I don't the beta(1,1) is a good prior. There is not an equal probability between 0 and 100 that a store has that percentage of good reviews. I think the the art

I don't the beta(1,1) is a good prior. There is not an equal probability between 0 and 100 that a store has that percentage of good reviews. I think the the article should recompute the beta parameters using mle over all amazon stores.