This might be the most unpopular thing I say to D2C founders: your ROAS number is probably the least useful metric in your entire marketing dashboard.
I know. ROAS is the holy grail. It's the number every agency reports, every founder obsesses over, and every Meta ads "expert" uses to prove their campaigns are working.
But here's the problem: ROAS, as most people track it, tells you almost nothing about whether your marketing is actually making money.
The ROAS Illusion
Let's do some basic math that surprisingly few people bother to do.
Say you're a D2C skincare brand. Your Meta ads show a 3x ROAS. For every ₹1 you spend on ads, you're getting ₹3 in revenue. Sounds great, right?
Now let's look underneath:
→ Average order value: ₹1,200 → COGS (product cost): ₹360 (30%) → Packaging and shipping: ₹180 (15%) → Payment gateway fees: ₹24 (2%) → Returns/RTO: ₹144 (12% of orders returned, average cost per return) → Ad spend per order: ₹400 (the "3x ROAS" part)
Total cost per order: ₹1,108 Revenue: ₹1,200 Profit: ₹92
That's a 7.6% margin. On a good day. Before you account for team salaries, software subscriptions, warehouse costs, and the dozen other overhead items that eat into that number.
Your "3x ROAS" is actually delivering razor-thin margins that could flip negative with a single bad month of returns.
Now here's where it gets worse.
What ROAS Doesn't Account For
1. Returns and RTO (Return to Origin)
In India, especially for COD orders, RTO rates run 15-25% in many D2C categories. That's orders that ship, get rejected, come back, and cost you shipping both ways plus handling. Most ROAS calculations don't subtract this. They count the order as revenue at the point of purchase.
A 3x ROAS with 20% RTO is effectively a 2.4x ROAS. That's a massive difference.
2. Attribution is broken.
Meta's attribution model counts conversions within a 7-day click or 1-day view window by default. That means someone who saw your ad, didn't click, but bought 18 hours later through a Google search gets attributed to Meta.
This isn't Meta being dishonest. It's how their attribution model works. But it means your Meta ROAS is almost certainly overstated. We've seen cases where switching to last-click attribution drops reported ROAS by 30-40%. The sales were real, but they weren't all from Meta.
3. New customer vs. repeat customer ROAS.
If your overall ROAS is 4x but 40% of those purchases are from repeat customers who would have bought anyway (through email, organic, or direct), your actual acquisition ROAS on new customers might be 2x or lower.
Blending new and repeat ROAS is the single most common way D2C brands fool themselves into thinking their paid acquisition is healthier than it is.
4. Post-purchase economics.
ROAS measures the first transaction. It says nothing about lifetime value. A customer acquired at 1.5x ROAS who buys 4 more times over the next year at zero acquisition cost is infinitely more valuable than a 4x ROAS customer who never comes back.
But most D2C brands optimize for the first number, not the second. Because it's easier. Because it's what the dashboard shows.
What to Track Instead
I'm not saying throw ROAS out entirely. It has its place as a directional metric. But it should not be your primary measure of marketing effectiveness. Here's what should be:
1. Contribution Margin Per Order
This is revenue minus all variable costs: COGS, shipping, payment fees, RTO costs, and ad spend. It tells you how much actual money you made (or lost) on each order.
If your contribution margin is negative, scaling spend makes the problem worse, not better. Doesn't matter if ROAS is 3x or 5x.
2. Blended CAC (Customer Acquisition Cost)
Take your total marketing spend (all channels, not just Meta) and divide by new customers acquired. This gives you the real cost of getting a new customer through the door.
Why blended? Because in reality, customers don't come from a single channel. They might see a Meta ad, search your brand on Google, read a review, and then buy. Attributing that customer to one channel is fiction. Blended CAC accepts this reality and gives you a number you can actually work with.
3. New Customer ROAS (Isolated)
Separate your reporting. What's the ROAS on campaigns targeting only new customers? What's the ROAS on retargeting existing customers? These are fundamentally different numbers and should inform different decisions.
If your new customer ROAS is 1.5x but retargeting is 8x, your blended 3x ROAS is masking the fact that new customer acquisition is barely breaking even.
4. LTV:CAC Ratio
The real measure of sustainable growth. If your average customer lifetime value is ₹4,000 and your CAC is ₹1,000, you have a 4:1 ratio. That's healthy. If it's below 3:1, you have a problem.
This is the metric that tells you whether it's worth acquiring a customer even at a seemingly high upfront cost. A ₹1,500 CAC sounds terrible until you realize the average customer is worth ₹8,000 over 12 months.
5. Payback Period
How long does it take to recoup your customer acquisition cost? If your CAC is ₹1,000 and the first purchase generates ₹200 in contribution margin, your payback period is 5 purchases. If the average customer buys once every 2 months, that's a 10-month payback.
Can you afford to fund 10 months of customer acquisition before you break even? If not, you need to either lower CAC, increase margin, or shorten the repurchase cycle.
The Real Dashboard
If I could redesign the marketing dashboard for every D2C brand, it would have exactly 5 numbers on the home screen:
→ Contribution margin per order (by channel)
→ Blended CAC for new customers
→ New customer ROAS (not blended)
→ LTV:CAC ratio (trailing 12 months)
→ Payback period in months
Everything else, impressions, CTR, CPM, even total ROAS, is a supporting metric. Useful for optimization but dangerous as a decision-making metric.
So What Do You Do With This?
Start by building a real P&L at the order level. Not a spreadsheet with revenue and ad spend. A full breakdown that includes COGS, shipping, returns, gateway fees, and ad spend per order.
Once you see the real numbers, one of two things will happen: either you'll realize you're more profitable than you thought (rare but it happens), or you'll realize the leaks in your unit economics that ROAS was hiding.
Either way, you'll be making decisions based on reality instead of a vanity number. And that's how you build something that lasts.