AI helps L'Équipe with better forecasting newspaper demand

Global consultancy BearingPoint has helped world-famous sports paper L'Équipe with tapping the power of artificial intelligence to better forecast the daily demand for its printed newspapers.

L'Équipe is a French nationwide daily newspaper devoted to sport. Cited as a trusted source by the world’s top sports journalists, the paper is particularly noted for its coverage of football, rugby, motorsport and cycling. The biggest-selling issue was 13 July 1998, the day after the France national football team won the World Cup.

On that day, L'Équipe sold 1,645,907 copies – however, demand for the printed press has significantly declined since the advent of the internet, and as a result the newspaper has had to evaluate how it produces and distributes its printed copies.

Continuing to overestimate sales would result in papers being left unsold, creating a financial deficit and unnecessarily large carbon footprint; however, underestimating demand would lead to a shortage of copies and missed opportunities to make additional sales.

While the importance of anticipating the actual demand for copies of L'Équipe was clear, however, it was easier said than done. Accurate sales forecasts of print editions are often impeded by the fluctuating nature of sports events – as well as the influence of generic factors like weather conditions. At the same time, the paper is currently sold in approximately 25,000 vendor locations throughout France, with each retailer having allocated a specific average sales volume.

In order to get on top of the situation, L'Équipe joined forces with BearingPoint to address the issue of sales forecasting. This saw the paper’s in-house data science and media team subsequently leverage BearingPoint’s own Nitro AI-driven solution – a set of managed services that use machine learning and an artificial intelligence approach to focus on print-media supply chain forecasting – which was developed specifically for publishers.

The project saw L'Équipe input vast amounts of historical data, including units sold by period, units delivered, holiday sales, and weather variations. Nitro used this information, as well as new data added every day, to produce forecasting models that predict the exact number of paper copies needed by every retail location for the following day.

As a result, L'Équipe has since managed to lower the rate of unsold units by more than 5%, cutting the number of wasted print copies by an estimated 2 million. At the same time, the paper’s distribution planners are now able to easily adapt to different scenarios, relating to the impact of current events on sales.

Eric Matton, Editor & Director of the print division of L'Équipe’s owner Groupe Amaury, commented, “With Nitro, we have an autonomous vehicle whose engine settings are optimized automatically, giving us more control. Our demand planners can fully develop their know-how and correct the projections by adjusting the forecasts made by the machine.”