Optimizing Cloud and Server Costs for a Fantasy Sports Giant

About the Customer

Our client, a prominent player in the fantasy sports industry, developed a popular application for NFL enthusiasts. This platform allowed users to create fantasy teams and earn rewards based on player performance.

Problem

• The client wanted to develop an app for both Android and iOS platforms. This app would enable users to view the game's updates in real-time and allow them to participate in bets as well.
• The client wanted to partner with various providers so that the user can redeem accumulated rewards from various bets.
• In terms of DevOps practices, the client was looking to achieve high availability and scalability; especially horizontal scaling so that each time a game is available, a new infra needs to be deployed.
• The client also wanted the instance to be destroyed once the game was over so that there was no extra instance running.
• The client was also peculiar about the infra setup cost and didn’t want a steep rise in the overall infra costs.
• A predictive analysis system in place was another requirement of this project. The system would give suggestions to users about the results of a bet.
• The client wanted to launch the app into D2C space for the NFL league that was going to start within 4 months of the completion of the ideation phase.

Our Approach

• An Android and iOS application was developed to allow users to have an interface in order to view the live games happening around.
• The client was able to explore a SAAS offering using the app and partnered with various sports agencies to monitor a player’s performance.
• AR was used to monitor the trajectory of the ball passed, the speed of the shots, and accuracy in terms of short passes and throughputs.
• Sports agencies even used the AR feature to view the placement of the players on the field.
• Moreover, using the AR model, sports agencies were able to view the live positioning and were able to pre-plan the game in terms of how to pass the ball from the wingman to the striker.
• The client was able to replicate specific scenarios using AR, such as how to respond when being attacked by the opposite team, by placing the players virtually on the fields and replicating different scenarios related to the trajectory of the ball being passed and shots being taken up on the goal post.
• All this setup was developed as a real time monitoring, therefore there was no delay and all the triggers were being triggered in the real time.

Benefits

• We were able to launch the app within the stipulated time i.e. 3 months and an additional month for the UAT with the beta customers.
• The client was able to onboard more than 20 agencies within a month of the app going live.
• Horizontal scaling was achieved which allowed high availability at a low cost.
• For the trial purpose, predictive analysis was a success and was able to achieve around 74% accuracy.
• Using the AR model, sports agencies were able to plan well in advance and boost the player’s performance by around 40 %. This would be enhanced with time.