Skip to content

Why Do Systems (Tend To) Fail on Monday

You wake up an 06:30, after a full relaxing (but somehow late ending) Sunday. You rush into your car and start commuting. The traffic lights system is annoyingly slow, delaying your usually course to the office. Some traffic lights are broken. No way you could shop for a drink at your favorite coffee shop – the queues are way too long this early on the week start. You get to the office – then you circle around 10 more minutes as usually to find a parking place.
When you get (finally) in front of your desk, the usual ERP interface starts 10 seconds later. The OS seems also very slow – lots of software updates pop up and urge you to restart your computer, and the office WiFi connection is painfully slow. Everybody’s running around in meetings and even the urgent “to do’s” are postponed. And the list can go on for pages.
What really happens on Monday mornings with our emails, sales organisations, computers and organizations? Why is everything so slow?
The answer came to me several weeks ago, when I was reading a piece of news regarding Netflix: it said that during prime time, when the highest-watched new episodes of their TV shows are aired, Netflix accounts for almost 40% of the bandwidth (or internet traffic) in the US. Which causes huge slowdowns in the internet traffic, to such an extent that some mobile operators even sued the company for over-using the internet infrastructure.
And this of course explains why Monday mornings are causing most of the work or human-used systems to slow down: the traffic is huge, be it the need to buy coffee, the urgency of having some products delivered or the high adrenaline of the organizations trying to get massive tasks done. Systems are usually designed to cope with normal loads (in terms of users, requests or processing other types of resources or information). They are NOT designed to cope with double the traffic during peaks.

Why not? You might ask. Well, to start with – this would cost a lot of money and the processing capacity would be idle for 80-90% of the usage time. Which is a waste, of course. Secondly, systems theory is less studied in computer sciences or operations management – and even less applied. As a new science, systems theory can actually uncover these kind of “bottlenecks”, as they are named in operations management.
But this is not all. Of course, one can ask – how can we solve this problem? How can we create systems that respond properly during peak times without spending a huge amount of resources on them?
Well, about this, maybe on a next blog post. Until then, I am looking forward to your comments and reactions!

Leave a Reply

Your email address will not be published. Required fields are marked *