Carrier Network Profiling - Part 1: The BasicsPosted January 28, 2020 by
Photo by Felipe Furtado on Unsplash
Understanding which carriers are relevant for a load has always been a hallmark of successful truck brokerage. With brokerages having now tens of thousands of carriers on their roster, maintaining useful and up-to-date carrier profiles is a challenge.
Many have resorted to asking their carrier representatives to keep lane preferences updated in a TMS, but the quality and recency of such data are spotty at best. Smaller and infrequent carrier relationships have even more pronounced issues in data quality where emerging trends between multiple carrier representatives can easily fall through the cracks.
A further limitation of carrier profiles maintained by brokerages by relying on their representatives is that those profiles are necessarily sparse in the number of features kept about a carrier. Most frequently, they are being held in the form of region or laneway preferences without much regard for differences in equipment, timing, or other potentially relevant features.
Smart Carrier Profiles
With TNX, we build and maintain carrier profiles using artificial intelligence via our Smart Carrier Profiles technology. It allows us to incorporate several different data sources and extract multi-dimensional carrier profiles.
Artificial intelligence allows for the inclusion of multiple data sources that have different characteristics between them. Some of the key data sets used in building carrier profiles are:
- Historical Load Data from the brokers TMS
- User Supplied Lane Preferences (if available)
- Carrier Equipment Data (from induction or setup forms, or other data sources)
- Multi-Channel Carrier Interaction
- Official Data (e.g., FMCSA records or similar government data)
Over the next few weeks, we will have an in-depth look at each of the data sources.
The most relevant data source missing is GPS data. There are two essential reasons for this: first, in a more spot transactional brokerage style setup, the availability of GPS data from each carrier is not guaranteed. Most of the time, it is still missing.
Second, and more importantly, the additional information provided by GPS data when load pickup and destination data is available (technically the Fisher information of the GPS data) is at most limited. Statistically, the difference between the actual route (as per GPS) and the planned or optimal route (from a route planning algorithm) is negligible.
Private View on the Carrier
A majority of the data that goes into the carrier profile is private data to a broker. As a result, the profile only covers part of the carrier’s book of work.
For example, when a carrier does a repeated job from Chicago to Dallas with the broker, the carrier needs to find a way to get back to the Chicago area eventually. However, how this happens may be out of the purview of the broker. It may be straight back with loads from a different shipper or broker, or it may be via a few other stops in an elaborate roundtrip.
In a future article, we will have a look at how TNX deals with such situations by, on occasion, offering the carrier back loads on the Dallas to Chicago laneway to figure out if the carrier profile needs to reflect that interest.
At TNX, we are excited about the progress we have made understanding carriers’ interests and preferences in a data-driven way. We are happy to share how we approach the challenge over the next few weeks.
If you are interested to see how Smart Carrier Profiles work for your carriers, and how we use the information learned here to power the TNX Smart Tendering approach doubling brokers gross margins, then please reach out to email@example.com.