the cost of extrinsic motivators
The best output comes from engineers who are intrinsically motivated, according to Dan Pink in his book “Drive”. He details strong arguments around how complex, creative, right brain work is hindered by extrinsic motivations (e.g. pay for performance schemes) and is bolstered by intrinsic motivations (e.g. the internal desire to achieve, the feeling of accomplishment).
As an industry we’ve already identified that many of the pay for performance style schemes and metrics are detrimental to the creation of stable software. Measuring software engineers on lines of code or tickets delivered is commonly accepted as flawed when the desired outcome is a correct and bug free product. Measuring performance on lines of code, both max and min, creates an environment where engineers start optimising for the metric, which is detrimental to the quality of the codebase.
As Dan Pink describes; all extrinsic motivators, and particularly those that are in for format of goals, narrow focus. This can be a good thing and can help ensure that as a team we are building the right product in the right way. But it can also be a bad thing. Narrowing focus hinders right brain creativity and limits the scope of what we can achieve by limiting the creativity of how we can achieve. As the complexity of the problems we are solving increases the value of goals can decrease. The narrowed focus blinkers the creative thinking required for innovative solutions, but more dangerously it can appear that the focus has exacted a better result than if we did not have such stringent goals.
In the example metrics of developer output, we ruled out the metric of ’lines of code’ as flawed as it encouraged optimising in the short term for the metric at the expense of long term success. But the same is true of goals. Why now do we consider goals to deliver high performance but other heuristics to be flawed? Goals become increasingly dangerous when the short term payout is a big one, such as a feature release. This kind of short-term measurable output goal can restrict our view of success and can encourage unwanted behaviours. We’ve redefined ‘success’ to be delivery of the goal, and not considered the impact of the corners we are prepared to cut to achieve it.
This kind of situation, often referred to as ‘crunchtime’ can contribute to the undesired outcomes. In well functioning teams, introducing artificial goals that create high-stress short-term deadlines often brings the team members together. After the goal is achieved there’s a feeling of ‘we did it together, what a success’. But achievement of the goal is a flawed metric that does not actually represent long-term success for the team. Crunctime encourages optimising for the goal, which encourages tactical and not strategic choices to be made. It encourages undesirable behaviours that do not respect the long-term outputs of the team.
Goals and short-terms extrinsic motivators can present in many different ways. There are some obvious cases like artificial deadlines set by sales teams for closing deals that are dependent on certain features. Or the less obvious cases that often happen just as a feature is about to land; news gets out of the imminent release and everyone wants to get their hands on the new code. This puts pressure on the final closing stages of feature releases, the stages that often ensure that the new code lands and is received well. Another example is career progression checklists that reward certain behaviours and focus team members on achievement of the progression goals at the expense of the team.
Instead we should allow the time for teams to deliver their best work, and ensure that extrinsic motivations do not creep in. Engineers who are intrinsically motivated, where learning from the task is the reward, do not produce undesirable outputs or unethical behaviours as the cost of those are borne by the individual and not the team.