Why is my Meta ROAS high but the business is losing money?

Ads Manager counts order value at checkout, including GST and orders that later come back as RTO. After you remove GST, RTO losses, COD fees, shipping and product cost, the real return can be far lower. Work out breakeven ROAS on the same basis as Ads Manager and compare against that.

How long should I wait before turning off an ad?

A practical rule is to wait until the ad has spent at least twice your target CPA and run for three days. If it has no purchases at that point, turn it off. Earlier than that, results are mostly noise.

COD reaches buyers who will not pay online, especially outside metros, but brings RTO and cash-handling fees. Test a prepaid incentive and COD confirmation, and compare contribution per order rather than order volume.

Do regional language ads work better than English?

Often, in the regions where that language is spoken, but test it. Write native copy rather than translating English ads, and judge on contribution, not just click-through.

Meta ads for D2C brands in India: angles, policy and a testing plan

Angles worth testing

Problem the product solves (stated about the product)

Hypothesis: Naming the everyday situation the product is made for, without describing the viewer, will lift relevance and click-through compared with generic brand messaging.

A lunch box that keeps dal hot until 1 pm.
Made for small kitchens with one shelf of space.
A charger cable built for the way you actually use it on the commute.

Proof you can check

Hypothesis: Specific, verifiable proof (ratings with counts, test reports, warranties) will convert better than vague claims.

Our rating and review count are on every product page. Read the low ratings too.
One-year warranty, claimed on WhatsApp with a photo.
Here is the lab report for this batch.

Founder story

Hypothesis: A founder explaining why the product exists, filmed simply, will build trust with cold audiences more than polished brand films.

I started this because I could not find a steel bottle that did not leak.
This is the workshop where every order is made and packed.
Why we chose to sell only on our own website.

Demo and comparison

Hypothesis: Showing the product working, side by side with the common alternative, will beat lifestyle shots for products with a functional benefit.

Same spill, two mats. Watch what happens.
Our bag next to a regular backpack, packed for a weekend.
Ten seconds to set up. Timer on screen.

Prepaid incentive

Hypothesis: Offering a small prepaid discount or free shipping on prepaid orders in the ad will shift order mix toward prepaid and raise contribution per order.

Pay online and save ₹50. COD is available too.
Free shipping on prepaid orders.
UPI or card at checkout gets you a free sample in the box.

Gifting

Hypothesis: Gift-ready packaging and messages will lift conversion during festivals and occasions compared with standard product ads.

Gift-wrapped, with a handwritten note, delivered across India.
A Raksha Bandhan gift that is not another box of sweets.
Send it directly to their address. We hide the price.

Formats

Short vertical video (9:16): Reels and Stories carry most inventory. A product demo or founder clip in the first two seconds does the work.
Static image with one claim: Quick to produce and iterate. Good for testing angles cheaply before committing to video.
Carousel: Use it for multiple products, multiple uses or step-by-step explanations.
Catalogue retargeting: Brings back visitors with the exact products they viewed, once you have enough site traffic.

Policy watch-outs

Personal attributes
Meta does not allow ads that assert or imply personal attributes such as health, finances, religion or appearance. 'Are you struggling with…' style hooks are a common rejection across categories.
Meta Advertising Standards: personal attributes
Substantiation and comparisons
The ASCI Code requires claims to be truthful and backed by evidence, and comparisons to be fair. Superlatives like 'India's best' need objective proof you can produce on request.
ASCI Code for Self-Regulation of Advertising Content
Dark patterns
The 2023 dark patterns guidelines prohibit practices such as false urgency, drip pricing, basket sneaking and confirm-shaming. They apply to your ads and checkout flow.
Department of Consumer Affairs: Guidelines for Prevention and Regulation of Dark Patterns, 2023
Creator and influencer disclosure
Paid or gifted creator content needs a clear 'Ad', 'Sponsored' or 'Collaboration' label under ASCI's influencer guidelines, including when run as partnership ads.
ASCI guidelines for influencer advertising

Seasonal calendar

When	Moment	Ad idea
January	Pongal, Makar Sankranti and Lohri	Regional campaigns in Tamil, Telugu, Gujarati and Punjabi, tied to the festival's foods, colours and gifting.
March to April	Holi, Eid and regional new years	Gifting and home refresh offers, with regional language creatives for Gudi Padwa, Ugadi and Baisakhi.
August	Raksha Bandhan and Independence Day	Sibling gifting with clear delivery cut-offs. Avoid misusing the national flag in creatives.
August to September	Onam and Ganesh Chaturthi	Regional campaigns for Kerala and Maharashtra with native-language copy.
October to November	Navratri, Durga Puja, Dhanteras and Diwali	The biggest window of the year. Build creative early, expect higher CPMs, and plan stock and delivery dates.
November to February	Wedding season	Gifting and occasion-led offers for guests and families.

A testing plan

Work out your breakeven CPA and ROAS first, after COD fees, RTO, shipping and GST. Set target CPA from that, not from what Ads Manager shows as a good ROAS.
Write a hypothesis for every ad before you make it. If you cannot say what you expect to learn, do not launch it.
Make new concepts genuinely different: a different angle, persona or format. Five colour variants of one static count as one concept.
Give each ad at least 2× target CPA in spend and 3 days before a verdict. Turn off any ad that spends 2× target CPA with zero purchases.
On winners, iterate one variable at a time (hook, first frame, offer, format) so you know what moved the result.
Judge winners on contribution after COD, RTO and GST, joined to delivery data from your logistics partner, not on platform ROAS.

Worked example: breakeven ROAS

Illustrative assumptions, not benchmarks. Replace with your own numbers. 18% GST is used as a general rate. Your rate depends on your product's HSN code, and many items moved between slabs in the September 2025 GST changes, so check yours.

Breakeven ROAS in Ads Manager

2.68×

Below this, every order the platform reports costs you money. Breakeven cost per purchase: ₹373.

Target ROAS for 10% margin: 3.35×
Target cost per purchase: ₹298
Shortcut “1 ÷ margin”: 1.83×
Orders actually delivered: 89.0%

Where an order’s money goes

Product cost₹267
Shipping + packaging₹85
RTO losses₹6
COD + gateway fees₹22
Left for ads and profit₹373
Per order placed, out of ₹753 realised revenue after GST and RTO.

Questions

Why is my Meta ROAS high but the business is losing money?: Ads Manager counts order value at checkout, including GST and orders that later come back as RTO. After you remove GST, RTO losses, COD fees, shipping and product cost, the real return can be far lower. Work out breakeven ROAS on the same basis as Ads Manager and compare against that.
How long should I wait before turning off an ad?: A practical rule is to wait until the ad has spent at least twice your target CPA and run for three days. If it has no purchases at that point, turn it off. Earlier than that, results are mostly noise.
Should I offer COD?: COD reaches buyers who will not pay online, especially outside metros, but brings RTO and cash-handling fees. Test a prepaid incentive and COD confirmation, and compare contribution per order rather than order volume.
Do regional language ads work better than English?: Often, in the regions where that language is spoken, but test it. Write native copy rather than translating English ads, and judge on contribution, not just click-through.