Last-Click Attribution Is Broken. Here's Why Brands Keep Using It Anyway.

Last-click attribution is one of the most widely used measurement approaches in marketing. It's also one of the most misleading.

And yet, most brands still use it as their primary source of truth for budget decisions. Here's why that happens — and what the data actually tells us.

What Last-Click Attribution Says

Last-click attribution is simple: whichever channel a customer touched immediately before converting gets 100% of the credit for that sale.

Customer sees a Facebook ad. Sees a YouTube pre-roll. Gets a retargeting ad. Clicks a branded paid search ad. Converts.

Paid search gets the credit. Every other channel gets zero.

The Problem With That Logic

Think about what that model is actually saying.

It's saying that the Facebook ad, the YouTube view, and the retargeting impression had zero influence on the purchase decision. That the customer would have converted at exactly the same rate without them. That the only thing that mattered was the final click.

That's not measurement. That's a convenient fiction that happens to favour bottom-of-funnel channels.

What the Data Actually Shows

When brands run proper measurement — Media Mix Modeling, incrementality testing, controlled experiments — the results consistently contradict last-click reporting.

TV drives online search volume. Brands that run TV consistently see correlated spikes in branded search and direct traffic. Last-click gives paid search the credit. MMM attributes it correctly to TV.

Upper-funnel spend creates demand that bottom-funnel captures. Paid social, display, and video build awareness that eventually converts through paid search or direct. Attribution tools see the conversion. They don't see what caused it.

Retargeting overclaims massively. Retargeting ads reach people who were already highly likely to convert. Last-click attributes those conversions to retargeting. In reality, many of them would have happened anyway. Incrementality testing routinely shows retargeting ROAS 40–70% lower than reported.

Why Brands Keep Using It

If last-click attribution is so flawed, why does almost every marketing dashboard still use it as the default? Three reasons:

1. It's easy to implement

Last-click requires nothing more than a tracking pixel and a standard analytics setup. Any marketing manager can access it without a data science team.

2. It tells a clear story

"Paid search drove $2M in revenue last quarter" is a simple, defensible number. The truth — a complex web of channel interactions, time lags, and attribution uncertainty — is harder to present to a board.

3. It confirms existing biases

Performance marketing teams are measured on channel-level ROAS. Last-click attribution produces channel-level ROAS numbers. It reinforces what people already want to believe about their channels.

The problem isn't that marketers are lazy or dishonest. It's that last-click attribution is the path of least resistance — and nobody gets fired for using the default.

The Real Cost of Getting This Wrong

Budget decisions made on last-click data systematically produce the same outcome: brands overspend on bottom-of-funnel channels and underspend on brand and upper-funnel.

In the short term, this looks fine. ROAS numbers stay strong because you're efficiently capturing demand that already exists.

In the medium term, demand starts to dry up. There's less new awareness entering the funnel. Branded search volume plateaus. Customer acquisition costs rise.

By the time the problem is visible in the data, the damage has been compounding for 12–18 months.

What to Do Instead

You don't have to throw out attribution tools. They're still useful for in-campaign optimisation and day-to-day reporting.

But for budget allocation decisions — especially for annual planning or any decision involving significant spend shifts — you need a measurement approach that captures the full picture. That means:

• Media Mix Modeling for understanding true channel contribution at the aggregate level
• Incrementality testing for validating specific channels or campaigns
• Brand tracking for measuring upper-funnel impact that attribution tools miss entirely

The Question Worth Asking

If you added up the revenue attributed to every marketing channel in your current reporting — paid search, paid social, email, affiliates, display — would it exceed your actual total revenue?

For most brands, the answer is yes. Sometimes by 2x or more.

That's not a data quality problem. That's a measurement problem. And it's the problem that's quietly distorting your budget decisions every quarter.