Attribution After Last Click: Measuring Paid Media in a Zero-Click World

On paper, performance looks strong. Year over year, ROAS is up, CPMs are competitive, and multi-touch reporting paints a positive picture. But zoom out and something does not add up. For a number of brands we work with, direct and organic search revenue — channels traditionally buoyed by performance marketing — are trending down at the same time.

Two things are happening at once, and they compound. Ad delivery is drifting toward older audiences. And the younger audiences being reached less often are the ones least likely to click at all. Together they produce revenue that is real, but lands in the wrong bucket or in no bucket.

The demographic drift

Across multiple accounts, the same pattern shows up: ad delivery skews older, year over year.

Where a year earlier we were reaching 25-to-45-year-olds at healthy frequency, Meta began serving more impressions to 45+, in some cases into the 55-to-65+ range, often with no targeting changes on our side. CPMs rise in those older cohorts because competition there is heavier. Delivery to younger demographics quietly falls away.

The mechanism is straightforward: Meta optimizes toward click signals. Younger users, Gen Z in particular, are increasingly a zero-click audience. They do not interact the way older cohorts do. They see an ad, screenshot it, and search later.

The screenshot economy

This is the behavior that breaks the model. A user sees an ad in feed, takes a screenshot or simply remembers the brand, then comes back through Google, Amazon, or direct — sometimes days later.

The purchase happens. The ad caused it. But the click that attribution depends on never occurred.

The result is campaigns that look excellent in-platform while direct and organic search revenue goes soft. That missing revenue is what we call the screenshot economy, and it is not a rounding error. If younger cohorts are not being served ads at scale, you are not only missing conversions today — you are draining tomorrow's search demand.

Why last-click makes it worse

Last-click attribution gives 100% of the credit to the final interaction before a conversion. That last touch is usually a branded search ad or a retargeting impression — the cheapest, most certain moment in the journey.

It ignores the awareness campaign, the social proof, and the content that created the demand in the first place. So the data tells you retargeting is your golden goose, while systematically undervaluing everything that filled the funnel.

The downstream effects compound:

• Overinvestment in lower-funnel ads — retargeting, branded search
• Underinvestment in awareness — social, creators, PR, content
• Misleading ROI expectations that make scaling decisions harder
• Stalled growth, because you keep feeding the bottom of the funnel without filling the top

You end up optimizing for short-term efficiency while suffocating long-term growth. And in a zero-click environment, last-click does not just under-credit awareness — it cannot see the conversion path at all.

Models that see more of the journey

There are three practical steps up from last-click:

1. Multi-touch attribution (MTA) spreads credit across touchpoints in the journey. Strong visibility, but data-heavy and dependent on tracking quality.
2. Position-based (U-shaped) weights the first and last touch heavily while still crediting the middle. The most straightforward upgrade from last-click, and a good first test.
3. Data-driven attribution uses machine learning to assign credit from observed performance patterns. Usually the most accurate, and the most demanding of your tracking infrastructure.

At higher spend levels — roughly $1M+ per month — we lean on Northbeam for MTA. It gives a clearer read on the purchase journey and distributes credit across the funnel rather than collapsing it into the last touch.

You do not need to replace your entire measurement stack at once. Start with directional tests: run last-click and a position-based model side by side, watch how budget allocation would shift, and bring the comparison to leadership to build buy-in before committing.

Value rules: a lever for rebalancing delivery

Better measurement tells you delivery is skewed. Value rules let you do something about it.

Meta's value rules let you bid up on priority cohorts — say women 25 to 45 — and bid down on less valuable ones, without resetting ad set learning. That last part matters. It is one of the few surgical levers available on a long-lived ad set.

On one large brand, a 20% bid increase toward the target demographic shifted delivery back toward younger audiences almost immediately. It is not a complete fix for demographic drift, but it is a real one, and it does not cost you the learning phase.

The metrics worth building into every account

Platform-default reporting will not surface any of this. A few custom metrics do:

• First-time impression rate — how much of your reach is genuinely new people rather than your existing warm pool. High ROAS paired with a low first-time impression rate is a warning sign: you are recycling an audience you already own and the account will stop scaling. On accounts under $150K/month we look for this to hold above roughly 50% on a 7-day window.
• Quality click percentage — landing page views divided by link clicks. If 100 people click and 50 arrive, something is broken: page load, or a hook that misrepresents the landing page.
• Quality session percentage — add-to-carts divided by clicks. Signals deeper funnel and CRO problems rather than ad problems.
• Blended MER — total revenue divided by total ad spend. The number that does not care about attribution windows.

