Kafka Statue

Are you measuring a team effort?

Productivity is used to evaluate how efficiently an organization converts inputs into outputs.  However, productivity measures can and often are misapplied for a variety of reasons ranging from simple misunderstanding to gaming the system. Many misapplications of productivity measurement cause organizational behavior problems both from leaders and employees.  Five of the most common productivity-related behavioral problems are:

  1. Attributing team efforts to individuals.  Most software development and maintenance activities are team efforts. Productivity measures usually reflect the output of the teams doing the work, and can’t be ascribed to an individual.  However, there is often a chronic need to ascribe team output, whether good and bad, to an individual.  For example, often productivity data is used in an individual’s review and evaluation. This can lead to everyone trying to game the metric.  The solution for this issue is just not to fall prey to using a team metric to measure an individual.     
  2. Measuring productivity can reduce productivity.  Measurement of productivity requires time and effort to count the inputs and outputs of a system. If an organization does not measure productivity they could either create more output with the effort they would have used to measure productivity or reduce payroll, in either case leading to an increasing productivity.  If the measurement of productivity is not directly tied to process improvement or cost avoidance it will have no return on investment. Avoid overhead that does not deliver significant ROI. The solution is to use productivity measures as part of process improvement or cost reduction program.
  3. Gold plating an output to increase productivity.  One of the most famous adages in management circles is that “you get what you measure.”  Holding teams or individuals accountable for productivity above all else will lead to pressure to add features even if not asked for in order to increase the development output.  This is often referred to as gold plating. Gold plating leads to delivering more functionality than needed often at the expense of needs that are higher value.  One fix I recommend is measuring productivity, value and customer satisfaction to ensure teams have more than a one-dimensional view of their performance. 
  4. Optimizing local processes without regard to the overall system. The focus of any system is to deliver the most value possible and to improve over time.  Maximizing the efficiency of a part of a system may not translate to the whole system becoming more efficient. For example, I recently observed a large organization with multiple software teams that efficiently developed software.  After developing and testing the software, the operations/implementation group performed a review and security testing prior to putting the implementation into production.  The operations team had a six-month backlog.  While waiting for review, two changes in the last month had to be withdrawn because the market need had changed before they ever made it to production. At the same time, a process improvement team was looking at ways to increase development productivity. Making the software development process any more efficient would only exacerbate the implementation group bottleneck.  In the end, the organization reallocated three development teams to support the implementation function in working through the backlog and streamlining the process.  As we identified when in our Re-read Saturday of The Goal, systems only increase output if you improve the throughput of the bottlenecks in the system.
  5. What doesn’t get measured gets overused?  Some organizations that focus on measuring productivity get very good at finding and utilizing resources that don’t influence the input side of the equation.  A classic issue is under-reporting of effort in time accounting.  For example, many organizations cap salaried employee’s time entry at 40 hours per week, even though they are often working 60 hours or more. The overtime is effectively hidden and because it considered free. Therefore, there is often little pressure to measure the overtime (until people start quitting).  Any resource that is considered free will be overused.

The last two behavioral issues are often the most common and can occur even when organizations don’t explicitly measure productivity. Every organization, whether they explicitly measure productivity or not, wants software development to deliver more functionality and cost less.  Organizations that don’t take a systems thinking view  can actually increase cost and reduce real productivity hurting the long term efficiency of the organization when they are trying to have the opposite impact.