A software house had been receiving numerous complaints about faulty disks. They assumed that these were caused by faulty processes in the production plant in Ireland and sent complaints in their direction. Nothing changed.
One engineer decided to visit Ireland and study their process. He found exemplary work and no likely causes of the faults. On his return he decided to conduct a type and frequency analysis of problems (all previously labelled ‘faulty disks’). He found that there were in fact many different problems, none of which was actually faulty disks.
Engineers had simply got into the habit of using received wisdom (it was normal to blame production), by using the complaint category ‘faulty disk’.