These are not really ad metrics. They are windows into the whole purchase journey. If DTC sales dip right after you launch on Amazon, check cart drop-offs — customers are often bouncing to buy on Prime, which means your ad dollars are funding someone else's recorded conversion.

Incrementality is the only real answer

Every model above still assigns credit to touchpoints it can observe. Incrementality asks a different and better question: how much additional revenue did this campaign drive beyond what you would have earned anyway?

That reframing matters most for the campaigns attribution treats worst. Impression-based and awareness campaigns look weak in Ads Manager — low ROAS, little last-click credit. Measured for incremental lift, they usually turn out to be the first touchpoint for new customers. They create demand, feed the funnel, and play defense: without them, competitors capture your existing customers with their own ads.

Practically, that means setting a budget, measuring lift through geo holdouts or a third-party tool, and letting the result guide spend — rather than chasing performance retroactively through a ROAS report. It is less tidy than a clean dashboard number. It is much closer to the truth.

Platforms with long consideration windows make the case plainly. Pinterest often drives sales well after the initial impression. Without incrementality testing, giving it proper credit is close to impossible.

Why this gets harder as you scale

This is not a small-brand problem. It gets more complicated with size.

Brands with large organic reach — a TV presence, a viral YouTube channel, an established social following — flood their own pixels with "already aware" signals. That makes it harder for Meta to distinguish and deliver to genuinely net-new customers. Without deliberate setup, your prospecting spend quietly becomes expensive retargeting.

The fixes are structural: new-customer conversion events, value rules, and suppression audiences. Exclusion targeting and tooling like Blotout help structure incrementality tests that separate true new-customer impact from the noise of traffic you already had.

Where to start

Attribution will never be perfect, and there is no single tool that fixes it. But the sequence is clear:

1. Audit your current model. If last-click is still the default, that is the first thing to change.
2. Instrument new-customer reach, not just ROAS, so demographic drift is visible before it costs you a quarter.
3. Use value rules to rebalance delivery toward the cohorts that drive downstream search demand.
4. Run incrementality tests on the channels attribution treats worst, and let lift guide budget.

The brands that keep growing are not the ones with a perfect dashboard. They are the ones whose operators stay curious enough to keep testing the assumptions underneath it.

FAQ

What is the screenshot economy in paid media?

It describes zero-click buying behavior: a user sees an ad, screenshots it or simply remembers the brand, then searches and purchases later through organic, direct, or a marketplace. The ad drove the sale but generated no click, so click-based attribution credits the revenue to organic or leaves it untracked. It is most pronounced among Gen Z audiences.

Why is my in-platform ROAS strong while total revenue is flat?

The most common cause is that platform-reported ROAS is measuring a warm audience you already own, while genuinely new reach is shrinking. Check what share of your impressions go to first-time viewers, and compare against blended MER — total revenue divided by total ad spend. Strong in-platform ROAS with a low first-time impression rate means the account is recycling existing customers and will stall.

What should replace last-click attribution?

Move up in steps rather than all at once. Position-based (U-shaped) attribution is the simplest upgrade and credits both the first and last touch. Multi-touch attribution gives fuller journey visibility at higher spend levels. Data-driven attribution is typically most accurate but needs strong tracking infrastructure. Above all of them, incrementality testing measures lift rather than assigning credit.

How do Meta value rules fix demographic skew?

Value rules let you bid up on priority demographics and bid down on lower-value ones without resetting ad set learning, which is what makes them usable on established campaigns. On one large brand, a 20% bid increase toward the target demographic shifted delivery toward younger audiences almost immediately.

Why do impression-based campaigns look unprofitable but still matter?

They are usually the first touchpoint for a new customer, so last-click gives them almost no credit. Measured for incremental lift, they do two jobs: creating demand that later converts through other channels, and defending existing customers from competitors' ads. Cutting them typically improves reported ROAS while reducing actual revenue.