L'Oréal

Demand Sensing

How to Forecast Sales at the Age of Data & AI?

How to forecast sales at the age of Data & AI?

Traditionally to forecast sales, you only use the past to project the future. In today’s world, this approach is no more relevant. Although forecasting is still a true craftsmanship, Tech helps to make it a science.

Our Beauty Tech program ‘Demand Sensing’, developed in our Tech Accelerator, re-invents demand forecasting process in a digital world, leveraging data, consumer insights and machine learning. It is a key enabler of the digital transformation of L’Oréal’s Supply Chain.

How does it work?

Accessing multiple high frequency sources of data through connected data platforms optimizes the sales comprehension and anticipation, allowing machine driven planning across the entire distribution network and ensuring the right stock is in the right place at the right time, automatically.
Currently being rolled out at Group level using Agile methodology, the project mixes people from Business, Supply Chain, IT and new skills of Data Engineers and Data Scientists in one team with a common ambition.

What is the magic behind?

L’Oréal has built the Beauty Tech Data Platform with Google that compiles all the relevant data and then uses algorithms and artificial intelligence to automatically create more detailed and more reliable sales forecasts. The data and underlying drivers are exposed to marketing, sales, digital, finance and supply chain departments, driving closer collaboration, breaking down the traditional silos and focusing exchanges on true business drivers rather than on absolute forecasts.
All inputs are integrated in real time enabling algorithms to propose the best decisions to demand planners at business pace. And that’s a real revolution!

The expectations are huge: improve sales forecasts accuracy, product availability, optimize inventory and reduce obsoletes, improve how to steer the business, identify more quickly any changes in trends and low signs of sales acceleration. Magic, right?

Back on Top