Coupon and cashback partners routinely intercept sales that content and influencer partners already closed, thanks to last-click attribution. How to detect the leakage, tell incremental coupon traffic from parasitic traffic, and lock down exclusivity with real network tooling.
Quick Answer
What is commission stacking in affiliate marketing?
Commission stacking occurs when multiple affiliate partners touch the same customer journey but the program's last-click attribution model pays commission only to whichever partner's link or cookie was active closest to checkout — typically a coupon or cashback site — regardless of which partner actually generated the original demand. The advertiser effectively pays commission on a sale that a different partner already earned through content, reviews, or influence.
# Commission Stacking and Channel Cannibalization: What Affiliate Programs Get Wrong
Most affiliate programs still pay on last click. That single design choice is the root cause of the most persistent profitability problem in performance marketing: a coupon or cashback partner intercepts a sale that a content creator, reviewer, or influencer already generated — and walks away with 100% of the commission. The advertiser pays twice for one sale: once in margin given up to the discount, once in commission paid to a partner who didn't create the demand.
This isn't a fringe issue affecting a handful of programs. It's structural to how most affiliate networks attribute conversions by default, and it shows up in three specific, addressable failure modes: coupon code leakage, browser-extension interception, and channel cannibalization between discovery partners and discount partners. Each has real prevention mechanisms available today — most affiliate managers simply haven't turned them on.
What Commission Stacking Actually Is
Commission stacking happens when more than one affiliate touchpoint contributes to a single sale, but the program's attribution model pays only the partner that gets credit for the final click — regardless of who actually influenced the purchase decision. In practice, the sequence usually looks like this:
- A content creator or comparison site introduces a customer to a product, builds trust, and earns the click that starts the session.
- Before checkout, the customer opens a new tab and searches "[brand] coupon code," or a browser extension like a coupon aggregator auto-applies a code at checkout.
- That coupon or cashback site's tracking cookie overwrites the original referring affiliate's cookie.
- The sale is attributed — and commissioned — entirely to the coupon site, even though it did none of the persuasion work.
Under last-click attribution, which most affiliate programs use by default, the coupon site that provided the discount code gets 100% of the commission even though the original content creator did the work of education and persuasion. The practical effect: an advertiser pays a coupon site to intercept a sale that was already decided, while the partner who actually generated demand gets nothing — and eventually stops promoting the program.
Why This Is a Compliance and Fraud Problem, Not Just a Fairness Problem
Affiliate managers often treat commission stacking as an internal fairness dispute between partner tiers. That undersells it. Three mechanics turn it into a fraud and compliance issue with direct P&L exposure:
Cookie overwriting and code leakage. Coupon codes issued exclusively to one publisher routinely leak onto public coupon aggregator sites, deal forums, and other publishers' pages. Once leaked, any customer can find and apply the code — and depending on tracking setup, the last-touch publisher captures the commission even if they never had legitimate access to the code.
Browser-extension auto-injection. Coupon-hunting browser extensions scan checkout pages and apply codes automatically, often triggering an affiliate cookie drop in the process — without the shopper ever having intentionally clicked an affiliate link for that visit. This is the mechanism behind the widely reported Honey litigation: independent trackers cited in industry reporting put Honey's merchant count at roughly 35,000 in December 2024, falling to just above 28,000 as the controversy unfolded — a reported net loss of more than 7,000 merchant partnerships — coinciding with a wave of creator class-action lawsuits alleging the extension overwrote other creators' affiliate links and claimed commissions it hadn't earned ([ppc.land](https://ppc.land/honey-loses-7-000-merchants-as-judge-lets-creator-lawsuit-proceed/)). Those figures come from third-party tracking rather than an official PayPal (Honey's parent company) disclosure, so treat the exact numbers as directional; PayPal has separately said the specific injected-code practice at issue in the litigation affected a small fraction of its traffic and had been deactivated. What isn't in dispute: a federal judge in the Northern District of California allowed the consolidated creator class action to proceed into discovery rather than dismissing it ([Courthouse News Service](https://www.courthousenews.com/paypal-unlikely-to-dodge-class-action-from-influencers-over-lost-commissions/)).
Non-incremental attribution. clean.io, a fraud-detection vendor with a direct commercial interest in the size of this problem, has publicly claimed — in its own press releases, not an independent audit — that third-party browser extensions like Honey, RetailMeNot, and Capital One Shopping cost merchants more than $3 billion annually in misattributed affiliate commissions ([Newswire](https://www.newswire.com/news/cleanio-prevents-affiliate-attribution-fraud-caused-by-third-party-21759110); [GlobeNewswire](https://www.globenewswire.com/en/news-release/2022/07/06/2474934/0/en/clean-io-Prevents-Affiliate-Attribution-Fraud-Caused-by-Third-Party-Coupon-Extensions-That-Cost-Merchants-3-Billion-Annually.html)). Treat that $3B figure as a vendor estimate rather than an audited industry benchmark — but it is directionally consistent with the broader pattern documented across the affiliate industry: last-touch attribution routinely pays commission on conversions that would have happened without the paid touchpoint.
