Feedback Control Response Times
In a previous blog post, I discussed how organizations cannurture innovation through the development of a “Learning” organization. In that blog post, I also emphasized the need for lines of multi-level communication that will allow ideas to come up from the lower levels.
To be effective, organizations need to have fast bi-directional communication between all levels. This needs to be fast enough to ensure that the organization responds in a timely fashion but not so fast that insufficient time is available for accurate and relevant facts to be collected and communicated.
In a previous blog post, I described how engineers developed the concept of feedback control to help them manage complex systems in dynamically varying environments. I also showed how this knowledge can usefully be applied to the management of organizational structures.
This dichotomy between speed and accuracy is fundamental to all feedback control systems. In what follows, I’m going to be showing you how engineers resolved this type of problem and how the techniques they developed in the process can be applied to improve the way in which every organization is managed.
In the picture above, we can see typical slow and fast feedback control response graphs. Notice that the slow-response graph takes longer to reach the command level than does the fast-response graph. Notice also that the fast-response gives rise to an oscillation and that, in this example, the oscillation stops over time. Clearly, in an ideal situation, the system would respond rapidly and reach the command level without oscillating.
What these graphs reveal however is that engineers have two choices when it comes to the design of feedback control systems:
(1) Slow-response designs in which the system will take longer to catch up with changes in the command level but which avoid oscillations and,
(2) Fast-response designs in which the system will catch up with changes in the command level more quickly but which involve oscillations.
In feedback control parlance, Gain and Damping are the two main variables of a feedback control system whose values determine whether it will have the characteristics of a slow-response or fast-response system.
The Gain of a feedback control system is controlled by the amount of energy that is available to drive the system’s response. For example, in a thermostat-controlled heating system, the gain is determined primarily by the size of the furnace.
When the room thermostat is set to a higher command temperature, systems with a larger furnace will be able to increase the room’s temperature to the new level more quickly.
As is shown in the picture entitled, “Thermostatically Controlled Room”, at any time during this process, the actual heat output of the furnace will be controlled by the size of the difference between the required temperature and the current actual temperature of the room. Because of this, as the room temperature gets closer and closer to the required temperature, so the heat output of the furnace will gradually be reduced. Using this approach, it is possible to design a low-gain heating system that will respond quite quickly without overshooting the required temperature.
While such a design is acceptable for the temperature control of a room, there are many situations where a much faster response time will be necessary. For example, jet fighters rely on feedback control systems that change the jet’s direction of flight rapidly in response to commands given to it via a joystick controlled by the pilot. Clearly, under these circumstances, the response needs to be more rapid (but not so rapid that the ‘G’ forces will be too high for the pilot).
To ensure rapid response, Jet fighters have jet engines that are large relative to their body size. This enables the jet to deploy high levels of energy in response to joystick commands. As a result, the gain of the system is high. Jet fighters’ feedback control systems do therefore generate oscillations but designers use a combination of clever control system design and damping to keep these to a minimum.
I mentioned earlier that, in fast-response systems, oscillations reduce because of damping. Additionally, when I described the behavior of a slow-response thermostat-controlled system, I noted that the difference between the required temperature and the actual temperature is used to determine how much heat should be generated by the furnace. To create a damping effect, engineers typically use the actual rate of response of the system to create an additional stream of feedback.
So, for example, when our jet pilot moves the joystick, the difference between the required and actual directions of flight is at its maximum and the jet engines are generating high levels of power. At this moment, the rate of change of direction is also at its maximum as will be the feedback that creates the damping effect. Using this technique, engineers can create a fast-response feedback controlled system which will respond rapidly to the commands of the pilot while at the same time damping the resulting oscillations to such a degree that the pilot is virtually unaware of them.
Organizational Gain and Damping
In a similar manner to the way in which gain and damping can be deployed in feedback control to improve the way in which they control complex systems functioning in complex environments, so also similar factors influence the capacity of any organization to respond effectively to the challenges created by a rapidly changing business environment.
Within the lower-levels of an organization, the amount of energy that will be generated as it responds to command and control instructions from higher up will depend on the effectiveness of three processes:
(1) Target Setting
(2) Measuring Results
Much has been written about whether these processes make any real difference in practice. In a recent article in The Economist (18th of January 2014), research being undertaken by Nicholas Bloom of Stanford University and John Van Reenen of the London School of Economics was described. They studied the performance of over 10,000 companies operating in some 20 countries and focused on three commonly accepted management techniques, namely setting targets, rewarding performance, and measuring results.
They conclude that effective use of these techniques is directly linked to improved productivity, profitability, growth, and survival.
What we have been able to ascertain, through the deployment of Talent Chaser, is that the proper use of these techniques achieve these improvements primarily through improvements to command and control structures within the organization.
Through the deployment of Talent Chaser’s Performance Appraisal and Task Action Planning Module, we have been able to demonstrate that, in an organizational structure, the damping effect comes directly through feedback control. Properly constituted Performance Appraisal sessions encourage employees both to acknowledge problems and to come forward with ideas for their resolution. Additionally, such sessions inhibit the development of blame cultures.
The feedback generated through this process, acts to dampen the organization’s response to command and control instructions in a way that encourages healthy debate and interaction between all levels of the organization. Our findings confirm those of engineers working in the area of feedback control, namely that, without such control, organizations can become subject to instabilities very similar to those associated with engineering control systems without proper damping.
These findings suggest that such feedback is not optional but at the very heart of the way in which highly responsive and stable organizations can be developed.