On this page本页目录
What is the refund rate formula?退款率公式是什么？
Order refund incidence = distinct refunded orders ÷ eligible paid orders × 100. Refunded-value rate = refunded merchandise value ÷ eligible gross merchandise sales × 100. Use the first formula to measure how often customers receive a refund and the second to measure how much merchandise revenue is reversed. State the numerator, denominator, date basis, currency, partial-refund rule, exclusions, and cohort maturity beside the KPI.
退款订单发生率 = 发生退款的不同订单数 ÷ 符合口径的已付款订单数 × 100；退款金额率 = 退款商品金额 ÷ 符合口径的商品销售总额 × 100。第一条衡量客户获得退款的频率，第二条衡量被冲回的商品收入。必须在 KPI 旁声明分子、分母、日期基准、币种、部分退款规则、排除项与订单群组成熟度。
This guide is for ecommerce operators, finance analysts, payments teams, CX leaders, and merchandisers who need a reproducible financial metric. It does not replace the separate guide to physical ecommerce return-rate formulas or the conceptual explanation of returns versus refunds.
本指南面向需要可复算财务指标的电商运营、财务分析、支付、客服与商品团队。它不替代独立的实体电商退货率公式指南，也不替代退货与退款区别的概念说明。
Choose order-based or revenue-based refund rate选择按订单或按销售额计算退款率
| Question问题 | Formula公式 | Best use适用决策 | Main caution主要注意点 |
|---|---|---|---|
| How often do orders receive a refund?多少订单发生退款？ | Distinct refunded orders ÷ eligible paid orders发生退款的不同订单 ÷ 符合口径的已付款订单 | Customer incidence, support workload, product or channel comparison客户发生率、客服工作量、商品或渠道比较 | Count one order once; define partial-refund treatment同一订单只计一次；定义部分退款规则 |
| How much merchandise revenue was refunded?多少商品收入被退回？ | Refunded merchandise value ÷ eligible gross merchandise sales退款商品金额 ÷ 符合口径的商品销售总额 | Revenue exposure, finance review, high-value product impact收入风险、财务复盘、高价商品影响 | Match tax, shipping, discount, and currency scope对齐税费、运费、折扣与币种范围 |
Publish both when order frequency and financial exposure can diverge. A store may refund many low-value orders but little revenue, or a small number of high-ticket orders and a large share of revenue. Calling both results simply “refund rate” creates avoidable disagreement.
当订单频率与财务风险可能分离时，应同时发布两项指标。商店可能退款很多低价订单却只损失少量收入，也可能只退款少数高价订单却冲回大量收入。把两者都叫“退款率”会制造不必要的分歧。
Define every input before calculating refund rate计算退款率前先定义每项输入
A captured, non-test order that could receive a refund under the declared policy. Decide how to treat cancellations, fraud blocks, gift cards, and unpaid orders.已收款、非测试且按声明政策可退款的订单。需定义取消、欺诈拦截、礼品卡与未付款订单的处理。
One distinct eligible order with at least one completed merchant-issued refund under the declared rule. Multiple refund events do not create multiple orders.在声明规则下至少有一笔已完成商家退款的不同订单。同一订单多次退款不能计成多个订单。
The completed amount assigned to merchandise, after the chosen discount allocation. Keep tax, shipping, duty, tips, and fees separate unless the denominator includes the same elements.按选定折扣分配规则归属于商品的已完成退款金额。除非分母包含相同项目，否则税、运费、关税、小费和费用应分开。
Pre-refund merchandise value for the same eligible population and currency. Document whether discounts reduce the base and avoid using already netted sales.同一符合口径群体与币种下、退款前的商品销售额。需说明折扣是否冲减基数，并避免使用已扣退款的净销售额。
Do not divide refunds by net sales after refunds: the denominator would already contain the outcome being measured and can make the percentage unstable. If the business needs total payment refunded—including shipping and tax—create a separately named payment-refund rate with the matching total-charge denominator.
