Data analytics
We are pioneering the use of data science and artificial intelligence (AI) across Croda, by exploring innovative ways to use this technology.
Beyond boundaries homeIgniting innovation for a sustainable future
Croda is pioneering the use of data analytics and artificial intelligence (AI) across its business - from internal operations, to helping customers achieve both commercial and sustainability goals. By exploring innovative new ways to use this technology, it has saved millions of pounds through targeted customer retention; reduced environmental impact by lowering the amount of land tomato farmers need to grow flourishing crops; and cut the time its scientists need to identify effective formulations for personal care products - from years to days. It is one of very few companies in this sector to investigate the potential applications of data analytics to this extent.
Below you will find a series of content as part of our ‘Beyond Boundaries’ campaign that explores how data and analytics is being used at Croda to provide positive environmental and commercial outcomes for the industries it serves.
Beyond boundaries: Data analytics
Croda x-rays every tomato seed and uses an AI model, trained on 10bn data points and utilising state of the art machine vision techniques to identify those that are unlikely to germinate normally (or at all) based on what the embryo inside looks like. Any that are identified as defective are withdrawn from seed batches sold to customers. The market standard for successful tomato seed germination is 85%. With this method, Croda customers achieve 95%. Not only is this a significant increase in the output from each investment customers make in their tomato seeds, but it reduces the amount of land they need to achieve these results by up to 10%.
Data analytics explained by our experts
Croda’s carbon footprinting application automates many of the elements of the manual process and it allows for testing of ingredients en masse. This has a huge impact on the speed at which they can be assessed — from one ingredient every few days, to more than 20,000 ingredients in just a few months. Accuracy also improves as a result.
Croda was previously limited to offering ‘site averages’ for the carbon footprint of its ingredients, it can now offer fair assessment for each one individually. The result is more informed customers, more accurate results, and a faster route to both.
Beyond boundaries podcast: Data and analytics
Croda’s team has built a machine learning model that can help it predict which customer contracts may be at risk. This innovative predictive model was trialed for a year in Western Europe for a single market segment. During this period, it correctly identified 70% of Croda’s ‘at risk’ customers and enabled the company to secure more than £2m of revenue by allowing the sales team to engage those customers with the highest probability of moving to an alternative supplier.