One thing about patterns is that most of us look for permanent patters. Ones that are visible all the time. We can get pretty sophisticated at that and can pinpoint patters of change in processes as well not just say in output.
What is a lot harder to figure out are temporary patterns. Regularly recurring ones are the easier kind. Irregular ones are tough. Because it is based on a gut feeling usually. You suspect there is one that happens but you do not have solid data to pinpoint it.
For situations like this, here’s my two cent: if you were right, what would be the impact on the bottomline? If there are many enough zeros, it’s worth investigating.