Beauty is one of the most crowded categories in Indian D2C advertising. Products are visual, repeat purchase is possible, and the barrier to launching a brand is low, so feeds are full of serums, sunscreens and lip tints competing for the same thumbs. That shapes the playbook. Targeting matters less than it used to, because Meta's delivery system does most of that work. What decides results is how much good creative you can produce, whether creator content is compliant and believable, whether your unit economics survive cash on delivery, and whether the conversion signal you send the platform reflects money you actually collect. I have written a broader piece on why Indian D2C needs a different playbook. This one is narrower and specific to beauty and personal care.
Creative fatigue is the main constraint
In beauty, a single winning ad rarely carries an account for long. The audience for a category like vitamin C serum is large but finite, the same people see many similar ads, and a hook that felt new in week one feels familiar by week four. The symptoms are familiar: frequency climbs, click-through rate falls, cost per purchase rises while nothing else in the account has changed. The fix is not a better single ad but a production system. I plan creative in batches around a few variables: the hook in the first seconds, the claim or benefit, the format (static, short video, carousel, before and after where permitted), and the person on screen. Each batch tests one or two variables so you learn something, rather than ten unrelated ads that tell you nothing. I also keep a simple log of each ad's launch date and the point where cost per purchase started to rise, which over a few months gives you a brand-specific sense of how long your creative lasts. If you want a starting estimate before you have that history, the creative fatigue forecaster on this site models it from frequency and audience size.
Influencers and UGC: what the rules require
Creator content is the default format in Indian beauty advertising, and two sets of rules apply to it. ASCI's guidelines for influencer advertising on digital media, in force since 2021, require a disclosure label that is upfront and prominent, and list accepted labels including #ad, #collab, #promo, #sponsored and #partnership. Reporting on the guidelines also describes a ban on filters that exaggerate a product's effect on skin, hair or teeth in promotional content, which matters a great deal for beauty. Separately, the Department of Consumer Affairs issued its "Endorsement Know-hows!" guide in January 2023 under the Consumer Protection Act, 2019. It requires endorsers to disclose any material connection with the advertiser in a way that is prominent and hard to miss, in both audio and video for video content, and advises them not to endorse products they have not used. In practice that means three things for a brand: disclosure written into every creator contract, a no-filter rule for any skin or hair result, and claims checked against what the product can substantiate before a creator says them on camera. Whitelisted creator ads run from the creator's handle carry the same obligations.
Claims: keep cosmetics claims cosmetic
The fastest way to get beauty ads rejected, or worse, is to make claims the product cannot support. Words like "cures", "treats" or "heals", and promises of permanent results, push a cosmetic toward medical territory. Platform ad policies restrict health and before-and-after claims, and Indian rules on drugs and cosmetics treat therapeutic claims differently from cosmetic ones. I am not a lawyer, and brands should have claims reviewed by someone who is, but the working rule I use is simple: describe what the product does to how skin looks and feels, back any number with a test you can produce on request, and keep a claims sheet that every creator brief draws from. That sheet becomes the single source of truth for performance creative, influencer scripts and product page copy, which also keeps the promise consistent from ad to checkout. Consistency helps conversion as much as compliance: a buyer who clicks on one claim and lands on a page making a different one tends to leave.
COD and RTO change what a good ROAS is
Many Indian beauty brands still take a large share of orders as cash on delivery, especially outside the metros and for first-time buyers. Every COD order that is refused or undeliverable becomes return to origin: you pay shipping out, often shipping back, and sometimes lose the product if it cannot be resold. That makes platform ROAS, which counts orders at the moment they are placed, an optimistic number. I set targets on RTO-adjusted ROAS instead, built from delivered revenue and including the shipping cost of returned parcels, which I explain in detail in a separate piece. Operationally, the usual levers are order confirmation by WhatsApp or call before dispatch, partial prepaid or small COD fees where they do not hurt conversion too much, prepaid incentives, and pin code level rules for areas with persistently high return rates. Each lever trades some conversion rate for better delivered economics, so test them one at a time and judge them on delivered revenue, not orders placed.
Meta CAPI: send the order you will actually get paid for
The default Shopify and Meta setup tells Meta that every placed order is a purchase. For a COD-heavy beauty brand, that teaches the algorithm to find more people who place orders easily, including the ones who refuse parcels. The better approach is to send Purchase through the Conversions API when an order is confirmed or delivered, and to send prepaid purchases immediately with deduplication between browser and server. Meta's documentation allows event_time to be at most seven days before the event is sent, so a delivered-only trigger can miss slow deliveries. I cover the full event plan, including fbp and fbc storage and deduplication, in my article on Meta CAPI for Shopify COD orders. For beauty specifically, the payoff is that campaigns optimise toward buyers who pay, which matters more in this category than most because impulse purchases are common. After switching, expect reported purchases in Ads Manager to fall, since you have stopped counting orders that never became revenue, and agree that with finance before launch.
Repeat purchase and the first-order trap
Beauty has a structural advantage over many D2C categories: products run out. A cleanser or sunscreen bought in March may be bought again in May. That means first-order CAC can sensibly run higher than first-order margin, provided you have evidence of repeat purchase, not hope. Measure it by monthly cohort from your own order data: what share of first-time buyers order again within 60, 90 and 180 days, and what they spend. Then set acquisition targets on payback over a period your cash position can handle. Two cautions. Discount-led acquisition tends to bring in buyers who repeat only when discounted, so check cohort repeat rates by the offer that acquired them. And repeat purchase is driven by product experience and retention channels such as email, SMS and WhatsApp as much as by ads, so performance marketing should not be credited with, or blamed for, all of it. The LTV to CAC calculator on this site is a quick way to sanity-check the numbers once you have them.
Sources
ASCI, The ASCI Code and guidelines: https://www.ascionline.in/the-asci-code-guidelines/ . Social Samosa, summary of the ASCI influencer guidelines (May 2021): https://www.socialsamosa.com/2021/05/all-you-need-to-know-about-asci-influencer-marketing-guidelines . Press Information Bureau via Department of Consumer Affairs, Endorsement Know-hows! (20 January 2023): https://consumeraffairs.nic.in/sites/default/files/PIB1892527.pdf . Meta for Developers, Conversions API server event parameters: https://developers.facebook.com/docs/marketing-api/conversions-api/parameters/server-event
FAQ
Yes. ASCI's influencer guidelines require an upfront, prominent label such as #ad, #collab, #promo, #sponsored or #partnership, and the Department of Consumer Affairs' 2023 endorsement guidance requires clear disclosure of any material connection with the brand. Brands should write disclosure into creator contracts, since the advertiser shares responsibility.
There is no universal number. Watch for rising frequency, falling click-through rate and rising cost per purchase on an ad that has not otherwise changed. Log launch dates and the point each ad started to decline, and within a few months you will have a brand-specific refresh rhythm to plan production around.
Not at the moment they are placed. Send a lighter event at checkout, then send Purchase through the Conversions API once the order is confirmed or delivered. Prepaid orders can send Purchase immediately. This stops Meta learning from orders that later become RTO.