@Anonymous wrote:Two things, one is whether the algorithm is set to prefer certain spendings like airline ticket/ hotel. Another thing is whether the number of transactions matter, say one hundred transactions adding up to $2K V.S. one transaction of $2K in the same store.
While this is pure speculation on my part, I'd imagine both would matter and be tracked. How important each are would depend on the issuer and their respective business models.
For a spend-centric model, volume, frequency and quality of spending would matter. Since a company such as Amex derives most of it's revenue from spending, I'd imagine one who spends $20K at a restaurant would be much more meaningful than the same amount spent at Costco. The former has on average the higest merchant fees, while the latter the lowest. Obviously, in this case, if we assume CLs are a form of reward for the customers, I'd imagine the higher profit one would matter more.
For issuers whose revenue is primarily based on outstanding balance revenue, I'd imagine the volume would matter more so than either frequency or type of transaction. Relative to interest on balance, merchant fees play a lesser role, therefore these issuers would be indifferent whether or not the aforementioned $20K were charged at a restaurant or Costco (assuming they accepted Visa/MC). Again, if we were to assume a CLI as positive reinforcement, these issuers would be more inclined to raise the CLs of those carrying larger balances.
Finally, both appreciate frequency of transactions, to some extent. With many merchants, along with a percentage fee charged by the issuer, there is also a fixed per transaction fee. If this fixed fee were 0.20 cents, a $1.00 charge would result in a 20% roe for the issuers. Make 100 of these charges, and they all add up pretty quickly.