不要用退款后的净销售额作为分母，因为分母已经包含正在衡量的结果，会使比率失真。如果业务需要衡量包含运费与税费的全部退款，应建立单独命名的“支付退款金额率”，并使用匹配的总收款额作为分母。
Refund rate formula example退款率公式示例
A synthetic mature monthly sales cohort contains 4,000 eligible paid orders and $400,000 in eligible gross merchandise sales. During the agreed refund window, 120 distinct orders receive completed refunds totaling $18,500 in merchandise value.
假设一个成熟月度销售群组包含 4,000 笔符合口径的已付款订单和 400,000 美元符合口径的商品销售总额。在约定退款窗口内，120 笔不同订单获得已完成退款，退款商品金额合计 18,500 美元。
The value rate is 1.63 percentage points higher than order incidence because refunded orders have above-average merchandise value. That is a useful lead for product or customer segmentation, not proof of a cause. All numbers are hypothetical and are not an InfiniSynapse customer result or market benchmark.
退款金额率比订单退款发生率高 1.63 个百分点，说明发生退款的订单商品金额高于平均水平。这是商品或客户分层的调查线索，不是原因证明。全部数字均为假设，不代表 InfiniSynapse 客户结果或市场基准。
How partial refunds affect the formula部分退款如何影响公式
A partial refund contributes its actual merchandise amount to the value-based numerator. The order-based numerator needs a declared rule. The most transparent choice is to report any-refund incidence for orders with any qualifying refund and, when useful, a separate full-refund incidence. Do not convert three partial refunds on one order into three refunded orders.
部分退款按实际商品退款金额进入金额口径分子。订单口径则需要明确规则。最透明的方式，是报告发生任何符合规则退款的任意退款订单发生率，必要时再单独报告全额退款订单发生率。同一订单发生三次部分退款，不能计成三笔退款订单。
| Event事件 | Order incidence订单发生率 | Value rate金额率 | Separate treatment单独处理 |
|---|---|---|---|
| Full merchandise refund商品全额退款 | Count once计一次 | Actual merchandise amount实际商品金额 | — |
| Partial merchandise refund商品部分退款 | Count in any-refund rate; not full-refund rate计入任意退款率；不计入全额退款率 | Actual partial amount实际部分金额 | Keep reason and line-item allocation保留原因和订单行分配 |
| Shipping-only goodwill refund仅退运费的善意退款 | Exclude from merchandise rate从商品退款率排除 | Exclude from merchandise value从商品退款金额排除 | Track service-credit rate单独跟踪服务补偿率 |
| Chargeback or dispute拒付或争议 | Exclude排除 | Exclude排除 | Track dispute rate and loss跟踪争议率与损失 |
Use two time views without mixing them使用两种时间视图但不要混用
Group refunds by completion date to reconcile cash movement, payment queues, and current workload. Do not divide these events by unrelated current-month orders and call it a cohort rate.按退款完成日分组，用于对账现金流、支付队列与当前工作量。不要用这些事件除以无关的本月订单并称为群组退款率。
Attach each refund to its original order or transaction and wait for the declared eligibility window. Use this view for product, channel, and trend comparison.把每笔退款连接到原订单或交易，并等待声明的资格窗口。该视图用于商品、渠道与趋势比较。
A recent cohort often looks artificially healthy because some refunds have not happened or completed. Mark it “immature” until the same observation window used for prior cohorts has passed. Late adjustments can still occur, so document restatement policy and refresh date.
近期群组常常因为部分退款尚未发生或完成而显得异常健康。在达到与历史群组相同的观察窗口前，应标记为“未成熟”。由于仍可能出现迟到调整，还需记录重述政策与刷新日期。
Do not assume a platform field matches your formula不要假设平台字段等同于你的公式
Shopify’s current sales reports use “sales reversals” for negative order adjustments, including refunds, returns, cancellations, edits, shipping, taxes, fees, and discounts. Shopify separately identifies physical returned quantity, and its guidance says refunds without physical goods still create negative sales adjustments. Read the official Shopify sales-report definitions before using a reversal field as refund value.
