Attribution Is Broken. Provisionary Economics Fixes It at the Root.
The Honey scandal did not create the attribution crisis. It revealed one that has always been there. Here is why last-click attribution was always going to fail — and what a structurally sound alternative looks like.
In December 2024, a YouTuber named MegaLag published a video that detonated across the affiliate marketing industry. His investigation revealed that Honey — PayPal’s browser extension, downloaded by 17 million users and endorsed by hundreds of creators — was systematically replacing legitimate affiliate cookies with its own at the moment of checkout. Commissions earned by creators who had researched, reviewed, and genuinely influenced purchase decisions were being quietly redirected to PayPal. No one was notified. No one was compensated. The work was real. The theft was structural.
The industry reacted with outrage. Class action lawsuits were filed. YouTube’s largest creators — LegalEagle, Wendover Productions, and others — joined the action. Reddit’s affiliate marketing communities erupted. The story dominated every industry publication for months. But amid all the anger, an uncomfortable truth was being avoided: Honey did not create this problem. Honey exploited a vulnerability that the industry had built into its own foundation. The last-click attribution model — the mechanism that determines who gets paid — was always a single point of failure waiting to be gamed.
Understanding Last-Click: The Flaw by Design
Last-click attribution is simple: whoever owns the final click before a purchase gets the commission. It does not matter how many creators researched the product, how many reviews influenced the decision, or how long the consumer considered the purchase because of content they consumed weeks earlier. The prize goes to the entity that occupied the checkout moment. That is it. That is the entire logic.
This model made sense in 2002. The internet was linear. A consumer saw a banner ad, clicked a link, and bought. Attribution was genuinely straightforward. But the internet evolved into something incomparably more complex. Today, a consumer might discover a product through a TikTok review on Monday, research it through a YouTube deep-dive on Wednesday, see a Reddit thread validating the purchase on Friday, and finally click a coupon extension while buying on Sunday. Under last-click attribution, the coupon extension earns the commission. The TikTok creator, the YouTuber, and the Reddit community that did every ounce of actual persuasion earn nothing.
Last-click attribution does not measure influence. It measures proximity to the buy button. Those are not the same thing. They have never been the same thing.
The numbers quantify the distortion. Global influencer marketing spend surpassed $30 billion in 2025, yet the overwhelming majority of that spend is measured against a last-click model that systematically fails to capture creator-driven value. Research from Bain and Company found that roughly 80 percent of consumers rely on zero-click results — AI overviews, social recommendations, LLM responses — in at least 40 percent of their searches. These influence points generate no trackable click. They generate conviction. And last-click attribution is incapable of measuring conviction.
The Structural Rot Beneath the Scandal
What Honey did was brazen. But it was not anomalous. Capital One Shopping operates on the same mechanism. Rakuten Rewards operates on the same mechanism. Any browser extension with checkout-moment access can inject itself into the final click and claim the commission — legally or otherwise. The last-click model creates an economic incentive to position yourself at the bottom of the funnel and harvest commissions generated by people above you. It rewards harvesting, not growing. It compensates presence at checkout, not influence at discovery.
The broader consequence is a systematic degradation of the affiliate ecosystem. When creators discover that their upper-funnel work — the reviews, the honest recommendations, the trusted audiences they spent years building — is routinely absorbed by lower-funnel actors at checkout, they face a binary choice: abandon the channel or become a lower-funnel actor themselves. Both choices damage the ecosystem. Brands lose the authentic discovery content that actually generates new customers. The industry fills with coupon sites and cashback platforms competing for the last click. Trust collapses from all directions simultaneously.
When you reward the last click, you defund the entire journey that made the last click possible. The industry has been paying for the harvest while starving the farmers.
Why the Proposed Fixes Don’t Fix Anything
The industry’s response to the attribution crisis has been earnest but structurally insufficient. Multi-touch attribution models attempt to distribute credit across multiple touchpoints in the conversion path. Data-driven attribution uses algorithmic weighting to assign value across the consumer journey. Server-side tracking replaces cookie-dependent measurement with more resilient infrastructure. These are improvements. They are not solutions.
