Portrait of a mechanic at work in his garage - car service, repair, maintenance and people concept.

Pricing prediction through machine learning and predictive analytics at a leading UK automotive repair organisation


The client offers vehicle maintenance and servicing through its Autocentre business. Like all vehicle service providers, the client is unable to present a customer with a fixed service price upfront, relying instead on a price estimate. On completion of services a customer may pay more than originally estimated resulting in a negative customer experience. Providing a customer with an estimate decreases transparency for the customer and the ability for Autocentre managers to forecast revenue accurately.


The analytics practice at Valcon designed and delivered a multi-algorithm solution that predicts the most suitable fixed price that a customer will pay upfront via the application of statistical, predictive and Machine Learning models. The solution calculates a ‘price corridor’ through statistical methods, the safe range in which to quote a price based on actual historical prices paid for services for a specific vehicle make and model. The solution predicts the likely workload for a given Autocentre on a given day leveraging Machine Learning. How busy an Autocentre is likely to determine where a final price will fall within the ‘price corridor’. The various pricing metrics for a given vehicle and service are presented for consumption via an API to the client’s consumer-facing website, supporting 3m bookings per year on average or 6m predictions. The solution is delivered on the Microsoft Azure cloud using the best-of-breed services available and utilising modern approaches to operationalise and productise analytical solutions from DevOps to proactive ML model management using Valcon’s proven architectural patterns and methodologies.


The client enjoys greater sales conversion due to transparent and predictable pricing, increased customer retention and accurate revenue forecasting at an Autocentre, regional and national level.

Case Studies