July 30, 2026
Affiliate Marketing Runs on Trust. CJ’s Estimated Conversions Put That Trust to the Test.
Affiliate marketing has one promise at its center: advertisers pay when performance happens. I have built a business on that promise, and it is the reason brands trust this channel with real budget.
That promise is now being tested.
CJ’s new reconciliation approach charges advertisers monthly for orders that were, in CJ’s words, “not tracked due to a tracking gap.” CJ defines that gap as the estimated percentage of commissionable activity its systems did not capture because of integration, consent, or technical limits. The example in CJ’s own documentation is straightforward. If an advertiser has $100 in tracked commissions and CJ estimates a 20% gap, that $100 is treated as 80% of expected activity. Estimated total: $125. Reconciliation charge: $25.
The math is simple. The logic behind the estimate is opaque. And the governance is an open question.
Tracking gaps are real. What we do about them is the actual question.
Tracking is harder than it was three years ago. Consent rules, browser restrictions, and the zero-click environments inside the LLMs all create real measurement gaps. No serious operator in this space should pretend otherwise.
That trend is not reversing. Which is exactly why the standard for billing on an estimate has to get tighter, not looser. The question was never whether gaps exist. It is what a platform is allowed to do about them.
When an estimate becomes an invoice, the trust model changes.
For as long as this channel has existed, a commission has pointed at something. A click, an order, an ID you can pull up. Attribution has never been clean, and brands, networks, and publishers have argued about the rules forever. But the argument was always about which action earned the commission. Not whether a transaction happened.
A reconciliation model based on estimated untracked activity introduces a different standard. It may be intended to compensate publishers for real value that imperfect tracking missed. But if the advertiser is being charged for modeled activity, the advertiser deserves transparency into the model.
That means more than a summary calculation. Advertisers should see the assumptions behind the gap percentage and the evidence behind partner-level allocation. They should also know whether the estimated publisher commissions are actually paid out, and whether platform fees are charged on activity nobody can point to.
Right now the advertiser sees the bill before they see the proof.
A tracking gap is not the same as an owed commission.
A tracking gap is a measurement problem. It should not automatically become proof that a specific commission was earned.
That distinction matters. Some transactions genuinely are not captured, because tracking is imperfect. It is a much riskier leap to bill an advertiser for estimated activity as if those conversions were known.
I want to be careful here. I am not arguing that modeling has no place in performance marketing. It does, and it will have more. But there is a difference between modeling for insight and modeling for billing. One informs a decision. The other takes money out of a brand’s budget. The second one needs a higher bar: published methodology, partner-level backup, and someone outside the platform checking the math.
When our team asked how the gap percentage is calculated so a brand could forecast the cost, the answer pointed us back to the support article. I do not read that as bad intent. I read it as a policy that shipped ahead of the transparency it needs. That is the part I would fix first.
Three questions to ask before the first charge posts.
There is a long list of questions worth asking here. Three of them are the ones a finance team will escalate on.
How is the gap percentage calculated, and how often does it change? Without that, the cost cannot be forecast, and an unforecastable cost becomes a budgeting problem long before it becomes a measurement debate.
Are platform fees charged on the estimated activity? If they are, that should be stated plainly, because the platform calculating the estimate also earns on it. Not sinister. It just needs to be in the open.
Can an advertiser dispute the amount before it locks? A charge with no dispute window is not a reconciliation. It is a deduction.
None of that is adversarial. It is how every other line in a budget gets reviewed.
Affiliate leaders are already defending this channel’s margin internally. Finance teams are forecasting against it. A platform-controlled charge that cannot be explained to either group becomes a problem long before it becomes a measurement debate.
This is about the policy, not the partnership.
I am not looking for a fight with a network. My concern is the standard that gets set when one of the largest platforms in the channel begins billing from a model.
Performance marketing works when the people funding it can understand what they paid for. If we drift toward reconciliation that nobody outside the platform can check, we give up the one thing that made affiliate different from every other line on the media plan.
What PartnerCentric believes
From where I sit, this is not a reporting issue. It is a leadership issue for every brand funding affiliate as a performance channel. Here is what we believe, plainly.
- Performance marketing has to stay accountable to transaction-level evidence.
Attribution rules have always been debated, and they should be. But the connection between a commission and an identifiable action is what separates this channel from everything else on the media plan. - Estimated gaps should not become blank-check billing.
If a platform estimates untracked activity, advertisers need the data inputs, the partner-level allocation logic, and the fee treatment before the charge posts. - Any model that creates advertiser charges should be independently auditable.
Models that move money should not police themselves. A neutral validation layer protects everyone in the transaction, including the network. - CJ has been a strong partner to this business, and this is a policy disagreement, not a verdict on the company.
The critique is of the policy, not the people or the company. Our job is to tell clients when a policy needs a higher standard, and this one does. - We hold ourselves to the same standard we are asking for.
Through Fuse Precision and our broader measurement work, we have spent years arguing that visibility beats assumption. A harder measurement environment is not an argument for more black-box math. It is an argument for better proof.
Shoppers need proof before they trust a recommendation. Advertisers need proof before they accept a charge. Same principle, and it is the one I would not trade away in this channel.
What I would ask for
The channel does not need to pretend tracking is perfect. It is not, and it will not be. But imperfect measurement does not get fixed by asking advertisers to accept numbers they cannot check.
Published methodology. Individual transaction-level reconciliation. Partner-level backup. Independent validation. Caps that prevent surprise. A dispute window before amounts lock. None of that requires abandoning modeled reconciliation. It requires making it reviewable.
Because the day an affiliate leader cannot answer “what did I pay for?” is the day this channel stops being different. That is the part worth protecting in the first place.
Frequently asked questions
What is CJ’s reconciliation charge?
It is a monthly charge to advertisers for commissionable activity CJ estimates its systems did not capture. CJ applies an estimated tracking gap percentage to tracked commissions and bills the difference. Using CJ’s own example, $100 in tracked commissions with an estimated 20 percent gap becomes a modeled $125 total and a $25 reconciliation charge.
Are platform fees charged on the estimated activity?
Yes. CJ has confirmed that CJ fees (CJ fees) are applied to the estimated transactions. That means the platform producing the estimate also earns on it, which makes the true cost of the policy higher than the reconciliation line alone. I do not read that as bad intent, but it is exactly the kind of detail that belongs in published documentation rather than in a support conversation.
Are tracking gaps real, or is this manufactured?
They are real. Consent requirements, browser restrictions, and zero-click discovery inside LLMs all create genuine measurement loss, and that trend is not reversing. The disagreement is not about whether gaps exist. It is about what a platform is permitted to bill for based on an estimate it controls.
Isn’t this just modeling, which the rest of marketing already uses?
Modeling for insight and modeling for billing are different things. One informs a decision. The other moves money out of a brand’s budget, which requires a higher bar: how the gap percentage is derived, transaction-level reconciliation, partner-level backup, fee treatment stated up front, and validation by someone outside the platform. Modeled reconciliation is not the problem. Unreviewable modeled reconciliation is.
What should advertisers do now?
Ask three things before the charge posts. How is the gap percentage calculated, and how often does it change? Whether platform fees apply to the estimated activity, and at what rate? Whether the amount can be disputed before it locks? Then track the charge month over month at the account level so an unexplained movement gets caught in the same cycle it appears, not two quarters later.
Ready for better?
Get In Touch