Meta Ads for D2C in India: Structure, Creative, Scaling

Keep the account structure boring

Modern Meta rewards simplicity. Fragmenting spend across dozens of tiny ad sets starves each of the data the algorithm needs to optimise. A consolidated structure, broad targeting, a healthy budget per campaign, and creative doing the differentiation, usually out-learns a maze of narrow audiences. Let the structure be dull so the creative can be interesting.

Creative is 80% of the result

On Meta, targeting has largely been handed to the algorithm, so the creative is the targeting, it decides who leans in. That means a volume-and-variety approach: many angles (problem, proof, offer, founder, UGC), many formats, tested constantly, with the winners scaled and the losers cut. A brand shipping ten thoughtful new creatives a week will almost always beat one polishing a single ‘perfect’ ad. Study what wins and make more of that.

Scaling without wrecking margin

1. Scale winners, don’t chase them — raise budgets gradually so the algorithm doesn’t reset learning.
2. Watch the margin, not the ROAS — as you scale, efficiency usually softens; know your break-even.
3. Feed the funnel — retention and email/WhatsApp lifecycle let you afford a higher acquisition cost.
4. Keep the creative pipeline full — scaling burns creative faster, so production has to keep pace.

What most brands get wrong

The most common mistake is treating Meta as a targeting problem and endlessly tweaking audiences while shipping one creative a month, then wondering why growth stalled. The second is scaling on platform-reported ROAS, which over-claims, so a brand ‘scaling profitably’ on-screen is quietly losing money per order. The third is cutting creative production to save cost, which is like a restaurant firing the chef to lower expenses, you’ve removed the thing that actually drives the result.

The contrarian take: your media buyer matters less than your creative team

D2C brands obsess over finding a genius media buyer. On today’s Meta, the media buying is largely automated. The durable edge is a creative system that produces winning ads faster than competitors. A competent buyer with a great creative pipeline beats a brilliant buyer with a thin one, every time. If you’re going to over-invest anywhere, over-invest in creative volume and quality, not in another layer of campaign structure.

What tends to improve, a realistic picture

• Business type: a growing D2C brand plateaued on Meta.
• Common problem: flat performance, over-complex account, thin creative output.
• Typical approach: simplify structure, ramp a weekly creative pipeline across angles and formats, scale winners gradually, and measure to contribution margin.
• What tends to improve: more consistent scaling and clearer profitability once creative volume rises and measurement is honest. Outcomes vary by category, margin and creative quality.

Audit your Meta account in 15 minutes

1. How many genuinely new creatives did you ship in the last 30 days?
2. Is your account a handful of campaigns, or a maze of tiny ad sets?
3. Are you scaling on platform ROAS or on contribution margin?
4. Do you know your break-even ROAS after all costs?
5. Is your creative pipeline funded to keep up as you scale?

Most plateaus trace back to question one. This is the acquisition engine inside Shopify & D2C marketing; model the maths on the ROAS calculator and read what a good ROAS really means.

The numbers to check before you touch the ad account

Meta will report a return figure inside the platform and that figure will almost always look better than your bank account does. This is not deception. The platform counts conversions it believes it influenced, using its own attribution window, which is a reasonable thing for a platform to do and a terrible thing to run a business on.

So the first number to fix is a blended one. Take total revenue for the month, divide it by total advertising spend across every channel, and you have a single ratio that no attribution model can flatter. Track it weekly. Watch what happens to it when you increase spend, because the useful question is never what the platform says a campaign returned, it is whether the whole business made more money in the weeks you spent more, and that comparison is the one your finance team will accept without an argument about attribution windows.

The second number is contribution margin per order after everything variable comes out. Product cost, packaging, shipping both ways, payment gateway charges, the discount you actually gave rather than the one on the price list, and the cost of orders that come back. What is left buys a customer. If that number is small, no amount of creative testing rescues the account, and the fix is a pricing or product decision rather than a media one.

Then hold three ratios against your own history rather than against anyone’s published average. First, what your customer acquisition cost is as a share of that contribution margin, which tells you whether you are buying profitably today or buying on the promise of a repeat order. Second, what share of revenue comes from returning customers, since a brand where that share is rising can afford a higher acquisition cost than one where every order is a first order. Third, how your average order value moves when you scale, because it commonly falls as targeting widens and a campaign that looks efficient at the click level can quietly be buying smaller baskets. None of those need an industry benchmark. They need last quarter’s data and an honest reading of it.

A worked example on a single product

Abstract advice about margin does nothing. You have to see it laid out, so here is one product, worked end to end, with invented numbers you should replace with your own before drawing any conclusion.

