The most volatile period in the history of the movie industry began on March 4, 2020, when James Bond blockbuster No Time To Die, the 25th instalment of the world’s longest running film franchise, was pushed back because of the Covid-19 outbreak.
“As an industry, we have already lost $5bn – more than ten per cent of the annual global box office,” says Dimitrios Mitsinikos, cofounder and CEO of London-based Gower Street Analytics, the company used by major Hollywood studios schedule movie releases in marketplaces across the world.
With cinemas closed, other films followed Bond’s lead, either by not choosing a release date or moving far into autumn – or to 2021 – and now, Hollywood studios are scrambling to reschedule blockbusters in order to recoup the vast budgets invested in their production. If lockdown restrictions lift, autumn 2020 could be rammed with big-ticket movies competing for the attention of moviegoers.
Picking the right time to release a global blockbuster is complex work at the best of times, and Gower Street Analytics could play a key role by finding the most lucrative release dates for studios, using a combination of data science and human expertise. “What makes it possible for our platform to still function under these extreme market changes is that human knowledge and algorithms work together to create the output,” Mitsinikos says. “Systems based purely on machine learning are not able to adapt as easily.”
Data-driven decisions aren’t new to Hollywood – in fact, they helped create it. In 1911, the Nestor Film Company relocated from New Jersey to Los Angeles, prompted by a report that suggested that better weather and more days of sunshine would make for a more profitable studio. It built the first movie studio in Hollywood at the corner of Sunset Boulevard and Gower Street, and the following year, it changed its name to Universal Pictures.
Gower Street Analytics, which takes its name from the location of that first studio, now has staff across seven different countries. Its software package, called "FORECAST", allows theatres and studios to simulate box office numbers – not only to find the optimum release date, but also to inform levels of staffing and stock, ensuring, say, that popcorn supplies don’t run out.
“Data is global, by day, by film, by cinema and in some cases by showtime,” Mitsinikos says. The software adds information about the cast and crew, genre, running time, format and age rating, and estimates about how widely particular films will be released.
Seasonal variations can make a big difference, too – in Russia, sharp temperature changes can cause drops of 90 per cent or more in cinema attendance over the course of a weekend. "While we cannot predict when this will happen, we can highlight the risk with a wider than usual estimated box office range – this is how we flag risk,” says Mitsinikos.
Even sporting events can play a role. “If a country participates in the World Cup, we need to know the probability of them progressing to the later stages,” he says. “We need to know fixture dates so our algorithms can incorporate their possible effect.”
In previous years, when, where and how a film got released would be based on how well a similar film performed in the past. FORECAST can simulate any film moving to any date on the calendar. “Our analysts have evaluated the kind of audience most likely to go to the cinema around specific days, such as Valentine’s Day,” Mitsinikos says. “Between the platform and our experts, we can generate real world simulations, even in unpredictable times. Different films react differently to changes, because of their target audience and other factors. Our platform can show whether that exact date improves a film’s commercial prospects or not.”
A complex process that used to take days now takes seconds – often taking place within a meeting. But release dates are not set in stone. A massive blockbuster is like a shark in the water – the other fishes will scatter or be devoured. “Twenty smaller titles that move on dates around you over a period of six to nine months can have the exact same effect,” he adds. “Each decision can make a difference of hundreds of thousands or millions of dollars in box office performance.”
To predict a film accurately in a normal year, there are more than 25,0000 release date changes that need evaluation. “Right now, we have more usage in the system than ever before,” Mitsinikos says. “The studios are trying to find new ways to date films when now all bets are off. Our simulations are an additional safety net.”
Gower Street, which – like many others in the industry – had to furlough some of its staff in March, is now using its data to explore ways in which the cinema industry can recover from the effects of Covid-19. “The market will be at least $10bn behind the norm when this crisis is over,” says Mitsinikos. “This recovery process is something that has never been done before, so it requires some out-of-the box thinking. But innovation is in our fabric. That’s how our R&D process works.”
One idea to revitalise the industry is to create a “Global Celebration of Cinema Experience”, a month-long film festival with greatly reduced ticket prices so cinema-goers can rekindle love for the big screen – a modified version of France’s annual Fête du Cinéma.
“Don’t forget that the biggest film admissions year is still 1946, the year immediately after WWII,” Mitsinikos says. “Yes, people were going to the cinema a lot more and the situation was slightly different. But cinema is a nice form of escapism. The whole industry may take that role again when people feel safe.”
This article was originally published by WIRED UK
