Dennis Stevens - Theory of Constraints and Big Agile
In 1984, Eli Goldratt first published The Goal: A process of ongoing improvement. The book describes the Theory of Constraints, a method for managing systems. It is based on the concept that at any time in a system there is one (of very few) bottleneck(s) slowing down the performance of the system. Performance is measured based on three variables. Throughput measures the units delivered to the consumer of the system. Operating expense is investment into the system to ensure its operation on an ongoing basis. Inventory is investment into the system to produce.
The Goal is to make more money now and in the future in order to meet the expectation of stakeholders and ensure the business continues to operate. You increase profit by combinations of increasing throughput while decreasing the ratios of operating expense and inventory against throughput. A key first step to accomplish this is to reduce inventory – or work in process. This will improve cash flow of the organization. Reducing WIP has the added benefit of exposing the bottlenecks in the system. Exposing these bottlenecks is the key to implementing POOGI (Process Of OnGoing Improvement). Goldratt provides five focusing steps to follow to help achieve the goal.
Five Focusing Steps
Identify the limited number of current constraints: At any point, there is a single point in the system that is the current bottleneck. For example in Agile development teams, it may developer productivity or a testing person. The way that you will identify the constraint is that it is the person where work is piling up in front of them and where downstream resources are starved for work. In the figure below the constraint is B. Work can only get through B at 3 units per unit of time. Since A is more productive work will pile up in front of B and C will be starved for work.
Two important concepts need to be presented here. The first is that the team of ABC can’t produce work any faster than the bottleneck, B. So the consumer is only receiving 3 per unit of time. The second really important concept is that A, B, and C don’t have to be different people, they can be state transitions of a piece of work. In an Agile team, A might be understanding the requirements, B might be develop and unit test the code, and C might be integrate and acceptance test. On an Agile team, everyone is involved in all the states, it is the work that is moving through various states.
Make sure output from the constraint is not compromised. The second step in the five focusing steps is to recognize that the work done at the bottleneck is precious and must be exploited. So focus on making sure that none of B’s work is wasted. In our Agile example that means improving the quality of how work is communicated in the transition from A to B. It also means focusing on quality at B so that C is able to use everything produced by B. Doing this step will result in improved performance of the system.
Subordinate the system to the bottleneck. This means slow down the work at A. While this might seem counter-intuitive B can’t work any faster than it can work. Often, manager’s focus on improving utilization rather than throughput and this focus actually exacerbates the problems. Producing more and more inputs to B creates stress on the people in the system, makes the system more difficult to manage, and increases the likelihood of waste at B. In our Agile example, A can spend more time clearly communicating what needs to be built and less time producing more detailed stories. A may consist of meetings where stories are communicated, estimated, and prioritized. On our Agile team, it is consuming time from the team that could be spent at A. So do less detailed estimating and prioritizing in each cycle. The effort is better spent at B.
Elevate the constraint. In many organizations, when the effort is to elevate performance of a system (or team) investment is made to elevate the entire team. In our example above, that may mean adding more subject matter expert resources as well as more testing resources. But the only benefit to improving throughput of the team in our example is to increase at B. Elevating the constraint can happen through improving the capabilities at B or shifting more manpower to B. In our Agile example above, it may be better to shift one of the testing people to focus on development. Even if they are initially not as productive as the top developers, they will contribute to elevating the constraint without damaging the throughput of the team.
After the system has stabilized go back to the beginning and identify the constraint. The final step is really to not let inertia become the constraint. Remember, this is POOGI – a Process Of OnGoing Improvement. You don’t go through the process once – you go through it continually.
Theory of Constraints and Big Agile
This model is very interesting because if provides a thinking process for focusing efforts on the next most important problem. If our Agile team example above, if this model is not kept in mind by the team, they may spend time putting in a cool new CI environment – but unless that environment raises the constraint at B it will not result in an improvement. So it isn’t the next best place to focus. The other interesting thing about this model is that it scales up through the organization through the orders of scaling. If you haven’t read The Goal and Goldratt’s other writings you need to get this model into your thinking toolkit.