Every one of these approaches attempts to more accurately measure influence after the fact. They are all still operating within the same fundamental architecture: an affiliate generates influence, a consumer travels an uncontrolled journey across an uncontrolled internet, and somewhere downstream, a measurement system tries to reconstruct who deserves credit for what. The measurement problem is real. But it is a symptom. The root cause is a model that attempts to assign credit for influence that was never captured in the first place.
No attribution model can accurately credit a TikTok video that convinced someone to buy if the consumer never clicked the affiliate link. No server-side tracking solves the zero-click discovery problem. No multi-touch framework compensates a creator for the trust they built in an audience that converted through someone else’s checkout link. These tools make the broken model slightly less broken. The model itself remains broken.
The Provisionary Economics Approach: Attribution Before the Journey
Provisionary Economics does not attempt to measure influence after a purchase. It captures and confirms the affiliate’s commercial relationship before the purchase is solicited. This is the structural difference, and it changes everything downstream.
Within this framework, an affiliate’s participation in growing a discount instrument establishes a protected attribution relationship at the moment that relationship is created — not at the moment of transaction. The commercial signal that links an affiliate to a consumer is recorded, confirmed, and protected before the consumer ever arrives at a checkout page. There is no checkout-moment cookie to intercept. There is no last click to steal. The attribution is not a footprint left on a consumer’s browser that any extension can overwrite. It is a confirmed commercial record established upstream of the transaction, within an architecture the affiliate participates in building.
The attribution window in this system is extended and structurally protected — not because a cookie was set with a longer expiry, but because the affiliate’s connection to the commercial outcome was established before the consumer’s uncontrolled journey across the internet could erode it. No browser extension, retargeting platform, or coupon site can arrive at checkout and claim prior credit for a relationship that was already confirmed in the system.
| The Attribution Problem | Last-Click Model | Provisionary Economics |
|---|---|---|
| When is the affiliate’s relationship established? | At the final click before checkout | Before the purchase is solicited |
| Can a browser extension override attribution? | Yes — by injecting the last click | No — attribution is upstream of checkout |
| What is the affiliate compensated for? | Proximity to the buy button | Confirmed commercial relationship |
| Does zero-click discovery go uncompensated? | Yes — no click, no credit | No — relationship is established independently of click tracking |
| Can coupon sites harvest upper-funnel work? | Yes — routinely and at scale | No — lower-funnel actors cannot override pre-established attribution |
What This Means for Every Affiliate Reading This
If you have spent years building an audience that trusts you, creating content that genuinely influences purchase decisions, and watching coupon extensions collect commissions on sales you created — you are experiencing the cost of a broken architecture, not a personal failure. The last-click model was not designed to reward the work you do. It was designed for a commercial internet that no longer exists.
The affiliate industry’s reckoning with Honey is healthy. But it is also incomplete. Filing class actions against PayPal addresses the most egregious symptom while leaving the underlying pathology untouched. The problem is not Honey specifically. The problem is any commercial architecture that allows the last actor in the funnel to claim credit for the entire funnel. Fixing that requires changing the architecture — not patching the measurement layer on top of it.
Provisionary Economics is a different architecture. Not a different measurement tool applied to the same broken system — a different system entirely, built on the premise that the affiliate’s commercial relationship should be established and protected before the consumer’s uncontrolled journey begins. Commission is tied to confirmed transactions. Attribution is protected from the moment it is created. The work you do to bring a consumer into a commercial relationship is the work that gets compensated — not the work of whoever happened to be standing at the door when the consumer finally arrived.
The Larger Point
The Honey scandal forced an industry-wide conversation that was long overdue. Attribution is broken. The mechanisms that determine who gets paid for influencing a purchase are systematically misaligned with the actual work of influencing a purchase. That misalignment costs legitimate affiliates billions of dollars annually, degrades the quality of the affiliate ecosystem, and ultimately harms the brands and consumers the system is supposed to serve.
The conversation the industry is now having is about how to measure influence more accurately within an existing model. The conversation that needs to happen is about whether the existing model is worth preserving at all — or whether the conditions that make accurate attribution perpetually elusive are structural features of last-click commerce that no measurement improvement can eliminate.
Provisionary Economics begins from the latter premise. The attribution problem is not a measurement problem. It is an architecture problem. And architecture problems require architectural solutions.
Provisionary Economics is the foundational framework of Voulay — a three-sided commerce ecosystem where affiliate attribution is established before the consumer journey begins, not reconstructed after it ends.