By Luis Lopez, AI transportation consultant, CEO of Go Hub.io Holdings Corp and subsidiaries, and host of the Freight Guru Podcast
Most shippers already own the data they need to buy freight better. It sits in carrier invoices, bills of lading, order systems, and spreadsheets that nobody has lined up side by side. Freight lane analysis is the work of turning those records into a clear picture of where freight moves, how often, at what cost, and with what service. It is the foundation for every good rate negotiation, RFP, and network decision, and it is also the first thing any AI tool will need before it can help.
As an AI transportation consultant, the first thing I look at is not the rate. It is whether the shipper can describe its own lanes accurately. This article walks through how to do that.
What a Lane Actually Is
A lane is a repeatable origin-destination pair for a given mode and equipment type. “Atlanta to Dallas, dry van, truckload” is a lane. “Atlanta to Texas” is not precise enough to price, because a carrier running into Dallas sees a very different market than one running into El Paso.
Decide on a consistent lane definition before you analyze anything. Common choices include:
- City or facility pairs for truckload freight with fixed shipping and receiving points.
- Three-digit ZIP pairs when destinations vary but cluster geographically.
- State or region pairs for a high-level view, useful for budgeting but too coarse for pricing.
Keep mode and equipment in the definition. Mixing LTL shipments with truckload moves on the same lane produces averages that describe nothing real.
Gathering and Cleaning the Data
Pull at least twelve months of shipments so seasonality shows up. For each shipment you want origin, destination, ship date, delivery date, mode, equipment, weight, pieces or pallets, carrier, linehaul cost, fuel surcharge, and accessorials as separate fields.
Expect cleanup. Typical problems include:
- The same facility spelled five different ways.
- Fuel surcharge rolled into linehaul on some invoices and broken out on others.
- Accessorials billed on a separate invoice that never got matched to the shipment.
- Missing delivery dates, which make on-time measurement impossible.
- Credits and rebills counted as extra shipments.
A short guide to the fields that matter most is in our piece on the data fields to clean before buying AI. The same cleanup pays off whether or not you ever buy a tool.
The Core Metrics to Read
Volume and Consistency
Count shipments per lane per week and per month. A lane with steady weekly volume is attractive to carriers because they can plan trucks around it. A lane with the same annual total but long gaps and sudden spikes is harder to service and will usually price higher. Consistency is something you can sell in a negotiation, so measure it.
Cost Per Shipment, Per Mile, and Per Unit
Look at total landed freight cost, not just linehaul. For truckload, cost per mile helps compare lanes of different lengths. For LTL, cost per hundredweight or per pallet is more meaningful. For the business, cost per unit shipped or as a percentage of order value is what leadership cares about.
Accessorial Share
Calculate what portion of spend on each lane comes from accessorials such as detention, liftgate, redelivery, or reweigh. A lane with a reasonable base rate and a high accessorial share usually points to a facility problem, a data problem, or a booking habit, not a pricing problem.
Service Performance
Measure on-time pickup, on-time delivery, transit time against plan, tender acceptance if you have it, and claims frequency. Cost without service is half the picture. A cheaper carrier that misses appointments may cost more once you count chargebacks and expedites.
Direction and Balance
Note whether your lanes run into markets where carriers struggle to find a return load. Freight heading into areas with little outbound volume often prices higher because the carrier may run empty afterward. Our explainer on backhaul vs. headhaul lane balance covers why this matters.
Questions the Data Should Answer
Once the metrics are in place, read them with specific questions in mind:
- Which lanes make up most of the spend? Usually a small group of lanes carries a large share of cost.
- Which lanes have stable, predictable volume suitable for contract pricing?
- Which lanes are irregular and better left to the spot market or a flexible arrangement?
- Where are accessorials or claims out of line with the rest of the network?
- Where is the same lane served by many carriers with no clear reason?
- Are there shipments that could be consolidated, shifted between LTL and truckload, or moved on different days?
The answers shape strategy. Stable lanes belong in a structured bid. Irregular lanes may be better priced with flexibility, a tradeoff discussed in our guide to contract vs. spot freight rates.
Common Mistakes When Reading Lane Data
- Trusting averages. An average cost per mile can hide a few very expensive moves. Look at the distribution, not just the mean.
- Ignoring seasonality. A lane that looks expensive in one quarter may simply have been priced in a tight market.
- Comparing unlike freight. A shipment needing a liftgate and an appointment is not comparable to a drop-trailer move on the same lane.
- Leaving out failed tenders. If the primary carrier rejected a load and a backup took it at a higher price, both facts belong in the record.
- Analyzing once and stopping. Lanes shift as customers, suppliers, and facilities change. Refresh the analysis on a regular schedule.
Where AI Fits
AI tools can help with the slow parts of lane analysis. They can standardize messy location names, match accessorial invoices to shipments, extract data from scanned documents, flag outliers, and summarize patterns in plain language. Forecasting models can project volume by lane when there is enough history.
What AI cannot do is fix data that was never captured. If delivery dates were never recorded, no model can measure on-time performance. If fuel was blended into linehaul, the tool can only guess at the split. Treat AI as a way to speed up a disciplined process, not as a replacement for one. Check its output against a sample of source documents before acting on it.
Turning Analysis into Action
Good lane analysis ends with decisions. Typical outcomes include a short list of lanes to bid, a set of facility issues to fix with internal teams, a carrier mix that is simpler and better matched to each lane, and a baseline to measure next year against. If you plan to go to market, our shipper guide to running a freight RFP picks up where this analysis leaves off.
Shippers that know their lanes negotiate from facts. Carriers notice the difference, and it usually shows up in better pricing and better service.
For more conversations on freight and AI, subscribe to the Freight Guru Podcast.
About the author: Luis Lopez is a Miami-based AI transportation consultant and logistics entrepreneur, the CEO of Go Hub.io Holdings Corp and subsidiaries, and host of the Freight Guru Podcast.
