Today is the penultimate Tim Woods day, along with the most obvious. Defects. There is a something wrong with the part, data, or finished product. Which means it has to go back for rework or be replaced. Which is…waste.
Defects can be created for any reason from accidents to malfuctioning equipment to loose tolerances to simple human error. Rather than go into all of the possible causes, let’s look at the total impact and the ways to address the causes of defects.
First, a defective part, data, software code, or anything else costs more than just the time it takes to do the process over again. A good rule of thumb that I learned in my software days is that if you find a defect, the cost is 1 (time unit, dollar, whatever). If someone else internal to the company finds the defect, the cost is 10. If a customer finds the defect, the cost is 100. Each step away from the step where the defect was caused creates an order of magnitude more work to fix it.
Think about any example from work. From banking, if someone submits a form that has incorrect information, when it is finally caught internally, there are emails and phone calls that go back and forth, otherwise smooth workflows that get interrupted because this is usually a Push process, and time spent by multiple people to correct the problem. If the customer finds the error, the same work has to be done and now the entire organization looks bad, someone has to apologize to the customer, and depending on the severity, other steps may need to be taken. That is no fun for anyone involved.
The same cost logic can be applied to blood draws at hospitals, auto parts, school bus drivers forgetting to pick up kids, a botched podcast and a thousand other examples.
Once you find a defect you need to
- find out why it was caused. Was it a training issue, mechanical issue, overly complex requirements?
- determine how many of these defects are being caused. Was it just the one time or is this a pattern?
- determine the impact. If you have multiple defects to follow up on or some other type of waste to address, which is the most urgent to fix?
- apply an appropriate remedy. This will be case specific, possibly requiring a change to the entire process or just one step.
- figure out how to apply and track accuracy measurements. This can be tricky sometimes, because requiring too much time at process steps for this can lead to over processing.
Defects, no one likes them. From typos to dam failures, they range from annoying to catastrophic. But with a little foresight and examination of your processes, you can tame them.