How Multiplex Achieved a 5.2%+ Revenue Lift Using AI‑Powered Forecasting

Multiplex is Ukraine’s largest cinema network – a company that brings millions of people together through film. With 26 theaters across the country and a loyal audience base, it has long mastered the art of creating experiences that last. Over the years, every ticket, visit, and purchase generated valuable customer data. But raw data, without analysis, remains an untapped resource.

Recognizing this, the Multiplex team set out to uncover the story behind the numbers. Which films actually drive return visits? How can scheduling improve performance beyond peak hours? And how can personalized communication strengthen customer engagement?

That’s when Multiplex partnered with S-PRO. The goal was clear: to transform years of scattered audience data into actionable intelligence – a system that informs smarter decisions in scheduling, marketing, and audience management.

The first goal was to understand audience habits and use those insights to recommend the right films at the right time. The team wanted to apply this knowledge in personalized campaigns – sending people recommendations based on what they watched and when they usually went to the cinema.

The second goal was to address the cinema industry’s uneven attendance patterns. Prime-time evenings were always packed, while other hours remained nearly empty. The team wanted to use data to understand which films could fill those quieter slots and keep theaters active throughout the day.

To achieve this, Multiplex first faced its biggest challenge – collecting and cleaning years of audience data. It was the company’s first project of this kind, and the task was to bring scattered information into order. Addressing these goals required a technical solution grounded in real business logic – one that could process complex data, make predictions, and support daily operations.

The first part was customer segmentation. The team analyzed customer data, found several clusters, and understood which audiences preferred which films and how often they visited cinemas.

The second part was uplift modeling. Using audience insights, the system modeled the cinema schedule – predicting which film to show and when, and how much potential revenue each combination could bring.

The third feature focused on marketing insights. Based on data, Multiplex could create targeted email campaigns with personalized recommendations for moviegoers.

The system also included a manual option. A manager could adjust the schedule and request auto-recommendations from the model. All customer data was processed under a data processing agreement. The data was anonymized – personal identifiers were removed, so the system never showed individual profiles.

Once the system was in place, it became clear how these insights could influence not just operations – but overall business performance.

We validated this by benchmarking against “twin venues” – other theaters with comparable historical data – to ensure the lift was driven by our optimization. For instance, while control venues followed standard attendance patterns, our data-optimized scheduling captured higher engagement, effectively boosting performance in traditionally quieter time slots.

The project is finished, and Multiplex already uses an AI-driven system that analyzes audience behavior, predicts attendance, and supports marketing and scheduling decisions. This approach helps the company identify demand patterns, optimize film distribution across time slots, and estimate revenue potential with greater accuracy. Adjustable for your business model and constraints

A stepped setup tunes every run to your specific rules before the model generates the optimised output. Pick the time window, the items to include, the constraints. Click once, the agent builds your plan. Cinema example shown: dates, films, constraints, review. In a warehouse: SKUs, locations, limits. In a factory: lines, shifts, batch rules.

Comparison to control data sets

We validate our forecasts by comparing them against real-world control data to isolate the true impact of our optimization.

Instead of just looking at a final number, we benchmark performance against peer assets or historical periods. For example, during the Multiplex AI optimization agent pilot, we measured the 5.2%+ revenue lift against a 95% confidence interval, established by twin venues – similar, control theaters that did not use the optimization – to confirm the growth was directly attributable to our system.

Optimisation after 5 iterations

Read-outs show the agent improving against your own baseline over time. Whatever metric matters to your business, tracked by day, hour, or segment. Cinema example shown: day-of-week lift versus the pre-test average and average tickets per session by hour.

The partnership between Multiplex and S-PRO became a milestone for the cinema industry. It proved that data can do more than describe the past – it can predict, guide, and drive growth. By combining deep domain knowledge with AI, we turned years of audience insights into a working system that supports real business decisions.

Beyond operational gains, the project showed how technology can reshape an entire industry. It opened the door for cinemas to manage audiences the way digital platforms do – through data, precision, and continuous learning.