Our affiliate program has reached a fairly mature and stable stage. That has given me more time to think about how to keep it healthy.
I started reviewing how every publisher in the program was promoting us. As I worked through the data, I noticed some unusual changes in clicks and revenue. A few small coupon sites, in particular, were generating far more sales than we expected.
Not long after that, David from Marcode emailed me. He said their monitoring had found a number of publishers bidding on our branded keywords in violation of our policy. That immediately caught my attention. It led to our first meeting, many more conversations, and eventually our decision to work with Marcode.
Once I began using the tool, I realized how many publishers were running branded search ads on Google that made them look like the official brand. That was when I started cleaning up the publisher base.
We have now worked with Marcode for six months. During that time, we found publishers running ads that we would never have thought to check. We reviewed the evidence, removed the violators from the program, and kept the publishers that were providing genuine value. The cleanup also exposed another problem: we had overlooked some publishers with real potential.
A few days ago, I invited David and Erica Li to an online discussion about this topic.
Erica Li previously managed UGREEN's affiliate program and has developed her own approach to affiliate compliance. David has worked with brands in Europe and the US, as well as many affiliate managers at Chinese companies. He has seen the same pattern repeatedly: brands know that noncompliant traffic exists in their programs, but their affiliate managers still struggle to act on it.
Most teams already understand the risks. They know that hijackers and brand bidders can weaken an affiliate program. The harder problem is acting on that knowledge when the commercial incentives point in the other direction.
Many affiliate managers at Chinese companies carry a channel revenue target or are expected to maintain the affiliate channel's share of total revenue. If they suddenly remove publishers that are generating sales through prohibited ads, the affiliate platform will show an immediate drop. Internally, that drop can lead to difficult questions from management. The cleanup may be the right decision and may improve profit, but the revenue report can make it look as if the manager is underperforming.
This tension came up throughout our discussion. I wanted to share what we learned and the practical steps we think other affiliate managers can take.
Affiliate platforms report attributed revenue
We spent more time on one question than any other: which data should a team use to judge the impact of a cleanup?
Conversion data from an affiliate platform matters. It tells us how many clicks, orders, and commissions a publisher generated. Comparing data from the period when a violation was active with data after the publisher was removed can give us a rough estimate of how much affiliate attribution was tied to brand bidding or hijacking.
But that data only shows how many orders those publishers claimed within the affiliate channel. It does not tell us whether removing them reduced the brand's total revenue. To answer that question, we also need to look at Shopify, GA4, and data from other channels.
Brand bidders usually intercept users who already intend to buy. Someone searching for a brand name is likely to reach the site through paid search, organic search, or direct traffic anyway. When a publisher stops bidding, those orders do not necessarily disappear. They may simply return to the channel that would otherwise have received the attribution, while the company no longer pays an affiliate commission on them.
For that reason, we compare two sets of numbers before and after a cleanup. The first comes from the affiliate platform and shows changes in affiliate-attributed revenue and commission. The second comes from Shopify, GA4, or the company's order system and shows total site revenue, order volume, and changes across paid search, organic, direct, and other channels.
If affiliate-attributed revenue falls while total orders remain broadly stable and ROAS improves, the decline should not automatically be treated as lost business. It suggests that some of the revenue previously credited to affiliates was not creating enough incremental value.
The comparison period matters as well. Whenever possible, avoid major promotions, stockouts, price changes, and site outages. If those events cannot be avoided, account for them in the analysis. Otherwise, too many variables move at once, and it becomes difficult to tell what the cleanup actually changed.
The longer you wait, the harder the internal conversation becomes
Brands often ask whether monitoring should begin on the first day of an affiliate program or only after the program reaches a certain size.
David's view is that brands should start as early as possible. Day One is ideal.
A new program with only a few publishers and little revenue does not necessarily need to buy a dedicated monitoring tool immediately. The team can start by manually checking its most important branded keywords, such as the brand name. Rotate countries, devices, search engines, and times of day. Search the brand terms on Google and review the results. New publishers can also be spot-checked from time to time. Once the program starts generating meaningful revenue, the team can increase the frequency and coverage of its monitoring and decide whether a specialist tool is justified.
Starting early has a clear advantage: noncompliant revenue has not yet entered the team's KPIs or the company's growth forecast. From the beginning, the team can explain to management that the affiliate program is meant to find partners who bring in new customers and create useful content. Publishers that make the program look stronger on paper by intercepting existing demand should not be admitted in the first place. Revenue generated solely by capturing branded search traffic should not be counted as program growth.
Cleaning up a mature program is much harder. If a publisher has contributed substantial attributed revenue for several months, removing it creates two uncomfortable questions. Why was the issue not caught earlier? And why has reported affiliate revenue suddenly fallen? The affiliate manager has to manage both the data and the internal conversation around it.
