitrage model. How?
First, the DevOps structure is more efficient than labor arbitrage. When we move the stack of functional expertise into a software-defined environment, we need fewer people to do the work. Thus, the labor arbitrage model becomes less important.
Furthermore cross-functional teams deployed in DevOps naturally want to be close to the business because they’re aligned by business services. The teams need tight communication. Therefore, the DevOps model naturally wants to come on shore. Business users naturally want these teams located in the same time zone and often the same building or in close proximity to the business, making the labor arbitrage equation less amenable. DevOps teams are tightly tied to business impact.
Microsoft and other companies have DevOps teams located offshore. However, in these instances, they move the entire cross-functional team offshore rather than having a distributed model where some components are onshore and some are offshore.
The idea of benefiting from a “follow the sun” strategy in labor arbitrage to increase productivity doesn’t seem to hold water in DevOps. In DevOps, constant and close communication dramatically add to productivity.
Finally, the objective function in a shared service organization is low unit costs per function. In contrast, the objective function in DevOps is high business impact at speed. Consequently, as companies shift into true DevOps models, they use dramatically less resources in a labor arbitrage model.