Earlier in August (2017), I was invited to attend an interview for the Manager Dealers Commissions & BI position at a world-class company that I would like to keep anonymity. Given the venue was largely far from my current residence, I couldn’t make it on-site. Apart from the distance, at that time, I had a hectic schedule (very difficult to explain). Meanwhile, a very mundane idea crossed my head to propose to be interviewed online instead. Fortunately, the recruiter appreciated it, and promptly granted me the same.
Regarding the position above, the company was looking for someone who has a genuine interest in working for them, and bears data analytics skill-set and experience with handling both incentive and commission management data. I was eligible for this position. Based on my work experience, I am familiar with building business dashboards from large-scale data, and developing CVM (Customer Value Management) strategies for customer retention. By the way, I have created an R package (cmsr) for calculating commissions of saleforce (super dealer, super vendor, etc.): inspired by weDo technology.
Throughout the interview, only 10% of the conversation has emphasized on my capability in commission management, whereas the remaining part upon big data challenges. I was asked to propose an effective approach to overcome churn issues. In the simplest way, I answered as follows: you should combine SQL with big data and machine learning (ML) not for tracking customers in the group but individually (isolation). Each customer likely behaves different each other, even if marketing segmentation policy could arrange them into the same strata. In an isolated-based ML analysis, we can capture from historical data the most flagship products that a single customer likes or has used in a given period. From this insight, we can thoroughly search for times that the customer likes to carry out his/her operations. In this sense, we’ll easily determine what make customers happy and unhappy. ML techniques that I evoked at the interview were Logistic Regression (LR) and Random Forest (RF). At this level, CVM should be able to initiate a well-focused campaign.
I successfully got the position but the starting date does not allow me to get over there.