Procter & Gamble
Almost five billion consumers around the world use Procter & Gamble (P&G) products. The company, which was founded in 1837 by an Englishman and an Irishman, owns personal care, health and hygiene brands such as Pampers, Herbal Essences, Olay, Ariel and Gillette. With so many household names in its portfolio, P&G is no stranger to innovation – and today, data analysis forms a crucial part of its research and development process.
With a wealth of data already at its fingertips, the company turned to the National Innovation Centre for Data (NICD) to not only help create new predictive systems for capturing and processing consumer insights, but to better understand the data science skills needed to develop such systems in-house.
Data capture on a greater scale
“While some organisations might approach NICD because they don’t know what to do with their data, we’re very good at getting insights from ours,” explains Andy Patton, P&G’s Research and Development Fabric and Homecare Digital Innovation Sector leader. “Instead, we wanted to know how we could generate even more data to allow us to develop new product capabilities and upskill our employees in the process.”
P&G’s Newcastle-based research and development team is continually expanding, with nine new digital development resources adding to the workforce of statisticians and modelling simulations experts. Traditionally, the team had gathered feedback from consumers about its products by using ‘claimed data’, collected from sources like questionnaires. However, in a bid to create a more reliable, predictive system, P&G wanted to move towards ‘measured behaviour’ – learning about customers' habits through direct observation, which can offer more accurate predictions on what people think about products.
After struggling to find systems that enabled them to collect data effectively through direct observation, P&G decided to develop a machine learning platform in-house, but they needed a steer on how to bring it to life and equip employees with the skills to do so, which is where NICD came in.
We needed someone with the knowledge to drive the project forward in a collaborative way
A collaborative effort
Over the course of several months, NICD played the role of a “critical friend”, one that helped to challenge the feasibility of the machine learning platform P&G required.
“We were very clear what we wanted the project to focus on, but we also wanted to learn how we could develop these new capabilities in the best way possible,” says Andy. “We needed someone with the knowledge to drive the project forward in a collaborative way. NICD helped us work through the challenges to meet our objective to gather more data.”
Thanks to NICD’s guidance, P&G has now advanced their knowledge of the measured behaviour system to use machine learning to make informed decisions about what consumers think about products, which in turn can be used to improve them. The company has also had a crash course in how best to develop similar systems in the future and the skills required to do so.
The relationships between NICD and P&G is ongoing, with plans in the pipeline for the two organisations to collaborate on a cloud computing assignment. “This project has changed the way we think about approaching the development of new data models,” concludes Andy. “What we’ve learned from NICD is that our principle-based approach is a welcome step up and we’re now applying that knowledge to a number of projects.”