Patxi's, a San Francisco pizza chain, cooks up an innovative ordering process
My local pizza restaurant just rolled out a new service that allows diners to select a pizza that's already in the oven, thereby cutting down the wait time. This restaurant serves deep dish pizzas that take 45 minutes to bake, so the advantages to the restaurant and to the diners are considerable. The diners get to eat sooner -- which for the target audience of families with children under 5 in my neighborhood is very important -- and the restaurant gets to turn over the table faster.
The diners view a tablet which shows the type of pizzas that are cooking and how long it will be until the pizza is ready. One of our tech savvy dinner companions wondered if the information was dynamically generated, while another friend who is a mother of two young children looked at the pizzas listed and she said "oh they must just figure that a lot of families eat at 5:30pm and you can't go wrong with a cheese pizza".
So I asked our waiter about the system when he returned with our pizzas. He told us they created an algorithm to determine what customers are most likely to order at any given day/time. The thing that stuck with me is that he said it took a year to launch.
I applaud their innovation in creating the algorithm, but I have to wonder - could they not have just used a bit of intuition to figure out that cheese pizzas sell well at 5:30pm? As they were gathering data and creating the algorithm, they could have been testing the pre-ordering system using a set of assumptions that they could test and validate. The risk of starting with testing would be some uneaten pizzas, while the risk of waiting a year to launch an algorithm is one whole year longer before you realize your goal of faster turnover.