Chicago based 3PL

How to Use Data Science to Increase 3PL Quote Win Rate.

An Overview

The Challenge

A Chicago-based 3PL wanted a way to quickly analyze why quotes were won or lost.

The Solution

The 2nd Watch data science team built a machine learning solution to determine which factors most impacted a quote’s success to easily analyze new quotes based on these factors.

The Outcome

The client now can adjust their quotes to increase their win rate. Based on their current conditions, even a slight increase can result in millions in revenue.



The math is simple: The more quotes a 3PL wins, the more money they make. The ability to win those quotes relies on a number of factors beyond just the quoted cost. In order for a quote to win, factors like these need to be taken into account:


  • The time between when a quote is created to when it expires (business hours)
  • Quoted cost
  • Total quotes for a customer since 2016
  • Total load miles for a customer since 2016
  • Total number of loads for a customer since 2016
  • Total load weight for a customer since 2016
  • Average market line haul rate
  • Average market price aggressiveness
  • Average market number of reports
  • Average market number of companies
  • Month
  • DnB annual sales categories
  • Day of the week

What if you could use data to determine which of these factors impacted the win rate the most of which have little to no effect on the win rate? By understanding these factors, a 3PL could make more informed decisions on quote creation leading them to an increase in their win rate.

When a Chicago-based 3PL came to us with this challenge, 2nd Watch brought in our data science team to explore how to use data science to increase the 3PL quote win rate (and profits).


The Goal

The goal of this project was to analyze our 3PL client’s internal and external data to understand which factors had the greatest impact on a quote being won or lost. We wanted to help our client leverage this information moving forward by predicting the win rate of current quotes. Ultimately, the client wanted a solution to optimize quotes to increase the 3PL quote win rate and profits.


The Solution

The complexity and amount of data that needed to be analyzed required a solution beyond standard analytical capabilities. The 2nd Watch data science team built a high-performing, computationally inexpensive random forest model to enable predictive analytics capabilities. This 3PL data science model was built with data from an existing internal Snowflake data warehouse as well as external industry data for our client’s customers.


“Improving the win rate relies on a 3PL’s ability to identify and analyze the top internal and external factors that influence which quotes are won and which are lost. 2nd Watch built a machine learning model to enable our client to quickly combine and analyze these factors to measure their influence on win rate.”


The Outcome

This solution identified six significant factors that greatly impact the likelihood that a quote would be won. Armed with this, our client’s team can focus on those six factors when creating quotes to increase the likelihood of a win. Factors that had a moderate or slight impact were also identified to give the team a more thorough understanding of the impact.

In addition to enabling the team to optimize quotes with this information, the 3PL predictive analytics solution can be used moving forward to predict the likelihood that a quote will be won. By having this information, our client can adjust their quotes to increase their win rate, where even a slight increase can turn around a massive return on investment.


The Bottom Line

In the initial stages of this project, 2nd Watch’s analysis found that even a conservative 1% increase in win rate had the potential to increase profits by $9 million for our client.