01

Ask for evidence only when it has a purpose

Every requested image, quantity update or document should support an explicit decision. If nobody can explain what the evidence unlocks, it is probably unnecessary collection.

02

Use a four-part evidence model

A useful evidence requirement states the subject, timing, acceptance rule and owner. This turns a vague request for proof into something a retailer can complete and an operator can review consistently.

  • Subject: the exact store, SKU, shelf or transaction.
  • Timing: when the initial and follow-up proof is due.
  • Acceptance: what must be visible or reconciled.
  • Ownership: who submits, reviews, corrects and approves.
Four-part evidence model for subject, timing, acceptance and ownership
Clear acceptance criteria reduce avoidable resubmission.
03

Make review proportional to risk

Low-risk completeness checks can be automated. Ambiguous media, identity mismatches, unusual stock variance and payment release should route to an authorised person with the full context.

AutoCompletenessFile, fields and due date
HumanAmbiguityContextual judgement
EscalateMaterial riskMoney, identity, dispute
04

Close the loop for the participant

A rejection without a reason creates support load. The participant should see the issue, required correction, due time and reviewer outcome. The final record should preserve both the original submission and the decision history.