Why your best last-click channel might not be your best channel
A budget review runs the same way in a lot of retail marketing teams. The channel report comes back, and it favors whichever channel sits closest to the sale. Paid search and retargeting post strong returns again this quarter. The channels further back, the ones that first put the brand in front of someone, show thinner numbers by comparison. Finance wants the growth story to hold, and the fastest way to defend it looks like doubling down on whatever the report already credits. That instinct is usually wrong, because the report only sees the last click, not what led to it.
Last-click attribution gives most or all of the credit for a sale to the final tracked interaction before someone buys, usually a paid search click. It became the default measurement method because it’s the easiest number to pull. Most ad platforms report it natively, without a brand needing to stitch together data across channels. A shopper first sees a brand on TikTok. Days later, she returns to search for it by name, clicks a paid search ad, and buys. Paid search gets the full credit. The TikTok activity that likely created the demand behind that search gets none of it.
The same last-click number can justify two opposite, and equally wrong, decisions.
Cut TikTok or a channel like Demand Gen for its weak last-click showing. The risk is losing the activity that made other channels’ sales possible in the first place. Keep funding paid search for its strong last-click showing, and the risk flips. A brand may just be paying it to close sales that other activity already set up.
Most marketers know the metric is flawed but keep using it anyway. A Snap and EMARKETER survey of 282 senior US marketers found 78.4% still relied on last-click or web analytics as their main method. Fewer than one in four were confident it gives an accurate read on a channel’s long-term impact. Nearly three in four were already moving away from it, or wanted to. The gap between what marketers believe and what they measure by is exactly what lets a flawed number keep driving real budget decisions.
A fuller measurement approach, often called full-funnel measurement, separates two questions that last-click collapses into one. Attribution decides which channel gets credit for a sale that already happened. Contribution asks what that channel, or any other, actually added to the outcome, whether or not it happened to be the last click.
In practice, that means pairing modeled data across the whole customer journey with incrementality tests. Those tests isolate what actually changes in sales when a channel’s spend moves up or down, rather than relying on click paths alone. That distinction is what lets a brand credit TikTok for the demand it built, even when paid search closes the sale.
Fospha’s Peak Playbook 2026 points to the same pattern in retail. Fospha tracked retail brand spend across all four quarters of 2025. The strongest-performing brands, the top 25% by ROAS, spent close to twice as much on upper-funnel activity as the rest of the market. That’s the Awareness and Consideration end of the funnel. The gap held steady in every quarter, which points to a standing budget choice rather than a seasonal push.
That’s a correlation. It doesn’t prove the upper-funnel spend alone caused the stronger returns. Brands with more room in the budget may simply be the ones able to sustain awareness spend year-round. But a full-funnel view is what surfaces the pattern at all. A last-click report would show none of it.
The shift is in the question asked before money moves. Instead of asking which channel closed the most sales last quarter, ask what each channel contributed across the full path to purchase. That might mean weighing a channel’s role in assisted conversions, or running a controlled incrementality test before cutting its budget.
A channel that can answer the contribution question earns its next dollar. One that can only point to its last-click credit hasn’t answered it yet.
Leave a Reply
You must be logged in to post a comment.