Coupon Sites Aren't Always the Problem — That's Exactly Why Detection Matters
It would be a mistake to treat every coupon or deal partner as fraudulent. The evidence on incrementality is genuinely mixed, and that's precisely why blanket bans and blanket tolerance are both wrong answers.
Inmar Intelligence's shopper research found that coupons lift average order value by roughly 19.7% for online purchases (versus 17.1% in-store), and that 92% of shoppers who used a coupon online said the offer changed their purchase decision — leading them to buy more, buy sooner, or try a brand for the first time ([Inmar Intelligence](https://www.inmar.com/blog/press/new-inmar-intelligence-data-show-strong-growth-digital-coupons-which-significantly)). That's a real incrementality signal — some coupon and deal traffic is genuinely bringing in customers who wouldn't have converted otherwise, particularly in price-sensitive categories where shoppers are actively comparison-shopping before they've committed to a brand.
The failure mode isn't "coupon partners are bad." It's that most programs have no mechanism to tell the difference between a coupon partner who is generating net-new, price-sensitive customers and one who is simply sitting at the bottom of the funnel harvesting cookie credit from sales that content and influencer partners already closed. Without that distinction, programs either overpay indiscriminately or ban coupon partners wholesale and lose the genuinely incremental volume along with the parasitic volume.
How to Detect Cannibalization in Your Own Program
Before deploying any prevention mechanism, quantify the problem. Four diagnostics, roughly in order of effort:
1. New-vs-returning customer split by partner type
Pull new-customer rate by publisher category (content, coupon/deal, cashback, loyalty/browser extension, email, influencer). If your coupon and cashback partners show a materially lower new-customer rate than your content and influencer partners, that's the clearest available signal that they're intercepting decided purchases rather than creating them. This comparison — tracking new-versus-returning customer rates and comparing return rates across affiliate types — is widely cited as the most reliable practical method for separating incremental sales from last-click cannibalization, short of running a formal incrementality test.
2. Path-length and multi-touch reporting
Most major networks (Awin, Impact, CJ) expose multi-touch or assisted-conversion reporting in some form. Pull it for a rolling 90-day window and look at how often a content or influencer click precedes a coupon-site click in the same converting session. A high co-occurrence rate is the fingerprint of cannibalization, not coincidence.
3. Session-timing analysis
If your platform or a third-party tracking layer logs click timestamps, look at the gap between the first affiliate click and the last one before conversion. Sessions where a coupon click lands in the final 60–120 seconds before checkout, immediately after a long content-read session from a different partner, are the pattern to flag for exclusivity or exclusion review.
4. Coupon code audit
Search your own exclusive codes against public coupon-aggregator sites monthly. If a code you issued to Partner A for exclusive use is indexed on five other coupon sites, you have active leakage — and every sale using that code outside Partner A's traffic is a stacking event, not a legitimate conversion.
Real Prevention Mechanisms, Not Just Policy
The good news: the major affiliate networks have built specific tooling for this exact problem. The gap is almost always adoption, not availability.
Coupon code exclusivity and attribution locking
Awin's coupon attribution feature (part of its My Offers tool) lets an advertiser assign an exclusive code to one publisher so that commission from every sale using that code goes to that publisher — even if no affiliate cookie is present, or if the sale would otherwise be attributed to a different channel. If a leaked code is used through another partner's link, the sale is still automatically routed to the exclusive publisher, which directly blocks the leakage-driven stacking scenario described above ([Awin](https://www.awin.com/us/news-and-events/awin-news/coupon-attribution)). Awin has also automated exclusive single-use code generation via prefix matching, so advertisers no longer need to manually register each code-partner pairing ([Awin](https://help.awin.com/docs/using-the-my-offers-tool)).
Last-click exceptions and rule-based override attribution
Rather than accepting whatever partner happens to hold the last cookie, rule-based attribution lets an advertiser define exceptions: content and influencer partners keep credit even when a coupon click follows within a defined window, or specific partner categories are excluded from receiving credit on sessions where a higher-value partner touched the journey first. This is typically configured at the network or tracking-platform level (Awin's advanced commissioning rules, Impact's commission engine, or a dedicated attribution layer sitting on top of the network) rather than through the network's out-of-the-box defaults.
Tiered and role-based commission structures
Several practitioner sources recommend commission structures that pay differently based on funnel position rather than paying every partner the same flat rate regardless of role. Affiliates who consistently appear as first-touch across the customer journey — the actual demand generators — get higher rates for acquisition; partners who typically show up last-touch get a lower rate reflecting their conversion-assist role rather than a full acquisition credit. This requires multi-touch data to implement correctly, which is why detection (above) has to come before restructuring.