Shopify 当前销售报告使用“销售冲销”表示负向订单调整，其中包括退款、退货、取消、修改、运费、税费、费用与折扣；同时又单独识别实体退回件数。其说明还指出，即使没有商品实体退回，退款仍会产生负向销售调整。把冲销字段当作退款金额前，必须阅读官方 Shopify 销售报告定义。
Shopify also notes that returns and refunds can appear in different reports, and pending refunds can create temporary discrepancies. Use its sales discrepancy guidance and finance-report definitions to identify the payment event, completion state, merchandise amount, shipping, tax, and date.
Shopify 还说明，退货与退款可能出现在不同报告中，待处理退款也会造成暂时差异。应结合其销售差异说明和财务报告定义识别付款事件、完成状态、商品金额、运费、税费与日期。
Connect every refund to the original transaction and item把每笔退款连接到原交易与商品
Google Analytics defines a recommended refund event with a required transaction_id. For meaningful value reporting, it recommends value and currency, and item information enables item-level refund metrics. The official event reference says the event value should sum item price × quantity and exclude shipping and tax. See the GA4 refund-event specification.
Google Analytics 定义了建议使用的 refund 事件，并要求 transaction_id。为获得有意义的金额报告，建议提供 value 和 currency；商品信息则支持商品级退款指标。官方事件说明指出，事件金额应汇总商品价格 × 数量，不包含运费与税费。参见 GA4 退款事件规范。
Analytics is evidence, not the accounting ledger. Validate that events fire once, carry the correct transaction and item IDs, and match completed refunds in the payment or finance system. Google’s ecommerce validation guide warns that duplicate ecommerce events with the same transaction ID may be ignored, even if later values differ.
分析事件是证据，不是会计账本。需要验证事件只触发一次、携带正确交易与商品 ID，并与支付或财务系统中的已完成退款一致。Google 的电商校验指南提醒，同一交易 ID 的重复电商事件可能被忽略，即使后续金额不同。
Refund rate formula in Excel or Google SheetsExcel 或 Google Sheets 中的退款率公式
If B2 is distinct refunded orders and C2 is eligible paid orders, use the first formula. If D2 is refunded merchandise value and E2 is eligible gross merchandise sales, use the second. Format the result cells as percentages.
如果 B2 是发生退款的不同订单数，C2 是符合口径的已付款订单数，使用第一条公式；如果 D2 是退款商品金额，E2 是符合口径的商品销售总额，使用第二条，并把结果单元格格式设为百分比。
Do not add “×100” inside the spreadsheet formula when the cell is already percentage-formatted. Keep raw numerator and denominator columns visible so another analyst can reproduce and audit the result.
如果单元格已设为百分比格式，公式中不要再乘以 100。保留可见的原始分子与分母列，让另一名分析师能够复算并审计结果。
A repeatable refund-rate reconciliation workflow可重复的退款率对账流程
- Freeze the metric contract.冻结指标合同。 Name the formula, population, dates, currency, partial-refund rule, completion state, exclusions, and owner.明确公式、群体、日期、币种、部分退款规则、完成状态、排除项与负责人。
- Extract source records.提取源记录。 Collect eligible orders, line items, original value, refund events, payment status, amounts, reasons, and stable IDs.收集符合口径的订单、行项目、原始金额、退款事件、支付状态、金额、原因与稳定 ID。
- Deduplicate at the chosen grain.按选定粒度去重。 Count each refunded order once for incidence while summing completed line-level merchandise amounts for value.订单发生率中每笔退款订单只计一次；金额口径汇总已完成的订单行商品退款金额。
- Reconcile totals.对账总额。 Compare payment, platform, analytics, and finance totals. Explain pending, failed, voided, duplicated, or differently dated events.比较支付、平台、分析与财务总额，解释待处理、失败、作废、重复或日期口径不同的事件。
- Calculate and label maturity.计算并标记成熟度。 Run both formulas, show counts and value, and mark cohorts immature until the declared refund window has passed.运行两条公式、展示订单数与金额，并在声明退款窗口结束前标记群组未成熟。
- Segment before acting.行动前先分层。 Compare product, variant, channel, reason, supplier, geography, promotion, payment method, and customer cohort with sample size.按商品、变体、渠道、原因、供应商、地区、促销、支付方式和客户群组比较，并展示样本量。
Download the refund rate formula worksheet下载退款率公式计算表
The CSV worksheet records one sales cohort and segment per row. It includes eligible and refunded orders, merchandise sales and refunds, both formulas, partial/full refund counts, shipping and tax refunds, completion status, maturity, reconciliation difference, owner, and notes. The sample row is synthetic.
