The PACCAR CRM Accelerator

Client: PACCAR Parts

In order to increase redemptions, it is crucial to understand the preferences of loyalty members and to demonstrate that we know them. 

 We employed AI/Machine Learning approach to understand the redemption behavior of members and uncover hidden trends, not easily discernable using standard techniques. Using Python-based Collaborative Filtering algorithm, we are able to identify the products, members are most likely to purchase, driving up the redemptions. 

 Our machine learning workflow comprises of the following steps:

 •Data: Historical data for the past 8 years, consisting of each individual member’s redemptions

 •Model: Using collaborative filtering algorithm, generate a ML model which can identify hidden traits present in the input data

 •Recommendation: Based on the model generated in #2, make personalized recommendations for each individual member

 14 million redemption transactions from 412K members were analyzed at scale using this customized recommendation system to recommend what to feature at individual level across digital channels. The result allowed to create hyper-personalized creative. 

 As the global pandemic hit and supply chain issues arose it was crucial that we get even more personalized with our content and offer recommendation. We identified industry types for our existing members through RigDig, a third-party fleet data source. We identified an industry utility score to identify relevant offers per industry and inserted them into the recommendation engine to include them in the offer landing page depending on industry.