The days of relying purely on human experience to market movies are over. From propensity to marketing automation, the future of cinema marketing is here - and we’re ready to make the progression as seamless as possible for movie marketers.
Our Movio Cinema product team is now shaping our cinema exhibition marketing solution of tomorrow, and, we thought what better time than now to discuss the coming year with Chief Product Officer, Peter Beguely, and Chief Data Scientist, Dr Bryan Smith.
In this interview, the duo share their vision for Movio Cinema in 2018, explore how movie marketers can get prepared for what’s to come and explain exactly what all the buzzwords mean.
What is the vision for Movio Cinema in 2018?
Peter: For 2018, we have three key focuses for Movio Cinema: propensity, identifiable moviegoers and marketing automation.
The first, propensity, is looking at the likelihood of a loyalty member watching a movie. This will replace the traditional approach with a more scientific one. It will streamline audience creation and segmentation, and ultimately increase the performance of campaigns and quality of guest experience. That’s the big focus for 2018 - using science and actual data alongside human experience, rather than relying solely on intuition.
The next concept we’re exploring is identifiable moviegoers - those individuals who may not yet be signed up to loyalty. We’d like to empower our users with the tools to curate, monetize and enhance guest experience for everyone interacting with their cinema, and not just their loyalty members. This will then allow them to track the digital interaction and transactional data of both their identifiable moviegoers and their loyalty members - which is not a big leap from where we are today. Additionally, these insights could also be used to convert visitors into loyalty members. This deeper insight into moviegoer behavior is going to provide users with the information they need to make the best decisions to effectively run their cinemas and all guest touch points.
The third focus is marketing automation. Most of us are very familiar with this concept by now, but we will be really concentrating on how it matters to the cinema industry. We will address the key workflow, events and guest touch points that are specific to cinema. This will allow end-users to create experiences that span online and in-cinema, as well as freeing up valuable time to refine marketing strategies and test new theories.
What is Data Science bringing to Movio Cinema in 2018?
Bryan: This coming year, we’ll be adding machine learning and predictive analytics into Movio Cinema. We began introducing this in May 2017 with the release of our Movie Insights module, a decision support system that allows users to build a list of comp titles based on past guest behavior, without having to use their own personal knowledge. In 2018, we’ll double-down on that. Users will be able to set up groups of comp titles that we'll then use to generate member propensities for movies. In other words, movie marketers will be able to identify not just whether a cinemagoer wants to see a movie in a binary sense, but how much they want to see it.
One of the biggest challenges cinema marketers face is figuring out which group of people would be most interested in seeing a movie in any particular period of time. The theatrical propensity will allow users to rank a member's preference for each new release, empowering them to send optimized and personalized communications. This will change the way movie marketers undertake marketing - they no longer have to figure out the best audience themselves! By simply telling the system which movies are coming out and what they are about, Movio Cinema will use collaborative filtering1 and deep learning2 to predict what an individual member's preference will be for each of those new movies. This will allow Movio Cinema to make even more accurate recommendations and allow clients to determine the best offer or communication to motivate member behavior at the least possible cost.
Another development we’re exploring with member propensity is around customer lifetime value and churn modelling. It's easy for a marketer to look at a survival curve3 and say, "After six months since somebody’s last visit, there's a 50% chance they’re not coming back - we need to send out an email reminding them to return." However, that doesn’t really dictate whether an individual has actually churned. For example, if a moviegoer comes once every three months, and hasn’t been in two months, there's probably nothing wrong. But, if you usually visit every week, yet haven't in two months, there probably is something wrong. We'll be designing new models that take advantage of some of these modern machine learning techniques, like random forest4 or neural networks5, that can predict whether or not a member has churned. This will allow marketers to identify high-value guests who haven't visited the cinema in a longer time than they should have.
Lastly, we’ll be using reinforcement techniques to optimize the timing of automated sends. When a user is setting up a guest journey map, it's not uncommon to decide that a follow-up email after a member’s first visit is a great way to get them back through the door. But the question is: when should you send it? An hour later? A day later? A week later? There are sets of machine learning techniques that we can use to figure out optimal delay in sending these types of automated emails.
What exactly is machine learning?
Bryan: I read a great explanation once that said; there are two ways to program a computer to tell it what to do. You can give it a set of rules and say, "If this happens, do this." Or you can use data to teach it. You basically tell it, "So this is what the data looked like, and then this is what happened", and it tries to learn a relationship between what the data looked like at the start, and the result.
When we're looking at audience propensity, we have a vast behavioral history - up to five or six years of data for many exhibitors. You can go back to a period in time and say, ‘Here's a member that saw movie A, B, C, and D, and then on March 31st, they went and saw movie F. There were six different movies that they could've seen, yet they picked movie F.’ Marketers can then do the same with millions of members and thousands of different movies, and Movio Cinema learns the pattern between what people had seen before and what they saw subsequently.
How could this vision impact the way cinema marketers operate?
Peter: There's a fundamental problem facing marketers when it comes to using personalization correctly. Traditional segmentation tools rely on human input, with their own bias, to curate content for many audiences. Movio's new propensity-driven audience modules eliminate the human bias and countless hours spent by marketers on segmentation, to allow them to focus on creativity in content. Having the time to create content is what holds back a marketer from delivering a truly personalized guest experience. It's about increasing the speed and accuracy with science, allowing marketers to get back to their creative strategy, and building great experiences for their guests.
How could the industry prepare for this changing landscape in targeting?
Peter: Cinema marketers should begin by thinking about identifiable moviegoers and the guest experience as a whole. They should ask themselves, “What are we actually trying to create? Are we focused on building an incredible guest experience, as well as setting up practices to benefit the business?” Now may be the time for an internal audit to re-define those key objectives and metrics for success.
Once there’s a strong focus on creating excellent guest experiences, the byproduct of that is what the business really wants: engagement, frequency and increased revenue. This can be achieved by taking stock of all data, and beginning to prepare, strategize, and do everything possible on a technical front. Try to harness as many data points as you can, using tools like Movio Cinema’s Web Tracker which gives you actionable insights about your moviegoers interactions on your website, that will help drive segmentation and strategy.
Once those things are in place, you’re in a great space to start building on propensity, automation and identifiable moviegoers.
How does 2018 vision fit in with the bigger picture?
Peter: Our vision is for a quicker, simpler, easier-to-use Movio Cinema. We want to be a hub of rich data and insights, ready to drive truly personalized guest experiences across all touch points, not just your email or communication channels of email, SMS and Push. We've designed all of our extension modules as building blocks. Each module we release takes us closer to this objective. They build upon one another, making the previous module more powerful and more insightful than it was before.
We will deliver a platform that makes marketers feel like they have their own in-house Data Science team guiding marketing efforts, automatically highlighting opportunities and identifying future risks. We want to empower all Movio Cinema users with the tools to deliver the magic of cinema to each guest.
1Collaborative filtering - A method for making predictions based on the assumption that moviegoers or movies that co-occur in similar contexts (ie. moviegoers that have seen the same movies, or movies seen by the same moviegoers) are similar to each other.
2Deep learning - A class of machine learning algorithms that use multiple layers of nonlinear processing units to extract features and learn multiple levels of representation.
3Survival curve - A curve that presents the surviving percentage of a population vs. time
4Random forest - A machine learning technique that averages the contributions of many decision trees to make predictions
5Neural network - A computing system inspired by biological neural networks that uses connected nonlinear processing units to approximate mathematical functions.