Joshua Arnold facilitated a workshop about Cost of Delay at the Lean Kanban France 2014 conference. InfoQ did an interview with Arnold in which he talks about the cost of delay: what it is, why it matters, the importance of quantifying it and some tips for getting started.
InfoQ: For people who are unfamiliar with Cost of Delay, can you explain what it is?
Arnold: Cost of Delay is a way to communicate the impact of time on the outcomes we hope to achieve. One way to think about it is to consider the project, product or feature you are working on and ask this question:
What would it be worth to us if we had it right now?
In situations where there is already a nominal delivery date, you might ask:
What would it be worth to us if we could get this a month earlier?
What would it cost us if it was a month late?
The answers to these questions tell us more than just the value, they tell us how urgent it is. Cost of Delay is a way of expressing both value and urgency.
InfoQ: Why does Cost of Delay matter?
Arnold: When developing new or improved products and services, understanding the Cost of Delay helps in three key ways:
Firstly, it helps us make better decisions. By making the economic trade-offs more visible and easily understood, we can make faster and more sound choices. Making tough choices where you're trying to balance lots of different variables is an inherent part of the creative process. Many of these choices affect how long something might take and how much it costs. There are also lots of trade-offs at the system level, where understanding the Cost of Delay is a key part of managing the flow of work. Whether it is experimenting with WIP limits, controlling the length of queues, optimising batch sizes at various points or trying to work out appropriate levels of capacity utilisation, Cost of Delay is a key ingredient in all of these decisions.
Secondly, Cost of Delay helps us to prioritise in a way that maximises value. Using CD3 (Cost of Delay Divided by Duration) we deliver more total value with the scarce capacity we have. Unlike in manufacturing, value and urgency is not homogenous and evenly distributed. Employing CD3 not only helps to surface our assumptions about value (so we can design better experiments to test them), it also encourages the breaking down of work. If you want to move on from prioritisation by politics or gut-feel, maybe you should try using Cost of Delay.
The third way that Cost of Delay helps is in changing the focus – from efficiency and cost (which encourages the wrong behaviours), to speed and value. It's no good asking people to not do something (like estimating cost or delivery dates) if you don't give them a viable alternative that actually helps them. By talking in terms of Cost of Delay, you get more of what you want and less of what you don't.
InfoQ: What makes it important to quantify the cost of delay?
Arnold: Unfortunately, we don’t have very good intuition about Cost of Delay. Because we are unfamiliar with it, we haven’t developed a sense of the patterns and parameters involved. As a result, we can’t quickly categorise using System 1 thinking – our gut-feel is typically wrong. We need some help from System 2 thinking – not much, just a little.
To illustrate: for many years Don Reinertsen has been asking people working in product development: “Is a month of delay worth a million dollars or a thousand dollars?” Sadly, 85% of people working in product development don't know. My own experience of using this at various organisations reinforce his findings.
When lots of people in the process are making different assumptions about the value of time, this of course leads to some rather strange decisions. One person will be spending lots of money to save a little time while another person is spending lots of time to save a little money.
At the organisational level, I talk to lots of senior managers who are overly focused efficiency, but in a way that is actually costing them dearly in delay costs. It’s not until you actually quantify the Cost of Delay that this becomes obvious. No amount of categorisation or using relative Cost of Delay helps with this. Unless you quantify the Cost of Delay, the price we are paying for queues remains invisible.
As Don Reinertsen puts it: “If you only quantify one thing, quantify the Cost of Delay”.
InfoQ: In the workshop you said that some people might resist giving estimates of value. What do you mean by this?
Arnold: In my experience, people with an I.T. background are usually more reticent about the idea of quantifying Cost of Delay. My theory is that these people have learned to be wary of requests to estimate anything as a result of having had estimates of time or cost abused. The natural reaction to this is to avoid any situation that feels remotely similar.
What is different about estimating value (as opposed to cost) is that we are just as likely to underestimate the Cost of Delay as we are to overestimate it. The other thing to consider is that these decisions will be made either way – it’s simply a question of whether we choose to trust the gut-feel of whoever is making assumptions about value, or whether we try to improve it. It doesn't need to be perfect, only better than the incumbent – our intuition – which we know to be quite poor.
InfoQ: How can people get started with Cost of Delay?
Arnold: We’ve found that it’s better to start with one team or individual who is suffering the most from either difficult prioritisation decisions, or poor trade-off decisions. The more complex the stakeholder map, the more likely they will benefit from doing this. You only need to find one brave soul who is prepared to give it a go.
It also helps to collaborate with someone from Finance who can help you sense check some of your assumptions. There are generally only a handful of patterns in most organisations, with a set of base assumptions that you can then reuse and refine these over time.
Teaching others the ropes is much easier when you’ve got some concrete examples. It’s critical that assumptions are shared and made visible to prevent gaming. This also helps speed up the learning process.
Before long, others will probably discover that understanding the Cost of Delay is really useful. Because it helps people with various difficult decisions it tends to start to spread organically.
InfoQ: If people want to learn more about Cost of Delay, how can they do that?
Arnold: Don Reinertsen first wrote about Cost of Delay way back in the 80’s. He has since written additional material and it is rare that a talk of his doesn’t mention Cost of Delay in some way: choose your favourite search engine and go hunting.
For a quick intro, there is a 3-minute video introducing Cost of Delay that you can find at costofdelay.com. We have also written up our experience of introducing it and using it across the portfolio at Maersk Line.
