The Last Click Gets the Credit. That Doesn't Mean it Did the Work.

Why marketers need to separate attribution from influence.

By Zahava Robinson, Director & Chief Technology Officer, Kick Cashback.

Marketers want clean answers. We want to know which channel drove the sale, which campaign deserves the budget and which partner should be paid. The problem is that customer journeys are rarely as clean as the reporting makes them look.

A shopper might first discover a product through a creator, search for the brand a few days later, compare prices, read reviews, visit a cashback site, click a coupon link and then complete the purchase after returning through another channel. By the time that transaction appears in a dashboard, a complicated journey can be reduced to a single line of attribution: the last click.

That is useful operationally, but it can also be misleading. Building in the cashback and affiliate ecosystem has given me a close-up view of how often the channel that receives the credit is not necessarily the channel that created the intent. The final click may have helped close the loop, but that does not mean it did all of the work.

Attribution records an event. It does not always explain influence.

Last-click attribution became popular for a good reason. It is simple, relatively easy to reconcile and gives advertisers, publishers and affiliate networks a clear basis for assigning a conversion. Someone clicks a link, makes a purchase, and the system attributes that transaction to the click.

The difficulty comes when we start treating that technical event as a complete explanation of why the customer bought.

Imagine someone first sees a product in a creator's video. That creator creates awareness and introduces the brand. A few days later, the customer searches for the company, looks through the website and leaves. Two days after that, they decide to purchase, search for a discount, click through a loyalty or cashback platform and then complete the transaction.

In a last-click model, the cashback or loyalty platform may receive the attribution. That platform may well have created value. It may have increased the likelihood of conversion, encouraged the customer to spend more, brought them back sooner or stopped them abandoning the purchase. Those are meaningful outcomes, but they are not the same as originating demand.

That distinction matters because if marketing budgets are allocated almost entirely according to last-click conversions, channels that operate near checkout can look disproportionately valuable while the channels that created the initial interest appear less effective than they really were.

The closer you are to checkout, the easier it is to look efficient.

There is a structural advantage to operating late in the funnel. Coupon sites, loyalty platforms, cashback services, browser extensions, branded paid search and other conversion-focused channels often interact with consumers when they are already close to purchasing. As a result, they can produce excellent conversion rates.

That does not mean those channels are unimportant. In many cases they are extremely effective. A well-timed incentive can tip a customer from consideration into purchase, shift spend from one retailer to another or increase basket size. But conversion rate alone does not answer the bigger commercial question: would that customer have purchased anyway?

The same issue appears in paid search. If a consumer already knows the brand and searches for it by name, the resulting paid click may look highly efficient. But the search ad may not have created the demand. It may have simply captured a customer whose intent was built elsewhere.

The reporting often rewards the channel closest to the point where the transaction becomes measurable. That can make the final interaction look more important than the earlier interactions that actually shaped the decision.

Customers see one journey. The industry sees competing claims.

There is another part of this problem that becomes especially obvious in cashback and affiliate marketing: customers generally have no idea how attribution works behind the scenes.

A shopper does not think about affiliate cookies, attribution windows or publisher hierarchies. If they click through a cashback platform, they expect cashback. If they use a creator's code, they assume that creator receives credit. If they found the product through a particular site, they assume that site was responsible for the sale.

Behind the scenes, however, the commercial reality can be much more complicated. A later click may overwrite an earlier one. A coupon extension may become the final attributed partner. Another marketing channel may take precedence. A retailer may determine that a different source was responsible for the transaction.

The consumer experiences one purchase journey, while the systems behind that purchase may be determining which of several channels gets to claim it. That gap creates a trust problem.

From the customer's perspective, discounts, rewards and incentives appear to sit neatly on top of the purchase. In reality, those mechanisms can be competing with each other for attribution. When that creates a mismatch between what the customer expected and what the attribution system records, the consumer-facing platform is often left trying to explain a system the customer never knew existed in the first place.

Attribution models also shape behaviour.

Measurement systems do not just report behaviour. They influence it.

If marketers disproportionately reward the final click, publishers and platforms naturally have an incentive to position themselves as close as possible to the moment of conversion. Sometimes that produces genuine value. Better offers, stronger loyalty products and smoother checkout experiences can all improve conversion. But it can also create incentives to capture attribution rather than create incremental demand.

A partner that introduces thousands of new customers to a brand may appear less efficient than a partner that interacts with customers in the final few minutes before checkout. Looking only at reported conversions may make the second partner appear more valuable. Looking at new customer acquisition, incrementality and the broader path to purchase may tell a very different story.

That is why attribution should not be treated purely as an accounting exercise. At its core, it is a question about causality: did this channel introduce the customer to the brand, influence the decision, bring the customer back, increase basket size, prevent abandonment, or simply happen to be the last interaction before payment? Those are different forms of contribution, even when an attribution platform records them all under the same heading: conversion.

The answer is not to get rid of last-click attribution.

Last-click remains useful. It is deterministic, relatively easy to administer and gives advertisers and partners a practical method for reconciling sales and payments. The problem is not last-click itself. The problem is treating it as though it tells the whole story.

Marketers need to use it as one signal among several. That means looking at new versus returning customers, assisted conversions, customer acquisition cost, basket size, time between first interaction and purchase, repeat purchase behaviour and, where possible, genuine incrementality.

Controlled testing is particularly valuable. What happens when a channel is removed from part of an audience? Do total sales fall, or does attribution simply shift somewhere else? Does a cashback incentive create additional purchases, increase basket size or accelerate conversion, or does it mostly attach itself to purchases that were already going to happen? Does a creator introduce valuable new customers who later convert through channels that receive the final attribution?

Those questions are harder to answer than asking which link was clicked last, but they are far more useful when deciding where marketing dollars should go.

We need to become more comfortable with imperfect answers.

Marketing measurement has never been perfect, and privacy changes, cross-device behaviour, fragmented customer journeys and increasingly complex technology stacks have only made attribution harder. Yet dashboards can still create an illusion of certainty.

A transaction appears. A channel is listed next to it. It is tempting to assume that channel explains the sale.

Sometimes it does. Sometimes the decisive interaction happened much earlier. Sometimes several channels genuinely contributed. Sometimes the final click was the reason the customer converted. And sometimes it was simply the last thing the customer touched before paying.

The challenge for marketers is not to find a perfect attribution model. There probably is not one. The better challenge is to be more precise about what our measurement systems are actually telling us.

Last-click attribution can tell us where the transaction finished. It cannot always tell us where the sale began.

About the author

Zahava Robinson is Director and Chief Technology Officer of Kick Cashback, a Melbourne-based cashback and affiliate technology platform. She works directly across attribution, consumer rewards and the operational systems that sit between retailers, affiliate networks and customers.