This company (CarFinance247) receives leads from google who complete an application for finance on their website. These leads were allocated to sales agents and worked randomly, there were too many leads and not enough staff to work all of them. In addition, there are low conversion rates and so it was difficult for Google to optimise ad spend on the clients behalf.

Task:

Allocate leads to the sales floor automatically so that the lead generation team did not have to call them first

Action(s):

• Investigate customer application and credit check data, present options for exploring further.
• Modelled the lead flow through the business to assign a functial value to leads based on their liklihood of getting a loan and at what APR% and hence commission
• Demonstrated the performance with a PoC and gained sign off from the client
• Built a service for hosting the model(s) on Azure Machine Learning Studio
• Built a model retraining pipeline
• Setup ongoing monitoring of th elive service & Blue/Green deployment
• Monitored the model release and improved the reporting as requested by the client
• Identified an additional use case of a lead valuation service - optimising google ad spend through value based bidding.

Result:

• Improved the number of leads being allocated to the sales floor by 30%, performing the same work as a lead generation team of 10 people
• Awating the results of the Value Based Bidding process