PartsTrader operates online marketplaces in New Zealand and the USA for automotive collision repairers to search and compare options and then buy the right parts, at the right price, at the right time from new OEM and alternative parts suppliers.
Each month, PartsTrader manages over 140,000 parts request tenders for over four million parts and more than 1.15 million quotes from suppliers. The marketplaces transact over $1.5 billion annually.
PartsTrader’s customers rely on them for accurate quotes for car parts and when requesting a quote from suppliers, repairers want the confidence they are getting the right part at the best price. PartsTrader wanted to enhance the competitive tender process with a fast, efficient way to predict the best quote price for their customers.
Data Insight assessed the situation and proposed relevant statistical and AI techniques to predict the best quote for new requests. By testing multiple techniques, Data Insight was able to find a simple, fast approach that achieved similar predictive quality to that of more complex and expensive machine learning methods.
Our Data Scientists found an efficient approach that could be easily refreshed across all parts and conditions each night with new transactions from the marketplace. By identifying car parts with similar changes in value over time, Data Insight was able to propose ways to harness more data for each prediction. This provided PartsTrader with a roadmap of approaches that would use statistical techniques, improved by machine learning, for future iterations of this model.
Throughout the project, Data Insight demonstrated a deep understanding of the PartsTrader business, allowing for the rapid development of an effective pricing model. The pricing model developed by the Data Insight team provides a quick and efficient way for PartsTrader to derive the most cost-effective prices for parts in their marketplace. On top of this, the model allows PartsTrader to identify and exclude outlier data, i.e. extreme prices, which historically has caused a number of issues for their customers.
Talk to us about how we can help you.
Get in touch: firstname.lastname@example.org