“More Magic for the ThemePark of the Future – A Revolutionary Real-time Big Data Project”
The Problem: Not-in-Time Data
An American Fortune 100 media company decided to take a big step by investing over a $1 billion dollars in the digital infrastructure of their popular theme parks.
The goal: a perfect 360° customer experience that would reach a whole new dimension through personalized customer service (such as addressing guests by name, providing personalized offers, sending congratulations for special events etc.) and optimizing resource planning for the park’s attractions. In short: more magic for the customer.
The client initially equipped park visitors with specialized RFID bands to wear throughout their stay: to be used as a hotel key, to provide entry tickets and skip-the-line verification, as well as to serve as virtual money for purchasing snacks, drinks and souvenirs. Every wish fulfilled with a magical turn of the hand!
Additionally, a multitude of sensors throughout the park continuously recorded the customers’ movements and this data was used to reduce wait times. This information was also leveraged by the theme park app, which assisted visitors in selecting and booking park attractions and shows. External information sources, such as weather data, also fed into the optimization of the customer experience. For example, when there was a risk of rain, indoor attractions were promoted.
However, the client soon recognized that implementing this “Next Generation Experience” was much more complex than initially assumed. More than 150,000 daily visitors generate a total of 3.8 billion data records which are stored in more than 40 disparate data sources. Of these, approximately 500 million are actually referenced for analysis purposes, daily.
The park’s legacy ETL and SQL-based business intelligence systems (in use since 2009) were not capable of handling this mass of data quickly enough. It frequently took more than 36 hours to provide meaningful analysis based on visitor data, by which point, the park visitors had already returned home.
Management was alarmed. A quick solution was needed to minimize reputational risk. Several big-talking vendors were engaged but failed to deliver a solution and ended up feeling small as they cleared the field. No improvement of the situation was in sight. Finally, in a courageous experiment, the client decided to involve a real-time big data specialist that was not well known in the US.
The Solution: Show Time For dataWerks!
We were given the opportunity to present our solution and familiarize ourselves with the client’s environment. We certainly understood why previous solution providers left the project in exasperation. The client’s systems landscape was complex and consisted of many legacy databases and heterogeneous silo systems that had been in use for many years.
It was, for example, not possible to connect visitor-based data across data silos. The system was just not able to provide reliable or meaningful information at all. The client’s Next Generation Experience initiative added more data sources, which generated even greater masses of data.
Query response times were completely unacceptable due to the high number of disparate data sources and the massive volume of data. Even though minimizing latency times had been a key focus for the client’s internal IT department for years, neither conventional data warehouse solutions, nor solutions from data virtualization market leaders were able to meet the client’s requirements.
To prove our capabilities, we were invited to participate in a Proof of Concept. Equipped with a single standard desktop PC with four cores and 16 Gigabytes of RAM, we flew to the US and successfully integrated with a pre-selected set of production systems. We created multiple data mashups and response times were more than impressive. All of their information was now available in real-time!
We hit the target right on: all client requirements were met 100%.
The Implementation: Off to the Races
Shortly after the contract was signed, the project was scoped, business and technical requirements were defined and the project was kicked off. We sent a developer to Orlando to implement this project onsite. His comprehensive experience and expertise would help ensure maximum efficiency.
Facing tight deadlines, we had a Hackathon for several weeks in our Frankfurt office. Four developers were added to the project and they adjusted their daily work schedule to synchronize with the US-based developer.
Many of the client’s data queries turned out to be incorrect and were corrected within a very short timeframe. In less than six weeks, dataWerks successfully implemented and tested 70% of all requirements, including a newly built client-specific dynamic business rules engine.
And Now: The Happy Ending
Finally, on 24 August 2016 at 7:00 a.m. Eastern Standard Time, the solution went live with Champagne corks popping. Our dataWerks team continued to monitor system operations 24/7, correcting small issues on the fly. Within four weeks, operations for our solution were fully handed over to the client’s operations team.
During the first year, system support was also transferred to the client in incremental steps. Today, the client is able to implement new system requirements, such as integrating new data sources or defining new business rules, without our involvement.
Ultimately, within a year, and with a little bit of magic from dataWerks, the client’s ambitious growth target of “increasing annual revenue by $500 million US dollars with an operating margin of 20%” was realized.
Before you arrive at the theme park, modified iPhones provide theme park employees with information about your birthday, language, culture and meal orders. Of course theme park employees can also locate you and your children wherever you might be.
Thanks to the dataWerks solution, the client continues to meet their growth targets year over year: 500 million U.S. sales growth with a 20% operating margin. If this growth trend continues, this billion dollar investment will pay off significantly.
Sources: Harvard Business Manager (2017); Forbes (2017)