This post is co-authored by Ye Xing, Senior Data Scientist, Benjamin Moulès, Developer Evangelist, Peng Xia, Data Scientist, and Tao Wu, Principal Data Scientist Manager at Microsoft.
Beyond evoking passion among fans around the globe, professional soccer is also big business. Top clubs draw tens of thousands of fans into stadiums each game and their annual ticket revenues can reach one hundred million dollars or more. Marketing and user engagement are critical in this industry since a few percentage points of difference in attendance can result in millions of dollars of revenue impact.
OpenField is a leading data management company whose solutions are used by one of the world’s best known soccer clubs as well as sports and performance venues around Europe. OpenField uses the Cortana Analytics Suite to transform insights into actions by enabling more efficient and effective use of data in their contextual marketing solutions.
Profit in the live entertainment industry is challenged by the perishable nature of its inventory. Maximizing the value of each available seat in a venue requires targeting the right customer with the right offer at the right time. More specifically, soccer teams must balance the need for revenue with the need for maintaining adequate attendance at the venue. Since the market for their events is generally fixed, ticket sellers are challenged by the problem how to target the right customers for a specific game, especially as someone attending a game one week may be less likely to attend a game the following week unless there is a compelling reason to do so. One particular soccer club experienced a significant increase in ‘no-shows’ by season ticket buyers. This scenario created a perception problem for fans watching their games on television and those in the stadium. No-shows can also be an indication that future renewal by season ticket buyers might be difficult, which, in turn, can have a significant impact on revenue.
One OpenField customer, a major soccer club, now uses machine learning to make predictions on things such as:
- The likelihood of a known customer buying a ticket to a given game.
- The likelihood of a known customer attending a game for which they have bought a ticket.
By utilizing predictive models, the soccer club can more effectively target potential ticket buyers and anticipate which current ticket holders are unlikely to come to a given match.
The predictive models for both tasks are binary classification models and were built using Azure ML and a knowledge of customers’ historical behavior. The boosted decision tree gave the best performance for both the tasks. In addition to Azure ML, the analytics pipeline leverages several other services such as Azure SQL Database, Power BI for visualization, and Azure Data Factory for data processing and movement coordination. The “customer likelihood score” will be output to the client’s on premise SQL Server data storage and the analysis will be available for visualization through Power BI. The complete pipeline is shown below.
OpenField’s End-to-End Analytics Pipeline
With predictive models such as these, OpenField can better target potential ticket buyers for upcoming games and proactively manage potential no shows. By using the Cortana Analytics Suite, OpenField helps customers turn entertainment venues into smart, data-driven revenue sources.
Ye, Benjamin, Peng & Tao
Contact Tao at email@example.com.