SONAR is the central hub for freight market analytics & forecasts.
What are the data sources?
SONAR aggregates data from over 1,000 sources in transportation and logistics. These sources range from IoT vendors, transportation management systems (TMS), telematics providers, on-truck sensors, payment providers, fuel vendors, third-party logistics providers (3PLs), shippers and carriers.
How much of the market does SONAR see?
It depends on the type of data; not only is each data set statistically significant, most samples are massive in size.
FreightWaves estimates that we see 85% of the electronic tenders in the market. For asset-tracking such as wait times, SONAR tracks approximately 800,000 assets, which include trucks, trailers and chassis. SONAR’s maritime data tracks roughly one-third of all container bills of lading (BOLs) around the world. In diesel volumes, SONAR sees about 80% of the volume of diesel gallons purchased at the rack.
SONAR benchmarking provides insight into carrier financial and operational performance data from over 220 fleets, representing around 60,000 trucks.
Across the board, SONAR offers the largest and most representative samples in the market.
What do you mean by the world’s fastest freight Forecasting dashboard?
Much of the data in SONAR is based on what has happened in the market within the past 24 hours. This perspective gives SONAR its near-time view of market conditions from data sets generated by executed TMS and telematics activity.
How did SONAR become so smart?
SONAR is more than just data, it’s intelligence. The intuitive SONAR 5.0 platform allows users to quickly understand market conditions. SONAR users also have the expertise of the FreightWaves Customer Success and Market Expert teams at their disposal.
The FreightWaves team of data scientists and analysts consume the data and create standardized methodologies using deep-learning models derived from the data. Often the data received is unstructured and “dirty,” requiring cleaning and filtering by these experts.
Once the data is cleansed and ready to be backtested against like data, it is published into a controlled environment where a computer simulates the data against other models. If the data passes authenticity tests, it is then put into a QA production environment.
Once it reaches QA, the market analytics team will use it to test in a real-world environment to answer questions. This can take months. If there are errors or issues, the methodologies are tweaked. If the answers are substantiated by backtesting and analysis, then the data gets added to SONAR for client use.
How does SONAR create freight rate forecasts?
The FreightWaves team has created deep-learning models to track how rates move in the freight market. We call this the Freight Market Waterfall Theory and built our forecast based on scientific principles of the Freight Market Waterfall Theory.
The Freight Market Waterfall Theory is outlined below:
Freight rates are dictated by routing guide compliance. A shipper that achieves nearly 100% of routing guide compliance will continue to optimize its spending by taking advantage of lower-priced carriers. If a shipper sees compliance in its routing guide break down, it will be forced to buy capacity in the spot market, often at higher rates.
If a market is decelerating (tender rejections are falling), spot rates are likely to fall until they hit the market floor.
The market floor is equivalent to the collective operating cost of carriers. Rates are unlikely to fall below this point.
If a market is flat (tender rejections are near zero), in the short term there will be continued downward pressure on spot rates until rates hit the market floor.
If a flat market continues for more than a few months, contract rates are likely to fall towards the market floor.
If a market is strengthening (tender rejections are increasing), there will be upward pressure on rates. There is no ceiling on rates, but if rates stay high for an extended period of time, new capacity will enter the market.
SONAR tracks routing guide compliance by tracking tender rejections and other data to determine if routing guides are likely to break down in the near future. This data is then compiled and compared against other data sets, including financial and operational data from hundreds of carriers, brokers and shippers.
Using data derived from hundreds of operating and financial metrics of over 200 carriers, we calculated the average operating cost of carriers across the market. We then backed this into the “base rate” across the market. This is basically the cost it takes to operate a truck in the market. This is considered the base rate.
Then we built an algorithm that multiplies the base rate against the tender rejection data to get the current market condition rate, using historical market spot rates.
The market condition rate is then trained against the HAUL index to allow for individual market conditions. If a market has a negative HAUL value, it’s considered a backhaul market and rates can follow below the base rate. If a market has a positive HAUL value, it’s considered a headhaul market.
Keep in mind that the rate for any given lane is determined by both the origin and destination; the price of a truck has as much to do with the attractiveness of certain destination markets as it does with the availability of trucks in a given origin market.
Adjusting the base rate by conditions in those markets gives the team a scientific model of “today’s rate” by origin/destination pair.
Today’s rate is then forecasted out a year by looking at the historical rates and future direction of the market, using SONAR data from thousands of sources. The rates adjust for seasonality variations in the model and other financial and operational components.
The model also becomes smarter over time as more data is fed into it. Significantly, our models for each market interact with each other, so a surge in volume in Atlanta, for example, will affect how we think about the availability of capacity out of Macon.