If your program is already established, a phased cleanup is usually easier to manage. This is also the approach I am using now.
Start with hijackers. They directly interfere with the user's original path to the site and can do more damage to the brand experience. Once those cases are handled, move to publishers bidding directly on prohibited branded keywords. Publishers with less impact or relatively minor violations can be addressed in later rounds, which reduces the disruption caused by removing everyone at once.
After each round, track affiliate-attributed revenue, total site orders, commission, and ROAS. The team can then redirect the commission it saves toward new placements, strong content partners, and publisher recruitment.
The affiliate numbers may look worse in the short term. Over a longer period, however, more of the budget goes to publishers that are willing to explain the product, create content, and protect the brand. That gives the program a healthier base for further growth.
Amazon needs a separate review process
Most affiliate compliance conversations focus on DTC programs. In our discussion, we also looked at Amazon's affiliate ecosystem, especially Amazon Attribution, which is easy to miss in a standard compliance workflow.
Amazon Associates and Amazon Attribution use different types of links. Amazon pays the commission for Associates, while brands need to pay closer attention to the Attribution links for which they bear the cost. The redirects and tracking parameters also differ from those used by common affiliate networks such as Awin and Impact. The point is to make sure your monitoring covers Amazon links to the same standard as your network links - not to treat Amazon as a blind spot you live with.
Some platforms that support Amazon affiliates leave recognizable letters in their UTM parameters, such as LV or PB. Those parameters can provide a useful lead. If the advertiser name or promotional domain is already known, the team can also filter by name and domain, see who is bidding on which keywords, and then investigate where the link came from.
Network compliance tools are a starting point
After speaking with Chinese brands, agencies, and consultants, David has noticed that awareness of affiliate fraud is growing. There is still a wide gap between knowing about the problem and building a process to deal with it.
Some brands assume that their affiliate network, such as Impact or CJ, is already handling compliance because it offers monitoring features. Those platforms are primarily built for tracking, payments, partner discovery, and program management. Brands should not expect them to manage every type of affiliate violation.
There can also be an incentive conflict. When revenue attributed to the affiliate channel increases, the platform may earn more in platform fees. It is unrealistic to ask the network to identify and resolve every violation on the brand's behalf.
The network's features can still provide a basic layer of coverage. Impact, for example, includes a limited allowance for keyword monitoring. Brands can combine those features with manual checks, then decide whether to add external monitoring based on program revenue and risk.
Compliance starts with publisher approval
Our friend Janus, who also works as an affiliate manager, takes a strict approach to reducing violations. When he inherits a new program, he approves only publishers whose identity and value he has verified. He rejects most applications with unclear sources or business models. The trade-off is slower revenue growth at the beginning, but it can prevent a large publisher review later.
Erica and I both think this whitelist approach works well for new programs. Approval is not permanent, though. Publishers can change their traffic sources or hand promotion to new sub-partners. Admission checks, spot checks, and periodic reviews need to continue throughout the life of the program.
The program policy should also be updated on the affiliate page and in the contract. Many older policies ban brand bidding and hijacking but say nothing about browser extensions, subnetworks, or AI-generated content.
Brands should state which branded keywords and misspellings are prohibited, whether brand-plus-generic keywords are allowed, and whether publishers may use coupon codes, toolbars, or browser extensions. The policy should also say whether publishers must disclose downstream partners and what will happen when the brand confirms a violation.
Clear terms only work if the brand enforces them consistently. If two publishers commit the same violation but only one receives a warning, publishers will quickly conclude that the policy is not taken seriously. Once the evidence has been verified, the brand can respond according to severity with a warning, a deadline to correct the issue, withheld or suspended commissions, removal from the program, or legal escalation. The team should document the process and allow publishers to appeal or request a review when the evidence is unclear.
What affiliate managers can do now
If your program is still new, start with these steps:
- Review each publisher's website, main traffic sources, and promotional methods. Do not approve every application just to increase the publisher count. If you do not recognize a publisher, email them and ask how they plan to promote the brand.
- Add brand bidding, hijacking, browser extensions, sub-affiliate activity, and AI-generated content to the program terms. Define what is permitted and how violations will be handled.
- Manually check your main branded keywords. Rotate markets, devices, search engines, and times of day.
- Save the ad page, time of discovery, location, redirect chain, and affiliate ID. Review every alert before taking action.
- Compare affiliate platform data with total site data before and after enforcement. Track commission, orders, ROAS, and changes in other channels.
- Redirect the budget you save toward partners that create incremental value, then measure the results of those placements, content partnerships, and recruitment efforts.
After this discussion, my view of affiliate violations became much clearer. Compliance is partly about traffic rules, but it also gives us a reason to reassess what the affiliate program is actually contributing.
The publishers we keep should be able to reach new customers, explain the product properly, and build content and trust that the brand can benefit from over time.
