If they profile all of the people in the store simultaneously, they can raise/lower prices based on time of day and item purchased. Then you won't see the prices changing constantly as people walk around the stores, they'll just change gradually throughout the day, when people aren't looking.

If all the people who work at Goldman Sachs next door to the store get off work at 6pm, the store can raise steak prices at 6pm while lowering the price of Oreos (or whatever). It doesn't even have to be based on economic class statistics, it can be tailored to the specific times when individual people shop at the store and which items they buy.