Yesterday I walked into a little shop named “Cozy” which sells clothes, shoes, and jewelry. Most of the items are hand-made and their prices fully reflected that. In one section there stood a sign “on sale from 20% to 50%”. “Getting 50% off is a good deal,” I thought. So I carefully looked through the items in the section, finally getting to a lovely pink shirt. The price on the tag was 60 dollars but with a blue sticker beside it. I brought it to the lady at the counter and asked her what discount I could have.
To my surprise, the lady pulled out a box containing about one hundred small pieces of folded paper. “Pick one,” she said, “the one you picked will decide the discount you have today.” Aware of my hesitation, she added:”We make sure that 30%-35% is the most common discounts.” I picked one, opened it, and it said “25%”.
Out of curiosity, I asked the lady why they decided to have this sale and how they designed it. She explained that they didn’t want to go through the trouble of changing price tags, and they just arranged the number of the different discount levels as the normal distribution. “In this way, it is like a game,” She said, “and some customers come back again and again to try their luck.”
I think this is a very good promotion strategy. Unlike the revised labelling method, drawing lots made information asymmetrical and skewed in favor of the sellers. I would never know whether they truly obeyed the normal distribution. Moreover, the fact that those discounts are discrete instead of continuous means more concentration around the mean (30% ad 35%) and smaller tails, making the chance of getting great discounts slimmer. For consumers, they will feel somewhat obliged to buy an item once they pick a discount, so they are easily “trapped”.
The story ends with me leaving my email address to the lady for further notice of their sales and walking out of the door with the pink shirt. The quality of the product is what really matters.