I'll admit a bad habit of mine: when I shifted budget between channels, I stopped at the dashboard's ROAS and CPA and almost never reconciled it to actual revenue on my own side. The problem is that pouring more into the "best ROAS" channel can quietly backfire — because a high ROAS often means the channel is re-engaging people who would have bought anyway, not winning anyone new.
TL;DR
- High-ROAS channels skew to returning buyers — platforms optimize toward the cheapest converters, who are usually existing customers
- Split new vs returning with two numbers — total orders and first-time orders; the gap is re-engagement volume
- Contribution margin flips the ranking — same ROAS, different product margin, very different profit
- Judge new-customer channels on a separate yardstick — harvesting demand ≠ creating it
Why a high ROAS can be a trap
ROAS is a channel's attributed revenue ÷ its spend. It looks clean, but it can't tell you whether that revenue is new or returning. Delivery algorithms chase whoever converts cheapest, and the cheapest converter is usually someone who already knows you and would have come back anyway. So retargeting and brand-search tend to post high ROAS — they harvest demand that already exists.
Cut the "low ROAS" channel and you may be closing the front door for the new customers who'd become tomorrow's repeat buyers. One brand killed its podcast ads for low ROAS — and watched search-driven revenue fall the next quarter, because those ads were creating the demand people later searched for.
Two numbers that do the splitting
You don't need a fancy tool. Per channel, pull total orders and first-time orders. The difference is the volume of returning-customer re-engagement.
A channel with 320 total but only 40 first-time orders isn't acquiring — it's reactivating. A plainer channel with a small gap is doing the real new-customer work. Notice how this can flip the ROAS verdict entirely.
Read profit, not revenue
ROAS measures revenue, not what you keep. Contribution profit = revenue − (COGS + shipping + fees + ad spend). Two channels, ¥500K spend each:
- Meta: ROAS 3x, 40% margin → ¥1.5M revenue → ¥100K profit
- Google: ROAS 2x, 70% margin → ¥1.0M revenue → ¥200K profit
Meta wins on ROAS and loses on profit. Revenue is vanity, margin is sanity — translate ROAS into profit before you rank channels, and the order changes.
Bottom line
A high ROAS can be re-engagement in disguise. Split each channel's total vs first-time orders to see who actually brings new buyers, and convert ROAS into contribution profit before deciding where to scale. Take three numbers — total orders, first-time orders, rough contribution profit — for your main channels, and the channel truly worth protecting stops hiding.
When you move budget, do you reconcile platform ROAS back to your own revenue and split new vs returning — or mostly trust the dashboards? Curious how others handle it.
Top comments (0)