The argument often heard now is that improvements have occurred in the past few years.
Well, I decided to check it. So I took a large newsagent file and ran in our newsagency software a report and then exported it to excel. Here is the chart.
As you can see, the minor titles the magazines are clearly not right unless you are talking about the very large ones circled in red.
This is what a statistician would expect. Say you sell one a week, if this week that person does not show because he missed his train they would be 100% off. Alternatively if you sold a hundred of that magazine, they would be 1% off because of that train. The more you sell, the more accurately they can predict the expected sales.
Overall much work needs doing on the low selling items.