Small-business outsourcing research
What share of damaged-order claims include a usable photo and order reference?
A protocol for measuring claim evidence quality, with clear tests for a usable photo, a linked order, and a recorded delivery condition.

Research question and scope
This protocol asks what share of damaged-order claims include a usable photo and a reference that links the damage to an order. The unit is one claim, identified by the order number and the date the claim was opened. The study measures whether the claim has a readable order label, a clear image of the damage, and a recorded delivery condition. It does not evaluate whether the customer deserved a refund or replacement.
Set the scope to one small business, one order system, and one claims queue. Record the channels that can receive a claim and the people who may approve one. If the photo requirements change during the window, keep the earlier definition for the full sample or analyze the change as a separate period. A new intake form can look like better evidence when it is only a new field.
Population and unit of analysis
Define the population as every damaged-order claim opened during the window, including claims that were denied, withdrawn, or closed without a payment. The unit is the claim, not the customer, because one customer can claim several orders. A claim that is reopened remains one unit with a reopen date. A second claim for the same order is a separate unit linked to the first.
Record the order value, the carrier, the product category, and whether the customer or the team supplied the photos. These strata matter because a low-value claim and a high-value claim carry different evidence needs. Include claims that arrived through a phone call rather than the web form, because the base rate is part of the finding. Test the unit definition on two ordinary claims before collection begins.
A customer can send a photo of a damaged item without the packaging and a second photo of the packaging without the item. Record the set as one claim and note which views are missing. Do not treat each photo as a separate unit, because that would reward the customer who sends many similar images. The unit is the claim, and the evidence is a property of the claim.
Definitions that make the measure reproducible
Define a usable photo as one in which the damage is visible and the image is in focus at normal viewing size. Define a linked order as a claim that shows an order number that exists in the order system and matches the customer. Define a recorded delivery condition as a note about how the package was received, including who received it. Define an incomplete claim as one missing any of those three elements.
Do not count a photo of a product catalog image as evidence of damage. Do not count a claim that only quotes the customer's description as linked unless the order number is recorded. A claim that was approved without the required evidence is still incomplete for this measure. Each definition should include one positive example and one boundary example so a second reviewer can apply it consistently.
Evidence standard and data dictionary
For each claim, capture the order number, the customer, the claim date, and the channel. Add the number and type of photos, whether the label is readable, whether the damage is visible, and whether a delivery note exists. Record the decision, the amount, and the reason. Keep the photos with the claim record rather than in a chat thread.
Also capture who took each photo and whether the customer confirmed the packaging was kept. Record any missing view and whether it was requested later. If the claim was denied, record the specific reason and the evidence that was absent. A second reviewer should be able to reach the same evidence assessment from the record.
If the business uses a template for customer photo requests, keep the version that was sent with the claim. A template that was updated after the claim can make the intake look better than it was. Record the date the template was used. The evidence should describe the claim as it arrived, not as it was later repaired.
Sampling and collection
Draw the sample from the claims queue rather than from a list of complaints. Record the export date, the filters, the row count, and the selection rule. Include claims from every channel and every product category so the result is not dominated by one promotion. If the population is ordered by time, a fixed interval spreads the sample across the window.
Collect the photos and the claim record as they existed when the claim was decided. If a customer sent a better photo after the decision, record the addition date and analyze the original evidence separately. Do not infer a delivery condition from the carrier's tracking alone. Note the collection date separately from the claim date.
Record the order value from the order system rather than from the claim message. A customer may round an amount or describe a replacement cost that differs from the invoice. Use the same source for every claim so the value bands are comparable. If the order was discounted, record the amount actually paid and the list price separately.
Analysis and reporting
Report the number of claims, the share with a usable photo, the share with a linked order, and the share with all three elements. Show the counts by channel, product category, and claim value. Present the share approved and the share denied, with the evidence status beside each. Use the same denominator for every rate.
Inspect the incomplete claims and describe the pattern without naming the customer or the employee. A cluster in one channel may point to a missing intake question, while a cluster in one product category may point to packaging. Compare the result with the day of week, because a weekend claim may be handled by a different team. Report the claims that were reopened separately.
A useful additional view is the share of claims where the missing evidence was requested but never received. That view separates an intake design problem from a customer follow-through problem. Show it beside the completeness rate and describe the difference. The owner can then decide whether to change the template or the follow-up rule.
Report the evidence status separately for claims that were decided within one day and claims that took longer. A fast decision on an incomplete claim is a different risk from a slow decision on a complete one. Add the median time to decision beside each group so the reader can see the trade-off. Keep the decision reason in the raw data for any later review.
Limitations and sources of error
The record cannot show whether the damage happened before or after delivery. It can show only what the customer photographed and what the team recorded. A clear photo is not proof of a valid claim, and a missing photo is not proof of a false one. The protocol also cannot capture a claim that was resolved in a phone call and never entered in the queue.
Photos can be cropped, filtered, or taken after the item was moved. A label can be readable while belonging to a different order. Record those risks and the treatment rather than assuming the evidence is honest. If the same person approves the claim and takes the photos, note the independence limitation and analyze that subgroup separately.
Review cadence and next action
Give the result to the owner and the person who runs the claims queue and ask them to choose one observable change. The change might be a required wide photo, an order-number field on the intake form, or a delivery-condition question. Record the owner, the effective date, and the measure that will show whether the change worked. Avoid changing the template and the approval limit at the same time.
Repeat the protocol with the same definitions after the change and compare the same fields. If completeness rose but the approval rate did not, the new evidence may not be the constraint. If approvals rose but denials became less consistent, the team may be relaxing the standard. Close the review when the owner accepts the finding and the next sample is scheduled.
Sources
- Federal Trade Commission, business guidance, Federal Trade Commission; checked 2026-11-07
- U.S. Small Business Administration, manage your business, U.S. Small Business Administration; checked 2026-11-07
- IRS, recordkeeping for small businesses, Internal Revenue Service; checked 2026-11-07
- U.S. Government Accountability Office, internal control standards, U.S. Government Accountability Office; checked 2026-11-07
Frequently asked questions
Does a favorable sample prove the routine caused the result?
No. The design is observational and supports a bounded operating decision, not a universal causal claim.
What should an owner review first?
Start with high-consequence defects, missing denominators, exceptions, and cases that required the owner to reconstruct the evidence.
Related research
How long do supplier bank-detail changes wait for independent verification?
A protocol for measuring the delay between a supplier bank-detail change and the independent check that confirms it.
How often do support macros cite a policy version that is no longer current?
A protocol for auditing saved replies against the current policy register, with a test for stale wording, missing versions, and untested macros.
What share of new assistant assignments have a named trainer and a first-week check?
A protocol for measuring onboarding coverage, with tests for a named trainer, a first-week review, and a record the next supervisor can use.