D2C e-commerce
E-Commerce ROAS Improvement Case Study (Meta + Google)
Quick answer
E-commerce ROAS improves when the product feed is clean, campaigns are consolidated rather than fragmented, creative is tested against a structured matrix, and the account is judged against break-even ROAS rather than a vanity target. Feed and creative usually move ROAS more than bidding changes do.
The brief
A direct-to-consumer store with 40–200 SKUs, running fragmented Meta and Google campaigns built up over time, with duplicate audiences and no clear view of contribution margin.
The challenges
- Fifteen small ad sets splitting conversion data so nothing exits the learning phase
- Product feed missing GTINs, sizes and product types, so Shopping coverage was poor
- Pixel and Conversions API double-counting purchases
- ROAS target set by feel, with no break-even calculation behind it
Account structure
- Meta: one Advantage+ Shopping campaign for prospecting, one retargeting campaign, broad audiences
- Google: one Performance Max asset group per product category, plus a branded Search campaign kept separate
- Retargeting split only by intent depth (viewed, added to cart, initiated checkout)
- Budgets consolidated so each campaign clears at least 50 conversions a week where possible
Tracking setup
- Meta Pixel plus Conversions API with a shared event ID for deduplication
- GA4 e-commerce events (view_item, add_to_cart, begin_checkout, purchase) with value and currency
- Google Merchant Center feed audited for GTIN, brand, product_type, availability and image quality
- Break-even ROAS calculated from gross margin before any target is set
Creative angles
- Problem-solution UGC video as the primary prospecting format
- Founder-story and behind-the-manufacture video for cold trust
- Static offer cards for retargeting and price-sensitive audiences
- Review and social-proof carousels for consideration audiences
Optimisation levers
1Break-even ROAS first
Divide 1 by the gross margin. At a 40% margin, break-even ROAS is 2.5. Every decision is judged against that number instead of an arbitrary 4x.
2Consolidation
Merge fragmented ad sets so the algorithm gets enough signal. Fewer, bigger campaigns almost always beat many small ones at modest budgets.
3Creative testing matrix
Test one variable at a time across hook, format and offer. Retire creatives on frequency and CTR decline, not on a fixed schedule.
4Feed quality
Complete attributes, accurate titles with the search term at the front, and clean images widen Shopping eligibility — often the fastest Google-side ROAS gain.
5Post-purchase economics
Add an order bump or a bundle so average order value rises. A higher AOV raises ROAS without touching the ad account at all.
Worked example: how the maths moves
Illustrative model, not a specific client's reported result. It shows how the levers interact.
| Gross margin | 40% → break-even ROAS 2.5 |
|---|---|
| Monthly ad spend | ₹2,00,000 |
| Average order value | ₹1,450 → ₹1,780 (bundle added) |
| Conversion rate | 1.4% → 1.9% (feed + landing page) |
| Blended ROAS | 2.1 → 3.2 |
| Contribution after ad spend | Negative → positive |
The account moved from below break-even to profitable with no budget increase — AOV and conversion rate did the work, bidding did not.
What this teaches
- A ROAS target without a margin calculation behind it is a guess.
- Consolidation beats granularity at Indian D2C budget levels.
- Feed hygiene is unglamorous and frequently the largest single Google-side win.
- Creative is the main variable on Meta; the algorithm handles most of the rest.
FAQs
What is a good ROAS for e-commerce?
There is no universal number. Calculate break-even ROAS as 1 divided by your gross margin, then target above it. A 60% margin store is profitable at 1.7; a 25% margin store needs 4.0 just to break even.
Should I use Advantage+ Shopping or manual campaigns?
Advantage+ Shopping usually wins once the account has enough purchase volume and clean tracking. Manual campaigns remain useful for tightly controlled retargeting and for testing new audiences deliberately.
How many creatives should I test per month?
Aim for six to ten new concepts a month at moderate budgets, refreshing winners with new hooks rather than replacing the whole set at once.
Run this campaign on a real ad account
PMAcademy's live programme walks you through builds exactly like this one, with mentor feedback.
Book the free live masterclass