Image: Damaged fragile parcel delivered to doorstep by Meanwell Packaging, licensed under BY 2.0.
Your parcel invoice carries a residential delivery charge on an address your sales team calls a distribution centre. Multiply it across a few thousand packages a month and the line is worth real money, so an analyst batches up three hundred of them, sends a dispute, gets a one paragraph refusal, and the file quietly closes.
That dispute did not fail on whether the address is a warehouse. It failed because the only thing attached to it was an inference about the address string, and a carrier has no obligation to answer an inference.
The definition belongs to the carrier, and it is narrower than you think
FedEx's help page for address book entries defines a residential address as a home or private residence, including locations where a business is operated from the home. Read the second clause twice. That your consignee runs a business at the address is not a rebuttal. It sits inside the definition.
What gets classified is the delivery point. Not the occupant's tax registration, not the signage over the door, not whether a forklift meets the driver. An argument built on any of those misses the object, which is why it bounces.
Timing makes the error expensive right now. Supply Chain Dive reports that FedEx's 2026 peak demand surcharge on Ground and Home Delivery residential packages runs from 50 cents to 80 cents per package between October 26, 2026 and January 17, 2027, peaking from November 23 to December 27. For enterprise shippers moving more than 20,000 residential and Ground Economy packages, the same reporting puts the Demand Residential Delivery Charge between 1.70 dollars and 9.35 dollars per package. A misclassified ship-to that cost a flat surcharge in March costs a multiple of it in December, and the window opens in two days.
Four things get called evidence, one of them is a record
Sort what your team attaches to a residential dispute by what the carrier can do with it.
| What goes in the file | What the carrier does with it | Grade |
|---|---|---|
| A map pin, a company website, a photo of the building | Nothing. None of them is a delivery type record, and none is dated to your ship date | C |
| A pattern in the address: a suite number, a street named for an industrial park, a ZIP you believe is commercial | Nothing. Text is an inference about a delivery point, not a record of one | C |
| The carrier's own address classification, pulled with a timestamp near the ship date | Recognises it, because it is theirs, subject to the disclaimer below | B |
| A USPS RDI business result for the delivery point, at the vintage covering your ship date | Reads an independent record from the organisation that serves that point daily | A |
The gap between row two and row four is the whole problem. Both feel like the same assertion to whoever writes the dispute. Only one is a record somebody else keeps and will stand behind.
What RDI gives you, and what it costs to use properly
The USPS Residential Delivery Indicator is published on PostalPro as a data product that verifies whether a delivery type is classified as residential or business. Addresses processed against it return, in the Postal Service's own phrasing, Yes for Business or No for Residential.
Three details in the product description decide whether you can use it in a dispute at all.
- It ships as hash tables rather than a flat file, keyed on 9-digit ZIP+4 codes and 11-digit delivery point codes. You cannot query it with a street address, so your ship-to has to resolve to a delivery point first.
- USPS says it should be used in conjunction with CASS Certified ZIP+4 address matching software. An address your master data never standardised produces no lookup key and therefore no answer.
- Updates arrive monthly, and PostalPro lists an annual licence fee of 600 dollars. The file you query in January is not the file that described the address in October, so cite the vintage covering the ship date.
One more property shapes how hard you can push. The tables enumerate business deliveries: ZIP+4 codes containing only business deliveries, plus the delivery point codes for business deliveries inside mixed ZIP+4s. A business result is an affirmative entry in a record. A residential result is the absence of one, and a default does not win an argument.
RDI also binds nobody. It describes how the Postal Service classifies a delivery point, not how FedEx or UPS rates one. Strong independent evidence at the price, and still evidence rather than a ruling.
The carrier's own record arrives with a disclaimer attached
FedEx publishes its classification logic in the Address Validation API documentation on its developer portal, and the detail there beats anything a vendor blog will tell you. The documentation states that the API has only four possible classifications for addresses: unknown, business, residential and mixed. Every address begins unknown and stays there until validation business rules move it. Mixed is reserved for a multi-tenant address holding both business and residential units, so it is a real answer about a real building rather than a system failure.
The derivation matters too. FedEx says the classification for a functional address is calculated independently of the address validation process, and is based on feedback by operational personnel, with commercial data sources used for confirmation only. The driver who serves the street is upstream of the database.
Then comes the legal disclaimer, which audit teams need to read before building a programme on this endpoint. FedEx states that the result reflects its current address-level business and residential classification, that it is subject to change, that FedEx does not guarantee its present or future accuracy, and that FedEx does not guarantee that packages shipped to these addresses will be invoiced under the provided classification. The documentation also marks the data proprietary and confidential, not for sharing with third parties without written permission.
So the carrier's own answer is strong, and it is fenced. Quote it back in a dispute with that carrier, with the date you pulled it. Do not build a public claim on it, and do not treat a business result as a guarantee the invoice was wrong.
Some of these charges are yours, not theirs
Before any of this, check what you told the carrier. A residential charge also applies where the shipper designated the address as a residence, so a flag set in your own order entry produces a surcharge no record will overturn. The carrier billed what your tender said.
This is the same failure described in a dimension you never recorded. The argument is lost at tender time and discovered at invoice time, and by then the only fix is upstream in the address master.
How Penny grades a residential line
Penny assigns a claim grade before she drafts anything, because a refused dispute is not free. It burns goodwill and trains the carrier to skim your letters.
Grade A is a delivery type record from the right vintage returning business, with a tender that did not flag the address residential. She drafts those, naming the record and the date. Grade B has the carrier's own classification and nothing independent, so she files it with the timestamp attached and expects a conversation rather than a credit. Grade C has only a pattern in the address string, and she marks it needs-data with the missing field named rather than sending it, which is the behaviour set out in an agent that says needs data.
She also rolls the grades up by consignee. Forty residential charges on one ship-to across a quarter is not forty disputes. It is one address record to correct, and the dispute only recovers what already shipped.
What to do next
Pull one month of parcel invoices and isolate every line carrying a residential delivery charge. Group them by ship-to, then answer one question per group: what record would you attach, and what date does it carry. Count the groups where the honest answer is nothing. That count explains every refusal you have had this year.
Bring that month to a walkthrough and we will run Penny across it live, grading the lines and separating the ones worth filing from the ones your own tender created.