Where to Find the CSV
In the carrier portal, navigate to Financials > Settlements. Each settlement period has a downloadable CSV. Download the CSV, not the PDF — the PDF is formatted for reading, not analysis. The CSV gives you raw row-level data that you can load into a spreadsheet, accounting software, or a database.
Amazon also lets you export load history separately from the Loads section of the portal. This gives you operational data (origin, destination, distance, driver, completion time) that the settlement CSV doesn't include. For full analysis, you want both: the settlement CSV for payment data and the load history export for operational data, joined on Load ID.
Key Columns in the Settlement CSV
The exact column names can vary slightly by export version, but the standard fields are:
- Load ID — Amazon's unique identifier for each load. This is your primary key for matching settlement records to operational records.
- Load Type — distinguishes trips from block legs. Important if you're analyzing performance by load type.
- Pickup Date / Delivery Date — actual completion dates, not scheduled dates. Use these for period analysis.
- Origin Facility / Destination Facility — Amazon facility codes, not addresses. You'll need to maintain your own facility code-to-address mapping if you want lane analysis by geography.
- Base Rate — the rate shown at booking. This is the number you compare against your cost per load.
- Adjustments — any detention pay, bonuses, or deductions applied after the base rate. Can be positive or negative.
- Net Amount — base rate plus adjustments. This is what you actually get paid for the load.
- Payment Status — shows whether the load has been paid or is pending
What's Missing From the CSV
The settlement CSV doesn't include distance (miles), driver name, or per-mile calculations. To get rate per mile, you need the load distance, which comes from the load history export or from your own ELD/dispatch system. Matching settlement records to your internal records by Load ID is the standard approach.
The CSV also doesn't break out the fuel surcharge component of the rate, because on Relay the FSC is embedded in the posted rate. This means direct apples-to-apples comparison with broker loads (which typically show base rate and FSC separately) requires some manual calculation — add the FSC from a comparable broker load to its base rate before comparing.
Building Useful Metrics From the Data
Once you have the settlement data in a spreadsheet, the calculations that matter most:
- Revenue per mile by lane — net amount divided by loaded miles. Group by origin-destination pair. This tells you which lanes are profitable and which are marginal after fuel.
- Revenue per driver per week — sum net amounts by driver (requires matching driver to load ID) and divide by weeks active. Benchmark across your driver team to identify utilization gaps.
- Average adjustment amount — are adjustments consistently positive (detention pay captured) or negative (deductions)? Negative patterns warrant investigation.
- Load count by period — track week-over-week load volume to spot availability trends in your lanes before they affect your revenue
Reconciliation Against Settlement Payments
Each weekly ACH payment should match the sum of net amounts in the corresponding settlement period. Pull the CSV for a settlement period, sum the Net Amount column, and compare to what hit your bank account. They should match exactly minus any amounts that rolled to a subsequent payment period due to timing.
If the sum doesn't reconcile, sort the CSV by Payment Status to identify any loads still in pending status. Pending loads won't be in the current ACH but will appear in a future payment. If loads show as paid in the CSV but the sum still doesn't match the ACH amount, contact Relay carrier support with the specific settlement period and the discrepancy amount. Bring the CSV — they'll ask for the Load IDs in question.
Automating the Analysis
If you're running significant volume on Relay, manually downloading and analyzing CSVs every week doesn't scale. Options range from building a simple Google Sheets import with scheduled refresh to using a third-party fleet analytics tool that ingests Relay export data directly. At minimum, build a spreadsheet template with your key formulas already set up so each week's CSV drop-in requires minimal manual work. The data's only valuable if you're actually looking at it.