The Client and Setting

The client was a leading ticketing firm. They sold tickets for a large range of live events, and had developed a great relationships with a number of firms on whose behalf they sold tickets. This gave them leeway in terms of how they would price tickets. 

Client Mandate

It is virtually impossible to know the demand and appropriate price for an event before we have sold a single ticket. We have had cases where we have sold out tickets within minutes of them going on sale. And for some other events, we have had the displeasure of seeing lots of empty seats.

We believe that the only way to get around this problem is to adjust ticket prices after we have a better idea of the demand – maybe some sort of dynamic pricing? This way, any mistakes we make in pricing will be less costly and we can provide greater value to our clients.

Problem Solving

Through the analysis of a large number of past events, we were able to understand patterns by which tickets got sold. Some events, for example, exhibited a pattern where most of the sales would happen on the day of the event. For some other events, the sales would be frontloaded.

The challenge in this case was to determine, after ticket sales for an event had started, what the “true demand” for the tickets was at the price at which it was being sold, and how changing prices could adjust that demand.  I used concepts from pricing of financial derivatives, along with a model to determine the price sensitivity for customers, to determine the true demand. I completed the model by incorporating a layer of risk pricing, and revenue management practices from the airline industry. 

The process of model development was iterative, as at each iteration of the model I presented to the clients the possible results of the model, and incorporated the feedback in order to improve it. 

The final deliverable was a comprehensive model, which the firm’s engineers then used to incorporate into their overall pricing system. 

Results and Outcomes

I handheld the firm’s engineers and internal data science team to make sure that the model was internalised and incorporated into the firm’s pricing product. As it happened, the model was to be launched as part of a big-bang event that had been scheduled in April 2020. 

With the pandemic resulting in postponement of the events, and live events not having yet made a comeback (at the time of writing this), we have not yet got a chance to live test the model and measure its impact to the client’s clients.