惯性聚合 高效追踪和阅读你感兴趣的博客、新闻、科技资讯
阅读原文 在惯性聚合中打开

推荐订阅源

博客园 - Franky
D
Docker
Jina AI
Jina AI
The GitHub Blog
The GitHub Blog
博客园 - 聂微东
B
Blog RSS Feed
大猫的无限游戏
大猫的无限游戏
M
MIT News - Artificial intelligence
Vercel News
Vercel News
Microsoft Security Blog
Microsoft Security Blog
博客园 - 叶小钗
爱范儿
爱范儿
D
DataBreaches.Net
Hugging Face - Blog
Hugging Face - Blog
IT之家
IT之家
Recent Announcements
Recent Announcements
U
Unit 42
腾讯CDC
奇客Solidot–传递最新科技情报
奇客Solidot–传递最新科技情报
宝玉的分享
宝玉的分享
量子位
Stack Overflow Blog
Stack Overflow Blog
Cyber Security Advisories - MS-ISAC
Cyber Security Advisories - MS-ISAC
Microsoft Azure Blog
Microsoft Azure Blog

DEV Community

Authentication Security Deep Dive: From Brute Force to Salted Hashing (With Java Examples) Why AI Systems Don’t Fail — They Drift Spilling beans for how i learn for exam😁"Reinforcement Learning Cheat Sheet" I Replaced Chrome with Safari for AI Browser Automation. Here's What Broke (and What Finally Worked) How Python Borrows Other People's Work The $40 Architecture: Processing 1 Billion API Requests with 99.99% Uptime Vibe Coding: A Workflow Guide (From Zero to SaaS) Most webhook security guides protect the wrong side. The scary part is delivery. Headless CMS for TanStack Start: Build a Blog with Cosmic EU Age Verification App "Hacked in 2 Minutes" — What Actually Happened Comfy Cloud’s delete function does not actually remove files Running AI Models on GPU Cloud Servers: A Beginner Guide Event-driven media intelligence with AWS Step Functions and Bedrock I scored 500 AI prompts across 8 quality dimensions — here's what broke How to Call Google Gemini API from Next.js (Free Tier, No Backend Needed) The Portal Protocol: Reclaiming Human Connection in the Age of AI How to Fix Your Team's Scattered Knowledge Problem With a Self-Hosted Forum Intro to tc Cloud Functors: A Graph-First Mental Model for the Modern Cloud Designing Multi-Tenant Backends With Both Ownership and Team Access I Built a Neumorphic CSS Library with 77+ Components — Here's What I Learned PostgreSQL Performance Optimization: Why Connection Pooling Is Critical at Scale Cómo construí un SaaS multi-rubro para gestionar expensas en Argentina con FastAPI + Vue 3 🚀 I Built an Ethical Hacking Scanner Tool – Open Source Project I Replaced /usage and /context in Claude Code With a Single Statusline A Pythonic Way to Handle Emails (IMAP/SMTP) with Auto-Discovery and AI-Ready Design I Collected 8.9 Million Polymarket Price Points — Here's What I Found About How Markets Really Move EcoTrack AI — Carbon Footprint Tracker & Dashboard Everyone's Using AI. No One Agrees How. 5 self-hosted ebook managers worth trying in 2026 Building Your First AI Agent with LangChain: From Chatbot to Autonomous Assistant
Clawing back affiliate commissions when a customer refund...
Mihir kanzariya · 2026-06-18 · via DEV Community

You pay an affiliate 30% for referring a customer. Three weeks later that customer refunds. Now the affiliate is holding a commission on a sale that no longer exists, and you have to decide what happens to that money. If you only thought about this after it happened, you have already lost the clean version of the fix.

I build affiliate software on Stripe, and the clawback case is the one teams skip until it bites. Here is how to design for it before it does.

Earned is not the same as payable

The first mistake is treating a commission as money the moment the sale clears. It isn't. A commission is a claim against a sale that might still reverse. Stripe makes that gap concrete: a customer can refund whenever they want, and a card dispute can land up to 120 days after the charge. Your payout schedule (say, monthly) runs on a completely different clock.

So you have two timelines that don't line up: how long a sale can still reverse, and how soon you pay out. Pay before the first window closes and you are paying real money on revenue you may have to give back.

Hold commissions pending before they become payable

The fix is a holdback period. A commission lands in a "pending" state when the sale clears, and it only becomes "available" for payout after N days with no refund or dispute. Set N from your own refund data, not a guess. If most refunds happen in the first 30 days, a 30 to 45 day hold covers the bulk of them without making affiliates wait forever.

This one decision removes most clawbacks entirely, because the commission never reaches an affiliate's wallet until the risky window has passed.

Store a ledger, not a balance

Here is the part that saves you later: never store an affiliate's earnings as a single number you edit. Store immutable ledger entries and compute the balance from them.

commission_entries
  id
  affiliate_id
  charge_id            -- the Stripe charge this relates to
  type                 -- earned | reversed | paid_out
  amount_cents         -- signed: + for earned, - for reversed/paid
  status               -- pending | available
  created_at

A refund does not edit the original row. It writes a new reversed entry that points at the same charge. The affiliate's balance is just SUM(amount_cents) over their entries. You get an audit trail for free, every reversal is explainable, and you never lose the history of why a number changed. When an affiliate emails asking why their balance dropped, you can show them the exact charge and the exact reversal.

When the refund lands after you already paid

Sometimes a refund or a dispute arrives after the commission went out the door, usually because a chargeback showed up 60 days later. Now the affiliate's balance goes negative. You have three honest options:

  1. Carry the negative forward against future earnings. This is the common one and the fairest: the next commissions they earn pay down the deficit before anything becomes payable again.
  2. Reverse the transfer. If you pay through Stripe Connect, you can reverse a past transfer, but it is aggressive and it will surprise people. Reserve it for fraud or for affiliates who have gone silent with a large negative.
  3. Write it off below a threshold. Chasing back a 4 dollar commission costs more in goodwill than it returns. Pick a floor and absorb anything under it.

Whatever you choose, decide it once and apply it the same way every time. Inconsistent clawbacks are how you lose good affiliates.

The Stripe wiring

Two webhooks drive this: charge.refunded and charge.dispute.created. Each one looks up the commission entries tied to that charge and writes a reversed entry for the matching amount. For a partial refund, reverse the same proportion of the commission, not the whole thing.

A detail worth stating: process these as triggers to re-read state, not as the final word. A refund webhook can arrive while the original commission is still settling on your side. Look up the current state of the charge and the existing entries, then write the reversal against what you actually find. (Idempotency matters here too: key the reversal on the refund ID so a redelivered webhook doesn't reverse twice.)

Show pending and available separately

This is a product decision as much as an accounting one. Affiliates hate surprise clawbacks more than they like fast payouts. If your dashboard shows one big "earnings" number that quietly drops when a refund lands, every reversal feels like you took their money. Show pending and available as two separate numbers, with the date each pending commission becomes available. Then nobody counts money that isn't theirs yet, and a reversal during the pending window is a non-event instead of a support ticket.

We keep payout fees at zero and we run this exact pending-then-available model, because the trust cost of getting it wrong is higher than any fee. (Disclosure: I work on Referralful, https://referralful.com/?utm_source=dev.to&utm_medium=content&utm_campaign=article , affiliate software built on Stripe.)

Design the ledger and the holdback before you write your first commission. Retrofitting an immutable ledger onto a system that stored a single balance is a migration you do not want to run while affiliates are watching their numbers.