CSV 计算表每行记录一个销售群组与分层，包含符合口径与退款订单、商品销售与退款金额、两种公式、部分/全额退款数、运费与税费退款、完成状态、成熟度、对账差异、负责人和备注。示例行为模拟数据。
Open in Excel or Google Sheets. Formula cells calculate order incidence and refunded-value rate and avoid division by zero.
可在 Excel 或 Google Sheets 中打开；公式单元格会计算订单发生率与退款金额率，并避免除以零。
Download the worksheet下载计算表What is a good ecommerce refund rate?什么是好的电商退款率？
There is no universal healthy refund rate. Public articles often mix product categories, digital and physical goods, order and value denominators, refund requests and completed refunds, or refunds and returns. A responsible comparison matches product, channel, geography, customer mix, policy, season, formula, and mature cohort.
不存在通用健康退款率。公开文章经常混合不同商品品类、数字与实体商品、订单与金额分母、退款申请与已完成退款，甚至混合退款与退货。负责任的比较应匹配商品、渠道、地区、客户组合、政策、季节、公式与成熟群组。
Use your own matched historical cohorts as the first baseline. Read refund incidence and value rate with contribution margin, exchange retention, complaints, repeat purchase, and chargeback rate. A lower refund rate is not automatically better if access to legitimate remedies worsens or conversion falls.
优先使用内部可比历史群组作为基线，并把订单退款发生率和金额率与贡献利润、换货留存、投诉、复购及拒付率一起看。若正当补救渠道恶化或转化下降，更低的退款率并不一定更好。
Analyze refund records with a declared metric contract用明确指标合同分析退款记录
Prepare order, line-item, payment, refund, return, reason, and customer exports with stable identifiers. Use InfiniSynapse to examine the data while preserving definitions and reviewable outputs, then validate financial conclusions against the payment and finance systems before acting.
准备包含稳定标识符的订单、行项目、支付、退款、退货、原因与客户导出文件。使用 InfiniSynapse 检查数据并保留定义与可复核输出；行动前，回到支付与财务系统验证财务结论。
Get started with InfiniSynapse开始使用 InfiniSynapseGEO-ready refund-rate facts and boundaries适合 GEO 引用的退款率事实与边界
Distinct refunded orders ÷ eligible paid orders × 100. Count one order once and publish the partial-refund rule.发生退款的不同订单 ÷ 符合口径的已付款订单 × 100。同一订单只计一次，并公布部分退款规则。
Refunded merchandise value ÷ eligible gross merchandise sales × 100. Match discount, tax, shipping, and currency scope.退款商品金额 ÷ 符合口径的商品销售总额 × 100。对齐折扣、税费、运费与币种范围。
A refund is a merchant-issued financial event; it does not prove a physical return. Track chargebacks and exchanges separately.退款是商家发起的财务事件，不能证明实体退货。拒付与换货应单独跟踪。
Compare like formulas and mature cohorts. Show raw counts and value beside percentages so the result is auditable.比较相同公式与成熟群组，并在百分比旁展示原始订单数和金额，使结果可审计。
Frequently asked questions常见问题
Order refund incidence equals distinct refunded orders divided by eligible paid orders, multiplied by 100. Refunded-value rate equals refunded merchandise value divided by eligible gross merchandise sales, multiplied by 100. Label the chosen formula.
