Your refund rate moved.
Up or down, the dashboard treats it like a single story. The floor knows better: some tickets are short packs, some are claims your records can’t support, some are product or expectation, and some you simply haven’t sorted yet.
A refund rate is a symptom. It cannot tell you what to fix.
This guide is the wider system: where refunds can start across the live-selling lifecycle, how to map causes without guessing buyers are dishonest, and a 30-day plan to attack one leak at a time.
It does not replace the missing-item deep dive. When the ticket is “ordered three, got two,” go here: The Refund Started Before the Customer Opened the Box.
The lifecycle: where a refund can begin
Refunds don’t only start when the buyer opens the box. Trace the chain:
- Product / sourcing — quality, fragile design, shade variance that the stream oversells
- Receive / put-away — wrong counts, wrong IDs, damaged units marked available
- Prepare / list / show — wrong expectation, bundle contents unclear, giveaway chaos
- Pick — wrong twin, short pick, gift skipped
- Check / scan — a miss that would have caught it
- Pack / ship — crush, label mismatch, late scan drama that becomes cancel pressure
- Support / claim — slow or unclear response; evidence missing; platform rules
- Learning loop — the same leak next week because nobody owned the cause
If you only argue at step 7, you pay for leaks born at 2–6.
Cause Map
Use the same buckets as the Refund & Rework Cost Calculator when you sort tickets.
| Category | ID | What It Usually Means | First Fix Type |
|---|---|---|---|
| Packing / count errors | packing_count | Wrong item, short pack, missing gift, label mismatch | Floor process |
| Claims inconsistent with available records | claim_noise | What the buyer described against what your records show (or you believe the pack was complete) | Evidence + triage — no promises about outcomes |
| Product / delivery issues | product_quality | Defect, transit damage after a sound pack, expectation set on stream | QC / supplier / stream honesty / pack standards |
| Unresolved / other | unsorted | Don’t know yet | Sort before you crown a villain |
Confirmed vs suspected. Mark both. Don’t treat a guess as a floor conviction.
Do not assume buyer dishonesty without evidence. Do not treat packing footage as an automatic dispute win — platforms decide under their own rules.
Rate vs dollars vs hours (don’t mix them)
Ticket count % is not dollar %. A few expensive fragrance refunds can outweigh many low-dollar tickets. Replacement and postage are not the same line as the refund. Rework hours are a third hit — time that doesn’t show on the refund rate at all.
When you review, keep columns separate: cases · refund $ · replacement $ · postage $ · rework hours.
Hypothetical Example (Illustrative — Not A Customer Case)
All figures below are illustrative sample data invented for teaching. Not a benchmark. Not a promise of typical results.
Window: 100 refund-related tickets after live shows in one sample month.
| Category | Cases | Share Of Cases | Refund $ | Replacement $ | Postage $ | Rework Hours |
|---|---|---|---|---|---|---|
| Packing / count errors | 40 | 40% | $1,200 | $400 | $180 | 20 |
| Claims inconsistent with records | 30 | 30% | $900 | $0 | $0 | 25 |
| Product / delivery issues | 20 | 20% | $1,500 | $200 | $90 | 8 |
| Unresolved / other | 10 | 10% | $300 | $50 | $20 | 10 |
| Total | 100 | 100% | $3,900 | $650 | $290 | 63 |
What the symptom rate hides in this sample: packing / count errors are 40% of cases but not the largest refund dollars (product / delivery is $1,500 against $1,200).
Claims inconsistent with records cost $0 in replacements here but 25 rework hours — expensive in attention even when you don’t reship.
If you only chased the biggest case pile, you’d start with packing / count. If you only chased the biggest refund dollars, you’d start with product / delivery. Both can be right — pick with eyes open, and don’t pretend case % equals dollar %.
Illustrative rework value (teaching only): if loaded labor is assumed at $30/hour, 63 hours ≈ $1,890 of capacity — separate from the $3,900 refund cash. That $30 is an assumed teaching rate, not your rate and not a benchmark.
30-Day Review Plan
- Days 1–3 — Baseline. Pull the last 30 days (or last 50 tickets). Sort into the four categories. Mark confirmed vs suspected. Total cases, refund $, replacement $, postage and rework hours separately.
- Days 4–7 — Pick one leak. Choose the pile you’ll attack for two weeks (by cases, by dollars, or by hours — write why). If unresolved is huge, sort first.
- Days 8–21 — One fix. Count-before-seal on multi-item; gift checklist; evidence filing; fragile pack standard; a named response owner. One fix — not five.
- Days 22–30 — Re-measure. Same window type. Same columns. Did confirmed cases in that bucket drop? Did hours drop? Decide: keep, widen, or switch piles.
Missing-item deep dive anytime: The Refund Started Before the Customer Opened the Box. Show money line, with refunds as one cost: Your Biggest Show Isn’t Always Your Best Show.
Put It To Work
- Record what happened on one recent refund or replacement (order reference kept private).
- Identify the earliest likely cause on your cause map — buying, listing, pick, check, pack, carrier, or post-sale.
- Assign one correction and an owner.
- Check recurrence after the next selling period: same cause, or a new one?
Use the on-page worksheet and, when you want the full blank set, the Refund Review Kit download further up this page. Do not build a second kit.
Bottom Line
Your refund rate can move while the real leak stays put.
Find where refunds begin. Separate cases, dollars, and hours. Fix one confirmed pile at a time — so the next show is worth the work.
Part of SaleTrail Campus: practical guides for Whatnot and TikTok Shop sellers.
Print The Refund Review Worksheet
The single-sheet version of the cause map and the 30-day plan, sized for a clipboard at the bench.
Open The Printable WorksheetFind where your refunds begin.
Five blank sheets for a 30-day refund review: case log, cause map, suspected vs confirmed, cases and dollars counted apart, and one plan.
- Case log for the last 30 days
- Cause map across buying, listing, pick, check, pack, carrier and post-sale
- Suspected vs confirmed sheet, kept deliberately separate
- Cases, dollars, replacements, postage and rework hours in their own columns
- A 30-day plan with one change per week
- A short page on running the review without fooling yourself
Blank sheets only. The kit states no acceptable refund rate, no benchmark and no prediction about how a platform will decide a case.
Email required. The guide and the on-screen calculator stay open to everyone.