ROAS is the headline number. It's also the easiest to read wrong.
What we see when we audit a store's paid media: blended numbers masking channel-level losses. New customer acquisition that's negative while repeat buyers keep the average up.
Three patterns we find in every store account that's scaling spend but not margin.
-
01
Reporting on blended ROAS while new customer ROAS is negative
Repeat customers are cheaper to convert and show higher ROAS. When blended into the same campaign, they mask the true cost of acquiring new buyers. An account can show 8x ROAS while new customer acquisition is at 2x — which is below the margin floor for most DTC brands.
-
02
Scaling spend into peak season without a margin floor
Peak season doesn't change the unit economics — it just amplifies them. If new customer ROAS is already thin, scaling budgets in November produces more revenue and worse margins. The accounts that survive peak well build a margin target before they touch budget.
-
03
Treating brand and non-brand campaigns as one cost center
Brand search converts at high ROAS because the demand already exists. Non-brand has to create it. Lumping them together makes the portfolio look better than it is and makes it impossible to evaluate what paid is actually doing for new demand.
What we build and what we measure.
- We split new vs. returning customer campaigns from day one. No blended attribution.
- We set a margin floor before recommending any budget level. ROAS targets come from contribution margin, not from benchmarks.
- We segment by channel, campaign, and audience type. Each gets evaluated independently.
- We audit creative for SKU-level performance. The top 20% of products usually drive 80% of the profitable ROAS.
- We build peak season plans around margin, not volume. Scale where the math works, pull back where it doesn't.
- We track dormant account patterns — buyers who lapse and return. Reactivation often outperforms cold acquisition.
- We report on revenue per visitor, not just ROAS. ROAS doesn't tell you if the landing page is the bottleneck.
| Metric | What it tells you |
|---|---|
| New customer ROAS | Whether paid acquisition is actually profitable |
| Blended ROAS | Overall portfolio health (useful, not sufficient) |
| CAC by channel | Which channel is cheapest for new buyers |
| AOV by cohort | Whether acquisition quality is improving or degrading |
| Contribution margin | The floor your ROAS target must clear |
| Revenue per visitor | Whether the funnel after the ad is working |
Two stores. Different margins. The same discipline.
A fashion DTC brand with solid product but scattered campaign structure. We rebuilt the account around margin-aware targets, separated new and returning buyers, and delivered 11.3x blended ROAS with a 12.6x peak campaign.
Read the case →A B2B uniform supplier with e-commerce infrastructure and a long repurchase cycle. We structured paid to separate channel types, isolated new account acquisition, and built a reporting layer the team could actually act on.
Read the case →From the people who ran it with us.
"Before Loocro, I was looking at overall ROAS and thinking everything was fine. When they broke it down by channel, I realized new customer acquisition was negative. That's a completely different problem than the one I thought I had."
"12.6x in the first peak. That's not a number I believed was possible going in. It took breaking the account apart and rebuilding it from margin up. The results followed the structure."
The reading behind the method.
Why blended ROAS hides the bleed.
The blended average is a number for the monthly slide. Segmented ROAS is the number that runs the business. Two real cases inside.
Read →How to build a segmented ROAS report in 4 steps.
The setup that separates brand from non-brand, new from returning, and product categories from each other. What you see changes everything.
Read →Margin-aware ROAS: the target that keeps you profitable at scale.
ROAS without a margin floor is just a ratio with no meaning. How to set the number that actually reflects your business model.
Read →The dormant account strategy for peak season.
Reactivating past buyers before peak costs a fraction of cold acquisition and converts at multiples. The playbook we use every November.
Read →The questions we get before every e-commerce engagement.
Our ROAS looks healthy but profit is flat. Where's the problem?
Almost always the same two places: new customer ROAS is worse than the blended average because repeat buyers inflate the number, or you're scaling into low-margin SKUs. Splitting ROAS by new vs. returning customers is usually the first thing we do.
How do you approach peak seasons like Black Friday?
We don't scale budgets in peak without a margin floor. Scaling into a period with thin margins because volume is high produces revenue that looks great and margin that doesn't. We build the peak plan around contribution margin, not ROAS.
We sell across multiple channels. How do you handle attribution?
We separate reporting by channel, campaign, and audience type before making any scaling decision. Last-click attribution understates the channels that create demand and overstates the channels that capture it. We look at assisted conversions and incremental lift.
We have a low AOV. Is paid media even viable?
Depends on your repeat purchase rate and LTV. A $30 AOV can work if the customer buys four times a year. We model CAC against 12-month LTV before recommending any spend level.
We've been running the same campaigns for months. Do they need a full rebuild?
Not always. Aged campaigns can still perform well structurally. What breaks down over time is audience saturation and creative fatigue. We audit the structure first before recommending any rebuild.
Book a 30-minute diagnosis.
We'll dig into your blended ROAS, your CAC by channel, your AOV by cohort. We'll show you where the math actually breaks — and what to fix first.
Book the Diagnosis →