Selling price is ₹1,500. Product cost is ₹450, packaging is ₹50, shipping out is ₹80, and the payment gateway takes roughly ₹30. That leaves ₹890 before anything goes wrong. Now apply the things that go wrong: assume one order in ten is returned or refused at the door, and each of those costs you the outbound shipping plus a return leg, say ₹160 wasted per failed order. Spread across ten orders that is ₹16 an order, so contribution lands near ₹874. Round it to ₹850 and be conservative, because there is always something you forgot.

That ₹850 buys a customer. It is your entire budget, and only if you are content to break even on the first order.

Say the account is acquiring customers at ₹700. On paper the brand makes ₹150 an order, which sounds thin until you look at repeat behaviour, because if a reasonable share of those buyers order again within ninety days and the second order carries no acquisition cost at all, the real return sits well above what the first transaction suggests. Rebuy changes everything. That is the whole argument for spending aggressively on a product people buy again. It is also the trap, since the same arithmetic applied to a product nobody rebuys means ₹700 to earn ₹150 and a business that grows revenue while losing money. Work out your own repeat rate before deciding which of those two businesses you are running, and if you genuinely do not know it, that is the most valuable hour of work available to you this week.

One more pass. Push acquisition cost to ₹900 and the first order loses money outright, which is survivable for a brand with strong repeat purchase and fatal for one without. The point of the exercise is not the numbers. It is knowing exactly which figure has to move before you are allowed to scale.

COD, returns and the part that is specific to selling here

Cash on delivery changes the arithmetic of a D2C brand in India more than any targeting decision will. Orders placed without payment are easier to place and easier to abandon, and a parcel refused at the door costs you both shipping legs plus the handling, which lands entirely on the orders that did convert. Your good orders subsidise the failures.

Measure it properly rather than guessing. Track the return-to-origin rate separately for cash orders and prepaid orders, and separately again by region, because the pattern is rarely uniform and the useful decisions are local ones. Some brands find that a modest prepaid incentive, an order confirmation over WhatsApp, or a simple address verification step moves the number enough to fund itself several times over. Others find that switching cash off entirely in the worst pincodes costs less revenue than the failed deliveries were costing.

Test it. Do not decide it in a meeting.

The second local factor is what happens between the ad and the payment page. UPI is how a very large share of Indian buyers pay, so a checkout that makes it awkward is losing orders that your media budget already paid to acquire, and the same applies to a page that assumes a fast connection or a card. Mobile is the default, not a variant. The third factor is delivery expectation, since a buyer in a metro who is used to next-day arrival reads a seven-day estimate very differently from someone in a smaller town, and that expectation gap shows up as cancellations rather than as anything visible in the ad account.

None of this sits inside Meta. All of it decides whether Meta works for you, which is why a brand struggling to scale profitably should audit the path from click to delivered parcel before it audits the campaign structure. The ad account is usually the least broken thing in the chain.

A 30-day plan for a brand starting from scratch

Most brands launch on Meta by putting a small budget behind whatever creative exists and waiting. A month later there is not enough data to conclude anything and the budget is gone. Here is a more disciplined plan. Same money, better sequence.

Week one is measurement, and nothing else. Get the pixel and the server-side connection working properly, confirm that purchase events fire once per order with the correct value attached, and build the blended revenue-over-spend sheet you will use to judge everything afterwards. Also write down, before spending anything, what contribution margin per order actually is. If you skip this week you will spend the next three arguing with numbers you have no reason to trust.

Week two is creative production. Aim for a handful of genuinely different concepts rather than fifteen variations of one idea, since colour swaps and text tweaks answer a much smaller question than testing whether the product is better sold on the problem it solves, a demonstration, a customer talking, or a straight offer. Shoot for the placement people actually use. Vertical, sound-optional, hook in the opening seconds.

Weeks three and four are the test. One broad campaign, one optimisation event that reflects a real purchase, and enough budget behind it to produce a meaningful number of orders rather than a scattering, because the most common mistake at this stage is splitting a small budget across six ad sets and ending the month with six sets of unreadable data. Leave it alone. Resist daily edits, since every change restarts the learning and the account never settles long enough to tell you anything.

At day thirty you decide with evidence. Which concept produced the cheapest genuine purchases, what your real acquisition cost is against the margin you calculated in week one, and whether the checkout and delivery experience held up under the first proper volume. Only then does scaling become a sensible conversation. Brands that follow this sequence spend the same money and end the month knowing something, which is the entire difference between a test and a donation.