Quote:
Originally Posted by Hrafn
I wasn't in fact disparaging Amazon's data but the publishers'.
Amazon has access to the data (at least on their own customers) on a completely disaggregated customer-by-customer basis, as well as being in a position to experiment on their customers by offer subsets different prices (as we know they've done in the past). This should put them in a position to build some very accurate and robust models. They would also be in a position to make a lot of money off this information, so I'd expect them to hire some very smart econometricians to design their models.
Thus my original point that I'd tend to trust Amazon's numbers over the publishers.
Addendum: it should also be noted that I was talking about level of aggregation in the input data that Amazon versus the publishers have available in creating their models. This is completely different from how aggregated versus detailed the reported results are.
|
My apologies. I completely misinterpreted your post. My only excuse is it has been a long day. Once again, my apologies.