With a case study centered on LEGOLAND properties in the US, the panel focused on ride operations, specifically objectives around ride occupancy and guest happiness. This was achieved through improvements for data accuracy, scale and insight, by digitizing processes, automating data collection and using analytics and AI. 

LEGOLAND, a brand close to home for Merlin with owners Kirkbi A/S, is geared towards families with 11 parks in operation globally and 4 more on the way. Within the US, these locations include the brand’s first North American site in California, in operation since 1999 and adjacent to a Merlin SEALIFE aquarium, in addition to a subsequent location in Florida which also incorporates a Peppa Pig attraction and a recently opened resort in New York which saw its debut in 2021, including a  themed hotel.

Adrian Thompson, Regional VP Improvement & Planning at Merlin Entertainments introduced LEGOLAND’s initial starting point, with complex, high volume procedures for ride operations, including safety checks, loading and management. Traditionally, operations have been performed with pen and paper, radioing reports back to a central command and collating clipboard results at the end of the day. 

Through digitization, automation and insight, LEGOLAND sought to improve the accuracy, real time nature and actionability of data. In addition to a core focus on safety, Merlin’s key performance indicators include metrics driving guest satisfaction, such as rides per visit and supporting metrics such as ride downtime and throughput. With this program, Merlin accomplished a 30% increase in average ride throughput, translating to an additional ride per guest outcome.

Owen Jones, Product Director at CmdCentr, showcased a dedicated operator app guiding workflows and automating manual tasks, with data digitized and centrally available. These processes covered everything from operator sign in and checks through to ride status and assistance and connecting with dispatches, capacity and throughput data. Owen touched on the need for simplicity in problem solving through design, noting an example where offering too many reasons for a ride downtime led to impacts in the quality of data collected. The team also witnessed learnings from having data available on previously hidden interactions, such as frequent call outs for water bottles to be delivered to ride operators that could be resolved through having refreshments available at the ride for staff rather than managers running water across the park. Owen stressed this project was as much about redesigning processes as introducing technology. 

Ali Vahabzadeh, CEO of Safari AI, discussed the switch from manual clicker counters to vision AI, discovering that traditional click counts were only 87% accurate. With vision AI, a live camera feed is used to detect the presence of visitors, enabling precise metrics on rollercoaster loading. Although this technology can be rolled out across existing camera equipment, new cameras were procured and installed at the ride sites for LEGOLAND. To protect visitor privacy, facial recognition is not used and visitor identity is not assessed beyond the video feed, leaving only the metrics useful to understand ride loading for capacity and throughput. With this development, Merlin discovered a fill rate opportunity with 10% of seats leaving the station unseated, resolved by a tweak to single rider entry procedures. Parks also switched to coordinating staff lunch breaks and other communications via headset rather than intercom or phone, enabling staff to be handsfree and avoiding delaying dispatches.

Angie Judge, CEO of Dexibit, rounded up the discussion of once data is digitized and automated, how to leverage it all. She explained how the challenge for visitor attractions is that their business models are complex, with lots of data coming from different lines of business, meaning many source systems, crossing various domains. Diving into one such example, Angie pulled up over 70,000 Google reviews from LEGOLAND’s three US properties, using Dexibit’s voice of the visitor AI to turn this freeform visitor feedback into enriched quantitative insight, covering aspects such as sentiment, emotion, theme, topic, complaint identification, suggested reply, prescribed action and more. Dexibit AI was also used to transform data, for example taking visitor descriptions of wait time and turning these into quantitative wait time perceptions in minutes to compare back to actual queue wait times from other sources. Angie also demonstrated this data at work within a conversational insight setting, where a data consumer can ask the voice of the visitor a question then have Dexibit AI respond, informed by real customer data. Within LEGOLAND, this led to discoveries to help smooth post opening operation of the New York location, confidence around priced propositions in Florida and an understanding of the sensitivities of age appropriateness – an important theme for LEGOLAND with a well defined appeal to a specific target market. This data provided actionable insight not just for operations but also other functions such as marketing. 

The panel wrapped by addressing a series of questions relating to the importance of navigating decisions around change such as this for understanding when to buy existing products, collaborate with vendors for new solutions or build in house; the need to treat change as iterative and important considerations for data privacy, security and culture for leadership.