So, a typical thing that marketers would do is go back and say, look, these individual store effects are just really complicated. It's really pressing my ability just to price at the chain level, so why should I go out and think about what these differneces are.
So, what I do want to do is just say forget about these differneces, let's just have the same model for every store. Now, obviously from a Bayesian perspective, this isn't going to be a good idea, because I know something about why these stores are different. I've got this demographic information, why can't I use it?
Well, what a marketer might do is they might say, well, let's go to some type of cluster models. So I know about the city stores are behaving something similarly, I know that suburban stores are behaving, so let's just split these models and say that all the city stores are similar all the suburban stores - well, again in this case I'm losing a lot of the information. It's going to be not an efficient way to do this.
Another possibility, is to say, well, look i would say that there's going to be some kind of fixed relationship between the demographics and price sensitivity. Well, the point is that they're all xx errors in model specification and errors in data measurement and what model, what variable should we include and not include. So the point is that you can't do these fixed effects, and if you do specify these fixed effects it's going to be sensitive to your specification.