I Killed a Campaign Showing 5x ROAS. Here’s Why.

The Campaign That Looked Perfect

On paper, everything about this campaign checked out.

Strong ROAS (5.1x). Low CPA (₹310). High click-through rate. Healthy impression volume. The creative was a 30% off sale, and it was converting beautifully.

But when we pulled the full picture:

Nearly 1 in 3 customers who bought through this campaign returned the product. And when we ran the real numbers:

The campaign was generating ₹0.28 in gross profit per rupee of ad spend, before team costs, before rent, before software. Effectively burning cash at scale.

Why ROAS Couldn’t See This

Meta’s pixel sees the purchase event.

It does not see the return.

When 620 customers returned their products 10–14 days after purchase, those conversions stayed in Ads Manager unchanged. Revenue stayed attributed. ROAS stayed at 5.1x.

The return hit Shopify. It hit the logistics partner. It hit the founder’s bank account.

Meta never knew.

And so the algorithm kept optimising toward this campaign. Kept finding more customers who would click a 30% off offer, buy, and return. The cycle kept running. The losses kept accumulating.

The Customer This Campaign Was Acquiring

Here’s the thing about 30% off campaigns with 31% return rates.

Customers who were attracted by a 30% discount have a specific profile:

• Price-driven, not brand-driven
• Higher likelihood of returning if the product doesn’t immediately wow them
• Lower LTV (they wait for the next sale rather than repurchasing at full price)
• Sometimes buying from multiple brands simultaneously, intending to keep only the best

The 5.1x ROAS campaign was efficiently finding and acquiring exactly these people. At enormous cost.

What Happened After We Killed It

The founder was nervous. Understandably. The campaign was their highest ROAS. Killing it felt like choosing to perform worse.

Two things happened in the four weeks after:

Contribution margin improved by 11 percentage points. Without a single change to the remaining campaigns.

The Lesson

ROAS rewards campaigns that generate attributed revenue efficiently.

It does not reward campaigns that generate profit.

A discount campaign that converts brilliantly, 30% off, high CTR, high CVR, will always show strong ROAS. The discount is priced into the conversion, not into the return.

The return, which is where the damage lives, happens after the attribution window. After the platform has already celebrated the conversion. After the ROAS has already been calculated and reported.

You need a metric that catches what ROAS misses.

The Metric That Would Have Caught This Earlier

POAS. Profit on Ad Spend.

If the founder had been tracking POAS per campaign, not just ROAS, this campaign would have been flagged in week two. Not month four.

The fix isn’t stopping discount campaigns. Discounts have a legitimate role in D2C marketing.

The fix is measuring discount campaigns on POAS, knowing your return rate for each campaign and applying it to the calculation. So when a “30% off” campaign has a 31% return rate that pushes POAS to 0.28x, you catch it before it costs you ₹20L.

What to Do Right Now

Pull your current campaigns. For each one, especially your best-performing ones by ROAS, ask:

1. Pull every active campaign, starting with your top performers by ROAS.
2. Find each campaign’s own return rate.
3. Run the POAS calculation with return-adjusted revenue.

The campaign you’re most proud of might be the most important one to look at first.