What is Flow? Thinking Differently About Work

Example of some impediments to flow

One of the concepts to recently come to the forefront of the agile movement is the concept of flow. The traditional metaphor of project and product delivery was one of engineering and construction. The idea that we are completing the work, and once we have completed all the work (usually expressed through a work breakdown structure) the job is done, as one might methodically complete a lego set.

While in an of itself Flow is not a new idea, it is a powerful one, and has been adopted by many frameworks and methods recently as an underlying principle.

Flow as an idea has three root origins. Firstly, Dr Goldratt’s Theory of Constraints, which famously visualised work in process as a boy scout troop with one slow hiker. Secondly, Lean and the Toyota method, with its concept of kanban, pull systems and value stream mapping. And finally Don Reinertsen’s work on Product Delivery, looking at the way the structure, size and form of work impacts effectiveness of delivery.

The most recent and impactful articulation of these ideas is in Mik Kersten’s book ‘Project to Product’. He articulates a simple but effective model for improving the way we work.

So what is flow? The simplest way is that it is a powerful metaphor, where we imagine work transitioning through various statuses from ‘not started’ to ‘done’. This flow might look like objects flowing through a pipe, or traffic travelling along a road. Sometimes it can be more complicated - such as people milling about and queuing to have a fun day at an amusement park. For the purpose of this essay I’ll use the metaphor of traffic as a relatively simple and relatable image.

In any system there is a constraint. We know this because the system does not product infinite work. We need to know where this constraint is so we can remove it. This widens the road. Now there will be a new constraint, and we move on to that. Management’s job is never done.

If we widen a part of a road that is not the constraint, not only will it not only improve overall traffic flow, it may well make things worse. Traffic jams are examples of this.

If traffic moves too slow, we will experience delays. If traffic moves too fast we will experience delays. If every car is late and over budget, the solution is not to train drivers better, or punish or reward them for performance. The answer is to lift our thinking and realise there is a systemic problem, and solve that.

Optimising one section of the road might not optimise the whole journey.

By conceptualising work in this way we can start to solve some of the systemic problems in our organisations that slow us down.

When we do not effectively manage flow we usually see the following problems;

1. Everybody is very busy but nothing seems to get completed

2. A large amount of work is going on, requiring a large team to track and monitor

3. Delays are common, often the norm

4. New work entering the pipe has a massive amplified disruptive effect on existing work

5. Emergency high priority work has difficulty being executed without huge damage to in-process work (visualise a combination ambulance/tank smashing through a traffic jam)

6. Stress and disengagement as delivery management are unable to meet their outcomes

In order to manage flow, Mr Kersten advocates for a set of flow metrics;

1. Flow Distribution. The different types of work underway, usually expressed as percentages

2. Flow Velocity. Also known as throughput. The number or volume of work items completed in a given period.

3. Flow time. The time to complete a piece of work, usually expressed as a distribution chart.

4. Flow load. The number of work items waiting and in progress at any point in time.

5. Flow efficiency. The percentage of time work is actually worked on within its total time to complete.

6. Flow predictability. How well the system predicts flow.

By leveraging these metrics, an organisation can optimise their portfolio of work from a flow perspective, and complete much more in the same time with the same people and tools. Increases of 30-50% are not uncommon.

How is this achieved?

1. Begin by tracking metrics and making the work visible, using the flow metrics and a kanban system.

2. Reduce the work-in-process and confirm this is also reducing flow time.

3. Use the free capacity from lower levels of utilisation to manage unexpected and emergency work so it does not destroy work elsewhere

4. Undertake value stream mapping to understand where there are delays and quality issues to increase flow efficiency, which will further improve Flow time and Flow Velocity. rinse and repeat.

5. Check your plans against the observed work to measure and improve your forecasting.

With any system, if it is not operating predictably, we cannot understand, control, and improve it.

The pyramid for high performance in technology is;

1. Transparency and visibility

2. Predictability

3. High performance

Without first traversing 1, and 2, no organisation can hope to reach 3.

If these problems resonate with you and you’d like to learn more, please reach out to russell@flowspring.nz to talk.