One area where prediction is now widely used is in pricing, especially in business-to-consumer (B2C) customer segments, where it is applied to improve margins and expand market share.
Companies that initially experimented with predictive pricing often leveraged it for customer centricity efforts, with customer personas becoming a key driver for price setting. This escalated the deployment of predictive analytics and machine learning models at a commoditized cost to capture supply and demand trends from a customer perspective and drive sales.
Today, B2C companies capture large volumes of customer, market and product data and its associated drivers and use this information to set price, plan supply strategy and shape demand forecasts. As just one example, Amazon has built and deployed dynamic (predictive) pricing models that make pricing and supply-demand planning both seamless and real time.2
Fortunately, chemical companies can mirror the B2C companies that are accessing on-demand computing and application services in the cloud and on-premise for data management, machine learning, artificial intelligence, predictive analytics and visualization. Furthermore, vendors are integrating these services into enterprise ecosystems, which makes it even easier for chemical companies to kickstart, test and extend predictive pricing programs for specific business use cases—from sandbox to industrial scale.
Chemical companies join predictive pricing trend
More recently, the need for predictive pricing in chemicals—which is still mainly a business-to-business (B2B) industry—has emerged due to four converging factors. The first is fierce competition in base and performance chemicals markets, including new customers, global competitors, cost-effective feedstocks and a dynamic supply chain. The second is the shift to customer centricity, which is shaping chemical product and end-use application development. Third is the growth in e-commerce channels designed to serve customers with an order-to-door experience. And last is the increasing business prioritization to manage and maximize margin.
Given these demands as well as the growing use of real-time predictions to make decisions in sectors such as financial, aviation and logistics, predictive pricing can play a central role for doing business in the chemical industry.
Steps to forge ahead
Sales, marketing and planning teams of chemical suppliers of all types can undertake a digital transformation journey using predictive pricing. Case in point: Accenture helped one global supplier capture and transform market pricing and driver data from sales personnel in the field into margin improvement. The goal was also to improve the company’s existing operational structure such as sales, supply and planning processes within global regions.
Accenture developed a machine learning/artificial intelligence model to predict pricing of a high-margin base chemical in a dynamic and highly competitive region. Key steps included:
- Capture pricing drivers (e.g., global, macro, strategic, supply-demand, feedstock)
- Identify dynamics of drivers
- Quantify and validate driver relevance
- Select drivers for building the model
- Build, train, predict and back-test price up to a six-month time horizon
Working together, we helped the company deliver a predictive pricing model with an accuracy of more than 80 percent in comparison to industry price benchmarks and validated it with regional market prices. The model was adopted to ultimately achieve the following key benefits: