- August 26, 2014
- August 25, 2014
- August 19, 2014
Those familiar with the hit TV comedy “The Big Bang Theory” most likely find the computer-like mannerisms of main character Dr. Sheldon Cooper to be funny and endearing. But when Revenue Edge President Stacy Westbay played the Friendship Algorithm – a scene in which Sheldon gets stuck in an infinite loop of defined parameters and outcome sets requiring human intervention to free him – it was a comical eye-opening look at striking a balance between predetermined algorithms and human interaction and when to effectively “game” the process.
Balancing human and technological ingenuity is all about equalizing, according to the Can You Beat the System: Human vs. Algorithm session at the 2013 Apartment Revenue Management Conference that also featured Carmel Partners Director of Revenue Management Jessica Mills and Simpson Housing Senior Vice President of Revenue Management Bryan Hilton, both of whom challenged that neither software nor human capabilities alone are as successfull as the two working in tandem.
Indeed, Carmel Partners often faces misconceptions from their leasing team including: “Residents who always pay on time should get a break on renewals,” and “Comps with higher rents means you can always raise your rents,” necessitating a balance between the algorithm and the human. “The amount of computer programming is huge,” Mills explained. “Humans couldn’t think of all the parameters to set, let alone replicate all that data. But there are things the system does not realize such as construction on the street outside. You need the algorithm but need the people to analyze that data.”
For instance, had it not been for the availability of the human element and their ability to analyze data outside of the system Carmel would have never learned their pet policy was a key driver for success in the company’s Denver markets. “The system can’t drill down into data to learn that,” Mills said.
But how much human interactivity is too much when it comes to pricing apartment homes? At what point should revenue managers trust in the system? And at what point should they game the process? “Look at the numbers,” advised Hilton. “Computers can’t predict change but are designed to predict patterns and trends. The system may take that negative history and forecast incorrectly a year later because of it.” He challenges to game the process by examining what happened in the past and determining where the market was. “If you ask all the right questions, you get to the correct pricing recommendation 95 percent of the time. Generally if you dig down enough you can find why you need to go with the system.”
Westbay put it simply, “You have to drive the system and be able to work the two together to avoid the infinite loop.”
For more, download the Can You Beat The System: Human vs. Algorithm presentation.
This information was presented at the 2013 Apartment Revenue Management Conference.