Predictive lead scoring model for a Fortune 100 global luxury automotive OEM

A Fortune 100 ranked global luxury automotives OEM sells and services to retail customers via dealers, franchisees to prioritize sales leads for targeted connect

Client had developed rule-based engine to calculate propensity of new sales leads to conversion but needed an ML-based solution.

TransOrg followed a phased approach to deliver the solution:

• Client’s sales funnel typically had the following key stages: Lead with/without demographics>>Enquiry assigned>>Test drive arranged>>test drive provided
• Collected & analyzed retail customers’ sales leads from a 1-yr period such as demographics, behavioral such as test drives, walk-ins, booking, purchase intent information
• Resolved data challenges such as inconsistency in mapping historical & demographic data at lead level, incomplete demographics such as age, industry, family status, designation, lead’s digital footprint & incorrect/outliers e.g., age>
• Identified significant variables such as response to campaigns, purchase time frame, lead freshness, car trade-in score, primary owned brand (luxury or non-luxury). Other variables such as gender, # of cars owned, expected purchase month, city, whether corporate were not found to be significant
• Different steps were carried out to prepare model which include:-Data collection, consolidation and cleaningSales funnel analysisFeature EngineeringModel building w.r.t different stages of sales funnel.Validation and refinementClassification based Machine Learning model was built to understand the
  dynamics of customer demographics and behavioural response to predict the
  conversion probability.

  Model validation and hyper tuning conducted when required.

• The predictive ML model has the best prediction among all the other models.
• Increment in mean precision rate of around 52% and mean recall rate of around 35%.
• Helped client to efficiently allocate resources, focus sales efforts on
  high-potential leads, and boost conversion rates, obtain valuable insights into customer behavior and preferences.