Browser-extension blocking and toggle-based prevention
Because extension-based coupon injection is the mechanism behind the largest, most litigated cases of commission theft, advertiser-side script tooling that detects and blocks unauthorized extension code injection at checkout (the category clean.io and similar vendors sell into) has become a standard line item for large e-commerce programs, separate from anything the affiliate network itself provides.
Comparison: Prevention Mechanisms by Platform
| Mechanism | Awin | Impact | CJ | Levanta (Amazon) |
|---|---|---|---|---|
| Exclusive coupon-code attribution | Native (My Offers / coupon attribution) | Available via commissioning rules, configured per program | Available via custom tracking setup, not a turnkey feature | Not applicable — Amazon Associates governs coupon behavior at the marketplace level |
| Last-click override / rule-based exceptions | Advanced commissioning rules | Commission engine supports custom rule logic | Requires custom integration work | Limited — Levanta manages creator-brand deals, not network-wide attribution rules |
| Prefix-based automated exclusivity | Yes | Not standard; manual code-partner mapping typical | Not standard | Not applicable |
| Multi-touch / assisted-conversion reporting | Available | Available | Available, more limited detail | Not applicable — single-network model |
| Third-party extension-injection blocking | Requires external tooling | Requires external tooling | Requires external tooling | Requires external tooling |
None of the four platforms managed brands operate on (Levoit, Cosori, TCL, Insta360 all run programs across combinations of Impact, Awin, CJ, and Levanta) solves commission stacking automatically out of the box. Every mechanism above requires deliberate configuration — exclusivity has to be assigned code by code or prefix by prefix, and rule-based exceptions have to be written by someone who has already done the detection work in the previous section.
Building a Prevention Workflow That Actually Holds
A prevention mechanism that isn't monitored decays. Coupon codes leak again after a partner shares them with an "affiliate network" of their own; new browser extensions launch faster than blocking tools catch up; and commission structures that made sense at one revenue tier stop making sense at the next. A durable approach treats this as an ongoing operating cadence, not a one-time settings change:
- ◆Monthly code-leakage audits against public coupon aggregators, with immediate code rotation when leakage is confirmed.
- ◆Quarterly new-vs-returning customer review by partner category, feeding directly into commission-tier adjustments rather than sitting in a report nobody revisits.
- ◆Documented exclusivity agreements with top-tier content and influencer partners, formalized in the network's coupon attribution tooling rather than left as a verbal understanding — the technical lock only works if it's actually configured.
- ◆A standing exception list for last-click overrides, reviewed whenever a new coupon or cashback partner applies to the program, so the default isn't "everyone gets last-click credit" but "here are the partners and conditions that get an exception."
The programs that get this right treat coupon and cashback partners as a distinct channel with its own rules, not as a lesser version of content partnerships. The programs that get it wrong pay every partner on the same last-click default and then wonder why their best content creators quietly stop promoting.
Frequently Asked Questions
What is commission stacking in affiliate marketing?
Commission stacking occurs when multiple affiliate partners touch the same customer journey but the program's last-click attribution model pays commission only to whichever partner's link or cookie was active closest to checkout — typically a coupon or cashback site — regardless of which partner actually generated the original demand. The advertiser effectively pays commission on a sale that a different partner already earned through content, reviews, or influence.
How do I know if coupon sites are cannibalizing my other affiliate channels?
Compare new-customer rates by partner category. If coupon and cashback partners show a materially lower new-customer percentage than your content and influencer partners, that's a strong signal they're capturing credit on sales that were already decided rather than generating incremental demand. Pair this with multi-touch or assisted-conversion reports from your network to see how often a content-partner click precedes a coupon-partner click in the same converting session.
Can I stop coupon codes from leaking to unauthorized publishers?
You can reduce leakage significantly but not eliminate it entirely. Awin's exclusive coupon attribution, for example, locks commission for a given code to the assigned publisher even if the code is used through a different or absent affiliate link, and its prefix-based system automates exclusivity for single-use codes at scale. Combine that with monthly audits of public coupon-aggregator sites to catch and rotate any codes that leak despite the technical lock.
Should I ban coupon and cashback affiliates from my program entirely?
Not automatically. Available evidence suggests coupon-driven purchases are more likely to involve new customers and can carry a higher average order value than non-coupon purchases, particularly in price-sensitive categories. A blanket ban risks losing genuinely incremental volume along with the parasitic volume. The better approach is detection first — quantify how much of your coupon-partner traffic is actually incremental — then apply exclusivity, tiered commissioning, or last-click exceptions selectively rather than removing the channel outright.
Which affiliate network handles commission stacking prevention best?
Awin currently offers the most mature native tooling for this specific problem, through its coupon/voucher attribution feature and prefix-based automated exclusivity. Impact and CJ both support rule-based commissioning and multi-touch reporting that can be configured to address stacking, but the setup work — and typically some custom configuration — falls to the advertiser or its agency rather than arriving as a default. No major network solves this automatically without deliberate configuration.
