Stats 101 - A sample of the total customers is sufficient to calculate the monthly average. As long as the sample size is sufficient (capital one credit cards) and the sampling method is not too biased (or in this case stedaily biased month-to-month). So yes this information appears highly valuable for forecasts as it gives customer amount and sales amount ahead of reporting.
sure, i'm not debating that this isn't stats, but it's a skewed subset, skewed towards those who can afford to have a credit card. then, stats is far away from knowing confidential information. I don't know...
At the end of the day I am not so sure the value of the information matters, even though in their case it turns out the data subset accurately reflected the larger picture (e.g. Capital One data was consistent with not just other credit card transactions, but cash as well).
Nevertheless, one may still be committing insider trading by trading based on non-public information even if the bet was wrong. Though is such a circumstance they may not have come up on the radar.