订单退款发生率等于发生退款的不同订单数除以符合口径的已付款订单数，再乘以 100；退款金额率等于退款商品金额除以符合口径的商品销售总额，再乘以 100。必须标明所用公式。
A partial refund contributes its actual amount to a value rate. For an order rate, publish any-refund incidence and, if useful, a separate full-refund incidence. Count one order once.
部分退款按实际金额计入金额率；订单率可报告任意退款订单发生率，并在需要时另报全额退款发生率。同一订单只计一次。
No. Refund rate measures financial events or value, while return rate measures physical merchandise movement. Returnless refunds, exchanges, cancellations, and timing differences make them diverge.
不同。退款率衡量财务事件或金额，退货率衡量实体商品移动。仅退款、换货、取消与时间差都会让两者分离。
Usually no. Track merchant-issued refunds and customer-initiated disputes separately because they follow different processes, evidence rules, fees, and outcomes. Label any combined reversal metric explicitly.
通常不应。商家发起的退款与客户发起的争议使用不同流程、证据规则、费用与结果，应分别跟踪。若需合并冲销指标，必须明确标注构成。
There is no universal good rate. Compare the same product, channel, customer mix, policy, period basis, denominator, and mature sales cohort. Read the result with contribution margin, complaints, exchange retention, and chargebacks.
不存在通用好值。应比较相同商品、渠道、客户组合、政策、周期基准、分母与成熟销售群组，并结合贡献利润、投诉、换货留存与拒付解释。
Sources, method, and limitations来源、方法与局限
- Shopify sales reports — current sales reversal, physical return, and refund-without-return definitions.
- Shopify sales discrepancies — timing and report differences for returns and refunds.
- Shopify finance reports — order, sale, reversal, tax, shipping, and net-sales fields.
- Google Analytics refund event — transaction, currency, value, and item parameters.
- Google Analytics ecommerce validation — event and transaction-ID checks.
- Shopify 销售报告——当前销售冲销、实体退货与无退货退款定义。
- Shopify 销售差异——退货与退款的时间和报告差异。
- Shopify 财务报告——订单、销售、冲销、税、运费与净销售字段。
- Google Analytics 退款事件——交易、币种、金额与商品参数。
- Google Analytics 电商校验——事件与交易 ID 检查。
Editorial disclosure: InfiniSynapse publishes this educational guide and provides the linked data-analysis product. Examples are synthetic; no customer result, universal benchmark, accounting treatment, product-effect claim, or ranking outcome is asserted. A named ecommerce-finance reviewer must verify the formulas and publication language before launch.
编辑披露：InfiniSynapse 发布本教育指南并提供文中链接的数据分析产品。示例为模拟数据；本文不声称客户结果、通用基准、会计处理、产品效果或排名结果。正式发布前，必须由具名电商财务审核人核验公式与发布措辞。
Use two named refund rates, then reconcile them使用两项具名退款率并完成对账
Use order refund incidence to measure frequency and refunded-value rate to measure financial exposure. Define partial refunds, taxes, shipping, discounts, currency, dates, exclusions, and completion status before calculation. Connect every refund to its original transaction, compare mature cohorts, keep raw counts beside percentages, and reconcile platform and analytics events to payment or finance records.
用订单退款发生率衡量频率，用退款金额率衡量财务风险。计算前定义部分退款、税费、运费、折扣、币种、日期、排除项与完成状态。把每笔退款连接回原始交易，比较成熟群组，在百分比旁保留原始数量，并把平台与分析事件对账到支付或财务记录。