Credit Cards: Why Is My Available Balance Zero?

Photo by CardMapr.nl

I recently make a large credit card purchase on a Thursday.  When I checked the credit card company website on Saturday, the transaction had been posted that Friday and was no longer pending.  However, even though my current balance was almost $1000 below my limit, an online transaction on Monday was rejected.  Checking the website again, I noticed that my available credit was zero, when it should have been the difference between my limit less the current balance and any pending transactions.

When I called the credit card support line on Monday, I was told that the Thursday transaction was still showing as 'pending' on their system but should clear that afternoon.  It appears humans are involved in clearing some pending transactions and they do not work weekends.  When I still had no available credit late on Monday, a supervisor said she had seen these 'double pending' transactions before - at best, it would clear overnight but I might have to wait until Wednesday morning.  Fortunately, my available credit was restored by Tuesday morning.   

I periodically see credit card transaction showing as posted and also as 'pending', but normally for only a day.  Why this happens and why it took five days in this case remains unclear.  No-one could explain why I did not get an alert that my available credit had dropped below $100, the default level for that credit card.  I suspect alerts are only triggered when a transaction causes my available credit to drop below the notification limit, and the 'double pending' condition does not count.  I escalated both issues.  

There is a tendency to divide the world into batch vs. real-time, such as https://www.wired.com/story/why-tech-bros-and-politicians-cant-really-connect/, but real world systems tend to be a complex hybrid of both, with a dash of humans thrown in for good measure.  These systems are often poorly understood, even by support staff.  As Don Norman points out in The Design of Everyday Things, we end up creating conceptual models that do not reflect how the systems actually work ("The System Image", p44).