Since the “Scrum” methodology became popular trend in the SW organizations, the problem of performance evaluations has been in the discussions. The basic reason for this is that most of the big companies still use conservative HR processes, where people have to be evaluated in order to reward/punish them and thus ensure that they are really performing.
The basic problem that Scrum raises up is: Should the company measure only the performance of the “Team”, or should it measure the performance of the “Individuals”? In Scrum philosophy the smallest basic building block of organization is “Team” and so it would be logical that only the performance of the Team is evaluated, not the individuals. In the systems thinking it is the cooperation of the agents that creates the ideal performance – optimizing the performance of the individuals could result the “system as a whole” performing worse (i.e. suboptimization). However, this is conflicting with the traditional HR systems – how could you know who are the individuals responsible for the good/bad performance of the teams if you don’t evaluate them? :)
One solution proposal for this are the peer evaluation systems. The idea seems to be that if the team members do the evaluation themselves, then it is actually part of the self-organizing phenomena of the “team as a system”. This supposedly leads to less danger for the hurtful suboptimization.
In this post I’m not so much interested to compare what kind of performance evaluation systems are good and what are bad. Instead I thought it might be interesting to see what kind of Ways of Thinking are leading to these different performance evaluation solutions.
Three ways of thinking about organizations
I have seen there to be three quite easily differentiated ways to look at the organizations. In reality there are also others, but these are the ones that can be quite easily recognized. If you are interested to learn more in-depth about these, you can read my essay from the recently published book LESS! Essays on Business Transformation (free eBook).
To make it simple, I call these viewpoints Mechanistic, Systemic and Complex. Most people intuitively use all of these viewpoints, but probably are drawn to utilize some of them more than others. My experience is that the mechanistic viewpoint is used the most, while the systems thinking is gaining more popularity. The complex viewpoint is still quite unknown to wide audience – although in practice the complex patterns of behavior are evident (and inevitable).
Mechanistic view looks at the organizations to be like a machine, a clockwork. The people with their roles and responsibilities are like cogwheels and the processes are the common design of the clock. The manager is like the clockmaker who can control the efficiency of the machine by supervising and giving orders to each of the cogwheels.
It is this thinking that is evident in the traditional performance evaluation systems. The efficiency is simply the result of the individuals performing their job. If everybody is fulfilling their roles and responsibilities properly, the clockwork is functioning well. If there are problems in the performance, each of the cogwheels can be evaluated by themselves and corrected/switched to a better one. This is done until the clockwork is functioning well again. If this isn’t enough, the overall design can be improved. In reductionist manner, the same principle can be applied to the managers also. Each of them are then responsible of the design/functioning of their part of the clockwork.
Systemic view looks at the organization from a different angle. Here the interactions between the parts of the system become more important than the parts themselves. The “teams” become more important than the individuals. Instead of controlling the individuals, the managers design boundaries for the teams. The roles and responsibilities become boundaries also – they are somewhat flexible enabling the people to adapt to the situation, but they also guide the behavior in certain limits. Managers can still control the organization with the same tools (roles, responsibilities, processes), but they become boundaries for the parts of the system and not direct control mechanisms for individuals.
The analogy used here is an organism. It is impossible to build an organism in a mechanistic manner – you can’t first draw a design for the organism and then take different kinds of cells to build it from the scratch (like Frankenstein monster). You have to grow the organism – take a seed, put it in the soil, give it water and expose it to sunlight. A beautiful bonsai tree starts to grow according to it’s inner design, but the gardener can prune it, steer the dimensions of the branches with robes, allow the sunshine only from certain angles, control the humidity of the air and soil, etc. Again the manager is in the position of designer, but instead of an architect he is now a gardener.
Scrum is a systemic methodology. Thus it isn’t very surprising that it develops inner conflict with mechanistic HR processes. Peer evaluation systems try to solve this conflict with “both-and” logic – teams are both organisms and individuals. Individuals are important ingredients of the team and thus it is possible to evaluate them. But instead of clockmaker evaluating them like cogwheels, the peers can be seen like parts of the organism who are guiding each others so that the bonsai is fulfilling it’s inner design. The manager can then observe the growth of the team and steer the development with his “gardener’s toolbox” to suit better with the goals of the bigger system. In isomorphic manner, this same philosophy can be then applied to managers themselves. The gardeners themselves can also be grown and gardened.
It is also worth noting that it isn’t uncommon to see one manager utilizing mechanistic thinking and his own manager utilizing systemic thinking – or vice versa.
Complex view takes yet another angle. It challenges both the notion of autonomous individuals and the view of organization as an organism. Complexity views the organization as a dynamic process of cooperation between interdependent agents. There aren’t really any fixed entities called “team” or “organization”, instead what we see are patterns emerging from this cooperative communication. “Organization”, “team”, “roles”, “responsibilities” and “processes” are conversational patterns used as tools in the daily politics of organization. People use these abstractions in their conversations as gestures that call forth responses. However, nobody can control all the responses he might get nor all the other conversation that are happening locally between different people. Because of this, the processes of organizing and product-making are always unpredictable even-though they are also somewhat predictable at the same time.
If one looks at the different performance evaluation systems from this perspective, it all looks quite different. The whole notion of “performance evaluation” becomes a conversational tool, a social object established in the conversational patterns of the organization. It is a tool in the organizational power games. Managers use it to establish hierarchical power relations, but also the subordinates use it at the same time to different purposes (e.g. legitimizing their own actions, gaining recognition, etc). Power relations are always complex and dynamic – they don’t follow the official hierarchical structures. Thus the subordinates might easily have power over their managers too.
When looking at the different performance evaluation systems from this point of view, it doesn’t really matter what kind of systems are used. What matters is how the people use those abstract concepts to build their own identities in the organization. Skilled individuals can use whatever system to initiate meaningful dialogues around the process of product-making. The results of these dialogues to the business can’t be predicted beforehand. They might initiate beneficial or negative patterns alike.