How to compare a mainland prepaid-card refund without treating the number as a verdict
Use this ordered guide before you accept a gym, beauty, or training-card refund quote. Re-read the official Interpretation text. This site does not upload contracts, does not call a court API, and is not legal advice (非正式法律意见).
- Step 1. Split the refund reason first. Payment within seven days may follow Article 14 principal return. Consumer-side reasons (no time, health, moving house) usually follow Article 19 original-price redemption even if Article 13 paragraph 2 allows termination. Operator relocation, unauthorized assignment, or closure follow Article 18.
- Step 2. Check the two Article 14 exceptions. If, when the contract was formed, you already received the same goods or services from this operator or from another operator, the seven-day no-reason branch closes. A more favorable contractual cooling-off term still prevails.
- Step 3. Choose session pack versus unlimited time card. Lesson packs use per-visit math. Unlimited time cards use the remaining-term ratio in Article 22, including the stop-service remaining term and the unused-expired paragraph 3 bar.
- Step 4. Record prepaid, gift amount or gift sessions, used count, list price, and discount price. “Buy 50 get 10” is gifted homogeneous sessions, not gifted money.
- Step 5. Non-consumer reason with a gift amount: preference ratio = prepaid ÷ (prepaid + gift), then redeemed value = corresponding list amount × ratio (Article 18 item 2).
- Step 6. Non-consumer reason with only a discount: redeemed value uses the discount unit price, not the original poster price (Article 18 item 1).
- Step 7. Consumer reason: redeemed value uses the pre-discount price. If there are gift sessions and no separate list price, the implied original unit is prepaid ÷ purchased sessions. If you claim the posted original price is clearly unreasonable and the merchant cannot produce pre-discount records, you may type a local market price yourself — the tool never estimates a market.
- Step 8. Apply Article 20. When discount or ratio redemption stays within prepaid but original-price redemption exceeds prepaid, the merchant’s request for the excess is not supported. The page shows a floor of zero and writes that the merchant may not demand a top-up. Do not read “trial refund 0” as “you must pay the shop.”
- Step 9. Unlimited time cards. Trial refund = prepaid × remaining term ÷ total term. If the operator already stopped service, count remaining term from the stop date. If you let the term expire without requesting service for your own reasons, Article 22 paragraph 3 yields trial refund 0.
- Step 10. Compare the merchant quote and any contract formula. Articles 14, 18 and 19 say a term more favorable to the consumer prevails. The wizard compares those two numbers; it does not treat a shop fee as a statutory deduction.
- Step 11. Export the comparison, the evidence list, and the next-path list. Keep the contract, payment proof, usage log, and chat screenshots. Negotiate first, then 12315 / a consumer association, then consider a small claim. Closed shops may have model judgments or group mediation — no promised result.
- Step 12. Read the disclaimer again. This is a public-text trial comparison, not a lawyer opinion and not a win guarantee. Zhangping’s statutory trial was 8250 yuan while mediation was 5600; Nanjing yoga’s statutory trial was 0 while mediation paid 2600.
Related: FAQ · Use cases · Home tool. Always prefer the official court publication for the binding wording. Copy the comparison table into your negotiation notes together with the official URL.
English context: After 1 May 2025, refund fights still show merchants quoting original prices plus penalties until the consumer appears to owe money. The gap this page fills is a single local comparison that branches on reason, handles gift money versus gift sessions, marks the no-overdraft rule, and lists paths — without uploading anything. It is not a clone of a pre-purchase value calculator, a generic complaint script, or a Taiwan-form contract estimator.