The Cost of Delay

I recently posted a video explaining the cost of delay

This is such an important topic. We all know the cost of delays when it comes to web requests — if a site takes too long to load, you’re going to lose customers. So we work hard to analyze what’s contributing to the delay, and apply best practices to monitor, alert on, and improve our performance and reduce delays.

But for some reason beyond a general frustration, we don’t really put much effort or investment into understanding and improving delays in software delivery.

I think this is for two reasons. First, the causes of these delays are subtle and invisible. As I say in my video, it’s not like an assembly line where you can physically see where things are getting jammed up.

But I think it’s also because we don’t really see how expensive these delays are. The costs are real and significant, and can even be an existential threat to the business

Here are some of the ways in which delays can cost you:

• The value you would get from releasing a feature earlier is lost
• The value of a feature degrades the longer it takes to get a feature get out
• The longer it takes to get a feature out the door, the longer it takes to get feedback, and this has a cost because the resulting adjustments are larger and more expensive
• Risk is increased the longer it takes to address the risk. As just one example, high performing companies were able to very quickly address the log4j zero day issue. Companies with slower release cycles took significantly longer, and were under much higher risk due to these delays.
• You can easily lose customers to competitors who are more responsive and agile and who are able to deliver more features in a given time period.

In subsequent videos and articles I’ll share principles, tools and practices that that I, and many others, have found to be very effective in both seeing and reducing delays in software delivery.

With these in your toolbox, it is my hope that you can optimize the performance of your software development processes with the same attentiveness, rigor and effectiveness that you use to improve the performance of your actual software.