Carrier Scoring/Matching Framework Led to Profit Margin Increase by 1.5% for a Leading 3PL Company

Business Objective

Our client is a leading 3PL trucking and logistics company. The client had a rule-based carrier selection and matching system which was complex, manual, and lacked a seamless carrier search experience. It was not a “One-Stop Platform” for carrier selection.

Hence the client wanted to build a Digital Freight Matching Platform that would:
• Craft a “One-Stop-Shop” profile for all carriers
• Provide an efficient carrier scoring and matching solution
• Identify an improved pool of carriers to choose from

The matched carrier should have the best balance between cost, service levels, and the likelihood of acceptance.

  • Difficulty in capturing various personal attributes from “Agent – Carrier phone call data”
  • Non-availability of truck level information leading to better acceptance scores
  • Identifying attributes which capture the profile of the perfect carrier
  • Integrating data from various data sources

Solution Methodology
  • Prepared and filtered data which included Zipcode cleaning, Market ID mapping, Removing NULL Origin and Destination Markets, Equipment mapping, Distance outlier treatment, RPM and Price guard rails, and Data assessment. Selected features for the input to the model.
  • Beyond basic cleaning and processing, performed posting/searches data processing, which included:
    • Assigning higher weightage to current day postings and lane postings
    • Performing EDA for postings, searches, distance and service time, carrier preferred lanes to identify useful insights
  • Scaled data to eliminate effects of variation in values. Computed and ranked Euclidean Distance between the ideal carrier and probable carriers
  • Based on the input features, the model computed Euclidean Distance and assigned ranks to suggest carriers
Business Impact
  • Increased profit margin by 1.5% by bringing in high propensity carriers who can carry the load at a competitive price

  • Crafted a near-perfect carrier profile from the learnings leveraged from past carrier behavior, lane history, postings, searches, and market situation

  • Added 60K qualified new carriers to the network in the first year

  • Observed relevance/lift in rankings of the carriers which accepted the load based on new solution vs old rule-based methods

  • Load booking conversion increased by ~10%

  • 15% load booked for carriers are identified as round trips (within or outside of client network) resulting in empty miles reduction

Copyright © 2023 Tiger Analytics | All Rights Reserved