What is system delay?

A system delay is the time that passes between a cause and its consequence. Hiring takes months to pay off, a price change takes weeks to reach demand, and a new factory takes years to build. Delays matter because decision makers keep acting on a signal that has not yet responded, and so push too far.

How it works

Delays sit inside feedback loops. When a balancing loop has a long delay, the correction keeps arriving after the problem has already been fixed, so the system overshoots the goal and then swings back the other way. In a reinforcing loop, delays can hide early growth or decline until it is hard to stop. Information delays, such as late reports, are as damaging as physical delays, such as shipping times. A useful habit is to estimate the delay before judging whether an action worked.

Because delays are easy to forget, they are among the most common reasons that careful plans fail. Making a delay explicit in a model, with a number on it, is often the single most useful step toward better decisions.

Example

Say a call center manager sees a backlog of 200 calls and orders 10 more agents, but training takes 6 weeks. In the meantime the backlog keeps growing, so she orders 10 more. When both groups finish training, there are 20 new agents for a backlog that 10 would have cleared. The center now has too much staff, and morale and costs suffer. The problem came from the 6-week delay, not from poor judgment about the backlog. This is an illustration with round numbers.

How to model it in Machinations

Place a Delay or Queue between the action and the node it affects, and set the number of steps it holds resources. For lead times on physical or staffing flows this is the main tool. Play the diagram with a Pool chart, then change the delay length with a Lever and watch the oscillations grow or fade.