Perhaps, but all that information can be obtained from census data at a much lower cost. Demographic and income data can easily forecast what type of people will be visiting the store and what they will be buying. So I do not see much advantage from collecting essentially similar kind of information through loyalty cards at a much higher cost. ....................
The cost is not higher from the point of view of large enterprises such as super market chains, or mega-scale retailers such as Target, Walmart and Best Buy.
Census data serves an entirely different purpose. Using census data, how can I (the chain owner) know how many people are buying Grandpa Beryllium's Retro-styled Grape-like Liquid Composite?
Answer: I can't.
Also, census data is slow moving - because of it's different mission, it's not gathered daily and cannot be used to capture daily and weekly trends. Predictions based upon census data would be much less accurate than predictions based upon sales data which is tied - almost in real time - to demographics.
Show of hands between thee and me - between the two of us, who has actually built one of these systems, watched it work, tracked its progress and thought through the implications?
...
There are many wasteful and foolish things in the world but loyalty cards - which provide sufficient ROI to justify their (not terribly large, all things considered) cost to a large scale business - are not one of them.
This isn't to say they're marvelous and we should celebrate ever more peering into our habits but, the cards are an emergent property of databases - i.e., almost inevitable under current conditions - and should be accurately understood.
.d.