Rate Insights provides users with load-level rating by looking at market conditions, specific load details, and accurate geographic information to predict rates. Traditional rate products took a blunt view of math and the factors that go into rating loads. In short, given an equipment type and an origin and destination, historical rating tools would simply spit out the average rates for that combination of three factors. Rate Insights leverages predictive models informed by booked and posted loads with from data across Truckstop’s leading digital marketplace and products (more).
Rate Insights uses non-linear machine learning methods to create a mix of boosted gradient models for point prediction, upper bound prediction, and lower bound prediction using over 50 variables about a load and the marketplace to make a prediction. The more descriptive a load is, the more accurate the prediction will be.
The machine learning models that power Rate Insight updates multiple times a day to account for sudden changes in the marketplace. It uses temporal techniques to factor in seasonality and places a larger emphasis on recent marketplace behavior.