🔑 When display ROAS outperforms search ROAS, the first question isn't 'should I shift budget?' — it's 'what attribution model is each channel using?' View-through attribution inflates display ROAS by 2–5x in most accounts. You're not looking at a performance comparison. You're looking at two channels measured by different rulers.
Display claiming a 4x return while search sits at 2.8x looks like a clear budget reallocation signal. It isn't. Search and display don't just perform differently — they measure performance differently by design. Treating them as equivalent will consistently steer budget in the wrong direction.
How Search Attribution Works
Paid search (CPC) operates on click-based attribution. You pay when someone clicks. You get credit when that click leads to a conversion within your attribution window (30 days for Google Ads, 7 days for Amazon Sponsored Products). An active user action is required before the attribution clock starts.
If they saw your ad and didn't click, it generates no cost and no conversion credit. Search attribution is inherently conservative — it refuses to claim influence it can't prove.
Search attribution is the most defensible metric in performance marketing. Every dollar of reported ROAS corresponds to a confirmed three-step chain: query → click → conversion. Nothing is inferred.
How Display Attribution Works
Display (CPM) is purchased by impression. At a 0.1% CTR, 99.9% of your ad spend would have zero attributed return under a click-only model. So display platforms use view-through attribution: a conversion window during which anyone who saw — but never clicked — your display ad can still be counted as a display conversion if they later convert through any channel.
| Platform | Default View Window |
| Amazon DSP | 14 days |
| Google Display | 1 day |
| Meta | 1 day |
| Google Search | Click-only |
The Double-Counting Problem
A user searches 'standing desk,' clicks your search ad, and buys. Your search campaign gets credit — correctly. But in the prior 10 days, that same user was also served a DSP impression while reading an article. They saw the ad, didn't click, moved on. Under a 14-day view-through window, your DSP campaign also claims credit for that same conversion.
Both campaigns are now counting the same revenue. Your search ROAS is not wrong. Your display ROAS is not technically lying. But your total reported ROAS is counting the same sale twice — once by the channel that closed it, and once by the channel that observed it from a distance.
The Test Every Media Buyer Should Run
Pull the attribution comparison report for your active display campaign. Look at ROAS under three scenarios: click-only, 1-day view-through, and 14-day view-through. In a typical account, ROAS will climb 2–5x as you extend the view window — even though the campaign didn't change a single creative or targeting parameter. That multiplier is attribution window selection, not performance.
What Display Is Actually Responsible For
Stop evaluating display on ROAS and start evaluating it on the signals it actually controls. The right questions to ask:
The Incrementality Test
The only honest number display can produce is a geo hold-out test: run display in five markets, keep five comparable markets dark, compare total category revenue lift after 30 days. In almost every incrementality test across retail media and programmatic, view-through ROAS overstates true incremental impact by a factor of 2–3x.
The display campaign isn't driving that revenue. It's claiming credit for the revenue your search campaign earned — and the attribution window is designed to let it do exactly that.
What This Means for Budget Allocation
Search captures confirmed intent and closes. Display builds ambient awareness and primes. They are not competing for the same job. Stop making direct ROAS comparisons between them — and start asking what each channel is actually responsible for in the conversion chain.
📚 Sources: IAB — Attribution Playbook: A Primer for Digital Attribution (iab.com); Google — About attribution models in Google Ads (support.google.com/google-ads); Nielsen — 2024 Annual Marketing Report: Attribution and Measurement (nielsen.com); eMarketer — Digital Advertising Attribution Report 2024 (emarketer.com).
