Enterprise Services Outlook Logo

The Evolution of an Analytical Framework for Ticket Pricing

By Alison Burnham, And Chief Information Officer, Score Big Inc


Alison Burnham,

At Score Big, we are lucky to be in a data-rich business where analytics can be used to provide a strong competitive advantage. We sell tickets to live events–sports, concerts, theater and family events. At any given time we have inventory for many thousands of events across the USA and many millions of tickets for those events. Our inventory comes from thousands of partners and in today’s highly dynamic ticketing market the prices for those tickets change in real time. This ecosystem of events, seat locations and ticket prices, is in itself a vibrant market where the performers (a sports team, an artist’s tour, a Broadway play, or an ice show) set the initial prices and the secondary market sellers take positions in the market by buying the tickets for resale. The final consumers of these tickets drive the economics of the market by deciding if and when to buy and at what final price. The price point and timing of the sale are dependent on each seller’s understanding of how to maximize their return on their investment. On top of this transparent market we run an opaque ‘Name A Ticket Price’ market that allows us to offer savings of up to 60percent from a performer’s initial price outside of the transparent marketplace in return for flexibility on exact seat location. This allows us to tap into the roughly 40percent of perishable inventory that currently goes unsold without devaluing the rest of the inventory with transparent sale offerings.

"We run an opaque ‘Name A Ticket Price’ market that allows us to offer savings of up to 60 percent from a performer’s initial price outside of the transparent marketplace"

Behind the scenes we have built a system capable of ingesting inventory data from all of these sources (both primary performers and secondary sellers), then pricing and curating that inventory to provide the best deals to our customers while preserving our own (and our partners’) margins. We determine what inventory can be offered in the ‘Name A Ticket Price’ format and what inventory should be sold at a fixed price. This has evolved from an initial set of hard coded rules in our system to multiple layers of home-grown pricing engines that allow our pricing team to optimize these rules without requiring development resources. The initial pricing rule capability was table based with a web front end for a business user – the underlying code read the values on the table to drive pricing calculations. Over time we found the very linear nature of tabular based rules too restrictive and felt a need to add more flexibility to our rules system. To accommodate that our development team put together a scripting environment in CoffeeScript (which we’ve coined ScriptBig) to allow us to write code easily managed by pricing analysts that is then compiled at run time into JavaScript and run to do the calculations required. We now have several components of our pricing engine managed in scripts with others still table based. Our pricing team do all of the management and Quality Assurance of this system giving them strong control and flexibility around pricing.

We have also built a reporting and analytics environment that houses snapshots of inventory, prices and sales data (both offers for ‘Name A Ticket Price’ and sales overall) for us to leverage time-based analysis in our pricing. Presently that additional layer is done mostly off-line by our pricing team and shared with our partners’ and their analytics teams. This analysis is done using Tableau which gives us a great ability to dive deep into the data and then present out our findings in both statistical and graphical terms. Our biggest challenge on this front is the size of the data we are managing and adding to the system on a daily basis. Our Business Intelligence infrastructure team is currently working with Tableau to find new ways to optimize this process. The next step on our pricing roadmap is to automate some of the analysis into a machine-learning capability built into the system that will replace roughly 80percent of the current off-line analysis. This will allow us to scale our partner base with our current analytics team, increase the speed with which we react to new pricing trends and have the pricing team concentrate on the additional 20percent of analysis that requires human expertise.

The increase in complexity of our pricing as well as curation and merchandising logic has led to us having to greatly beef up our processing power, memory and fine tune our databases. We continue to work on that by investing in larger and better servers, but we realize that eventually we will hit a point where this will no longer be enough. In parallel with our efforts to continually improve our current infrastructure we are also investing in a project to move most of this computation into a new Azure-based system utilizing the capability of Microsoft Service Fabric to manage the processes in the cloud. We anticipate being able to transition to this new platform in the first half of 2016. Overall, the ticketing industry has thrived due to a large number of buyers and sellers who have years of experience and knowledge of their own market demand. However, automation and analytics in this space has been historically sparse. We plan to continue leverage our systems and in-house analytics expertise to address this gap and provide a competitive edge to ScoreBig and our partners.