One of ClassPass’ most sophisticated features to date is SmartSpot. Our mission is to help our partners fill more class inventory and increase their visibility on ClassPass to drive more traffic to boutique studios. SmartSpot is a proprietary dynamic inventory algorithm created by a team of engineers that detects open spots in high-demand classes that are fully booked on ClassPass, but may otherwise go unfilled by class time. To allow studio operators to maximize their utilization without any manual effort, SmartSpot automatically observes and adjusts for changes in schedule behavior over time to open up unused spots, while always saving room for last minute walk-ins at the studio.
We spoke with Galen Hancock, Data Scientist at ClassPass; Chris Starace, Lead Data Scientist; and Jeff Bladt, Director of Inventory, who were each involved in developing and implementing the SmartSpot algorithm, to learn more about how the feature came to be and its impact on ClassPass studio partners.
What inspired SmartSpot?
Jeff: We knew that roughly 70% of spots at studios in the fitness industry overall (not on ClassPass) were going unfilled. That meant there was a real opportunity to explore ways to increase engagement with users and revenue for our partners. Our team looked at other industries for inspiration, in particular the travel industry. Airlines and hotels have been using dynamic algorithms for years to help increase utilization for their products, matching the right customer to unused inventory.
70% of spots at studios in the fitness industry overall were going unfilled.
Chris: I drew from my experience building financial models at a hedge fund to help create the SmartSpot model. As in finance, this model makes predictions based on a time series set of estimates, walking through time to set logic-based expectations around how many users a studio can anticipate having in a certain class. The algorithm looks at historical data including time of day, seasonality, and variance in fill rate based on a number of factors to predict which spots we may be able to monetize on our site.
How “smart” is SmartSpot?
Chris: For each class, SmartSpot analyzes tens of millions of time-stamped data points to calculate how many spots a studio will fill and, with some buffer, the number of spots available to ClassPass members. The model is optimized to maximize the studio fill by opening up additional spots to ClassPass without overfilling classes using a genetic algorithm.
Jeff: The product was built with our partners’ best interests in mind as a way to drive more traffic for our studio and gym partners. SmartSpot looks specifically at classes that are already in-demand on ClassPass — and that therefore are not showing up in user search — but that aren’t filling up on the studio’s end. It’s constantly checking the inventory and will adjust accordingly so that there’s always a buffer to allow for walk-ins and direct traffic.
What would you tell a studio partner who was interested in giving SmartSpot a try?
Galen: ClassPass is continuing to invest in search and brand exposure to drive users towards studios who have inventory available. SmartSpot is a great complement to this. The system will automatically detect if a studio historically hasn’t filled spots that we believe we can monetize on our site, without the owner having to manually manage this.
Chris: It’s important to note that, by SmartSpot opening up a limited number of spots on ClassPass, this is not taking spots away from a studio’s direct user base, but rather just opening up the option for ClassPass users to book and building awareness about what the studio offers. The intent for the algorithm is always to supplement a studio’s inventory, not overtake it. Furthermore, high-demand studios are less likely to be discovered in users’ search for classes because their classes are often booked. SmartSpot opens up the possibility to still be discovered by new users on ClassPass without cannibalizing the studio’s own traffic. It also works in unison with the other efforts ClassPass is prioritizing to help aid in studio discovery, such as home cards highlighting specific classes, email recommendations to users, and a revamped search page that emphasizes classes related to a user’s interests.