The Doubtful Value of Predicting Spot Rates for Bid TenderingPosted March 07, 2019 by
I’ve recently come across the press release of a major transport service provider touting their new feature of predicting – among other things – spot rates:
“Proprietary algorithms provide trend analysis of spot rates, driving conditions and other factors that impact bids.”
They continue to mention implementing machine learning techniques to succeed in this kind of analysis. Certainly, they have invested some time and effort in developing such a feature.
The critical question though is if there is any value in predicting spot rates in any sophisticated fashion to bid on spot work for execution 24-72 hours from now. How much better is the advanced model from the naïve default assumption that spot rates will remain constant over the very short term.
Looking at DAT data over the last two years it becomes apparent that spot rates don’t move all that much week on week. More than 90% of moves are below 5% in change. Furthermore, the most significant changes happen around unpredicted weather events (hurricanes, extreme temperatures, etc.).
Spot rates naturally move much over more extended periods as any market participant can tell from the experience of the last 24 months. Getting such long-run predictions right is immensely valuable, but it is only useful to RFx style contract procurements, not for the immediate execution in the spot market.
Another factor that diminished the value of predicting spot prices for the short term is the fact that they are far from being the most relevant factor impacting the bid of a carrier on a load. The most pertinent factor is his cost structure, and here particularly the non-revenue costs of deadhead miles. Whether a truck needs to drive 100 miles to pick up a load or is nearby is substantially more relevant for the bid price than any other factor.
As a result, bids differ by carrier quite substantially, with 10-15% differences between offers not uncommon. Identifying and leveraging this difference then is the critical item shippers, and transport service providers need to be good at: how do you design your tendering process in such a way that you reliably find the truck with the best economics and run a competitive and efficient enough process to participate in the better economics.
Check out TNX’s Smart Tendering feature for a cutting-edge implementation of such an approach.