Digitalisation has allowed cinema audiences to consume content with more flexibility choosing where, how and when they watch. ‘Smart’ digital platforms can also profile users for customised content. This leaves cinemas struggling to sell seats and retain a customer base. The EU-funded ITDS project used artificial intelligence and data analysis to predict likely box office outcomes. The project website Movieday, contains a platform, called ITDS, that hosts a portfolio of undistributed films. Using artificial intelligence and data analysis to predict likely box office outcomes, the platform proposes films for individual cinemas based on these projections and analysis of local audience characteristics. To date, 50 000 people have purchased 140 000 tickets, to see 110 movies, across 300 associated cinemas.
Reducing risk for cinemas
ITDS’s research shows that 83 % of films shown in cinemas are big-budget Hollywood blockbusters, but these represent only an estimated 8.8 % of movies made. The remaining films are often not shown in cinemas as distributors pursue a safe return on investment. “Cinemas have struggled to adapt to modern content trends. Despite adding 3D features and showcasing bigger blockbusters with more elaborate special effects, they still seem paralysed by video on demand (VoD). They have not adapted their business model to the 21st century,” explains Antonello Centomani, project coordinator. With new filmmakers viewed as too commercially risky by the bigger distribution companies, largely because they have not yet developed a fanbase, the project team saw a gap in the market. Research indicated that cinemas are also typically missing opportunities when they fail to promote films with local cultural connections or allow audiences to influence which films are screened. The project’s game-changing innovation was to develop promotional campaigns which drive the presale of tickets through Movieday. “ITDS connects the right parts of the value chain. Our accurate customer analysis introduces new talent to audiences, making that talent attractive to cinema owners, by reducing commercial risk,” says Centomani. By learning about its customer base, Movieday can more effectively craft a sense of community for shared filmgoing experiences. The prediction engine, powered by machine learning and its set of algorithms, can analyse customer behaviour to identify the best schedule for a given movie, further maximising attendance.
Following the crowds
About 35 % of people attending Movieday events do not usually attend cinemas, and screenings of the promoted films record an average of 110 moviegoers (an average venue capacity being 200). Targeted promotions have filled even the weakest time slots, achieving a 100 % audience increase for mornings and afternoons, and 80 % for the first days of the week. Research indicates that participating cinemas have increased their annual profit margin by 5 %. Streamlining the film scheduling process, while building audiences, has the potential to benefit venues beyond cinemas, such as NGOs, schools, libraries or bars – fitting well within the aims and objectives of the EU’s Digital Single Market Strategy. More broadly it also supports community spaces, resources and activities, while also boosting local economies. In order to fully implement the AI engine and data-analysis engine into their prototype, and scale up internationally, the project team are currently seeking further investment.
ITDS, cinema, movie, film, video on demand (VoD), distribution, screens, audience, filmmaker