case study

Student Lead Scoring Model


Business Impacts

Optimized marketing spend

Identification of optimum students

Increased ROI

Customer Key Facts

  • Location : North America
  • Industry : Education Management

Problem Context

An American higher education provider wanted to optimize its marketing spend on students for their campus-based and online programs. Therefore, they wanted to transition from using a third-party owned lead scoring model to a custom, in-house model that would enable them to identify, evaluate, and take action on the students on the basis of the lead quality or likelihood to enroll in a course.



  • Ensuring data consistency and unification between the two datasets to be utilized for modeling
  • Feature engineering and evaluation of feature importance to ensure satisfactory modelling results
  • Re-engineering the client’s existing student binning mechanism logic to ensure the usability of model results
  • Ensuring ease in process flow replication since model deployment was done by the client themselves on a different platform

Technologies Used

Google's BigQuery

Google's BigQuery

Google Cloud Storage

Google Cloud Storage

Google Compute Engine

Google Compute Engine



Identifying Quality Leads for Maximized Resource Usage Efficiency and Optimized Marketing Spend


Quantiphi developed a Machine Learning-based, custom Lead Scoring Model that generates the probability scores of respective leads that enroll in either the campus-based or online programs. These predictive conversion probability scores are then used to categorize the leads into different Bins A-G, in decreasing order of their likelihood to convert. This provides the customer with insights from the underlying factors per lead that affect the likelihood of enrollment. Google's BigQuery helps to analyze, clean, transform, and process the data. Feature engineering was also performed on the final dataset in BigQuery. Model training was performed using different classification models to identify the best performing model using instance running on Google Compute Engine.


  • Improved lead quality
  • Strategic decision-making

Thank you for reaching out to us!

Our experts will be in touch with you shortly.

In the meantime, explore our insightful blogs and case studies.

Something went wrong!

Please try it again.