The Ecomm Analyst

If you’ve ever wondered why your awareness campaigns look like garbage in your reports while your branded search shows a 12x ROAS, last-click is the reason.

Shopify defaults to last-click. So does GA4 for most reports. So does almost every native channel platform’s “view this in your dashboard” experience. The model is simple. Whichever channel touched the customer last gets all the credit for the sale. The blog post that introduced them six weeks ago gets zero. The Instagram ad they saw three times gets zero. The Google search where they typed your brand name gets the full revenue value.

You see the problem.

What last-click rewards

Last-click rewards channels that close, not channels that open. Branded search always looks like a star because it captures people who already know your brand. Direct traffic looks magical for the same reason. Email looks great because customers tend to click an email before checking out.

None of those channels actually created the demand. They just caught it. The channels that created the demand (display, paid social prospecting, organic content, influencers, podcasts) get systematically underweighted.

If you make budget decisions based on last-click reports, you end up cutting the channels that fill the top of your funnel and pouring more money into the channels that harvest existing demand. For a quarter or two, your blended ROAS goes up and you feel like a genius. Then new customer acquisition stalls and you can’t figure out why.

A practical example

Here’s a pattern I see often. (Numbers below are illustrative, not from a specific client.)

A brand spends $40k on Meta prospecting, $15k on Meta retargeting, $8k on Google branded search, $2k on Google non-brand. Last-click attribution shows branded search and retargeting as the top performers. Prospecting looks weak.

If you weight the data differently and look at customer journeys instead of just the closing touch, prospecting is showing up in roughly 60% of the journeys for new customers. It just rarely closes the deal. Cut it, and over the next 8 to 12 weeks branded search volume drops because fewer new people are learning the brand exists.

What I use instead

I don’t pretend that any single attribution model is “the truth.” I do want to see the same data through multiple lenses before I make a budget call.

In ThoughtMetric I keep three views open. Last-click, to see what closes. First-click, to see what opens. And a multi-touch model that distributes credit across the journey. When all three agree that a channel is wasted, I cut it. When they disagree, I look at the journey data and figure out where in the funnel that channel is contributing.

The other thing this gets me is honest retargeting numbers. Retargeting always looks incredible under last-click. Under a model that gives partial credit to the channels that drove the original visit, retargeting’s contribution shrinks to something more honest. That doesn’t mean you turn it off. It does mean you stop scaling it past the point where there’s any incremental demand left to harvest.

When last-click is fine

There are cases where last-click is a defensible default. If your business is mostly bottom-of-funnel (you sell a high-intent product, customers come to you knowing what they want), last-click gives you a reasonable read on which channels close. Replacement parts, B2B reorders, branded merch for an existing audience. In those cases, the close is the whole story.

For everyone else (which is most of ecom), running on last-click alone is how you slowly strangle your own growth.

Bottom line

Look at every channel through at least two attribution lenses before making a budget decision. If you only have one lens, you’re optimizing for the wrong thing.