Social commerce is the part of ecommerce where discovery, and sometimes the checkout itself, happens inside a social app rather than on your website. Instagram product tags, WhatsApp catalogues and order confirmations, TikTok Shop, YouTube's shoppable links, Pinterest pins that carry live price and stock. If a shopper can go from seeing a product to paying for it without typing a URL, that's social commerce. If they see an ad, click, and land on your product page, that's just paid acquisition for a regular store. The distinction matters because the two need different work.
Most Indian D2C brands we talk to think they're doing social commerce because they spend eight lakh a month on Meta. They're not. They're buying traffic. Real social commerce is a set of surfaces you have to maintain: a product feed, a catalogue, a chat flow, a creator programme. Some of those pay back well. One or two are a waste of your quarter.
What's actually available in India, and what isn't
Native in-app checkout has never been broadly available here. Meta's checkout on Instagram and Facebook has been limited to a small number of markets, chiefly the US, and Meta has been trimming its shopping features rather than expanding them. So in India, a tagged product on Instagram is a very good link. That's it. The shopper still lands on your Shopify store and still has to get through your cart.
TikTok Shop, which is the strongest social commerce product anywhere right now, isn't an option in India at all. If you sell into the UK, the US or Southeast Asia, it's worth a serious look. If you sell only domestically, stop reading about it.
What you do have in India:
- Instagram and Facebook product tagging, driven off your Shopify catalogue sync
- WhatsApp catalogue plus a business number, which is where most of the actual conversation happens
- YouTube product links and affiliate-style creator programmes
- Reseller networks and marketplaces built on social sharing
- Live selling, mostly on Instagram Live and YouTube, still improvised rather than productised
WhatsApp is the channel nearly everyone under-builds
In India, the highest-intent social surface isn't a feed. It's a chat window. A shopper who DMs to ask whether the kurta runs small is worth ten who liked the reel. Yet most brands answer those messages from a phone, at night, badly.
What a properly built WhatsApp layer looks like: catalogue synced from Shopify, so prices and stock are current; abandoned-cart recovery going out as a template message forty minutes after drop-off; order confirmation, dispatch and delivery updates; and a COD confirmation message before the order is handed to the courier. That last one is the single highest-return automation available to an Indian ecommerce brand. Asking a customer to tap "yes, ship it" filters out impulse and fraud orders before you pay for a forward leg.
Costs are per message and marketing templates are the expensive category, so there's a real ceiling. A brand blasting every new-arrival drop to 60,000 numbers will find out fast. Utility messages tied to a real order are cheap and welcome. Keep the split honest.
The product feed decides more than your creative does
We've picked up accounts where ROAS jumped inside two weeks with no new creative, purely because the catalogue stopped being broken. This is unglamorous and it works.
Things that quietly break a Shopify-to-Meta feed:
- Variants missing size or colour attributes, so Advantage+ campaigns can't build sensible product sets
- Prices that don't reflect GST the way the checkout does, which gets items disapproved or, worse, served at the wrong price
- Out-of-stock items still marked available because the sync runs on a schedule your inventory doesn't respect
- Titles written for a human browsing your collection page instead of for a matching algorithm — "The Mira" tells Meta nothing
- No GTIN or brand field on products where you actually have them
Fix the feed before you touch bids. And if you sell jewellery or anything else where price moves daily, understand that a static feed will always be wrong by the time it's crawled. Rate-linked pricing has to be handled on the store first, then pushed out.
Live commerce: good for three categories, oversold for the rest
Live selling works when the product needs demonstration, has genuine variety, or benefits from scarcity in real time. Sarees and unstitched fabric. Gemstones and gold. Sneaker drops. Plants. In those categories a two-hour live with a decent host can do more revenue than a fortnight of static ads, because the objection being answered is "what does it actually look like" and video answers it.
For a skincare brand with nine SKUs, a live stream is an expensive way to talk to forty people. We've told founders to skip it and put the money into creator seeding instead. Nobody enjoys hearing that after they've bought a ring light.
Creators: pay for the link, not the follower count
The version of influencer marketing that shows up in P&L is affiliate-shaped. A unique discount code per creator, a trackable link, a commission on delivered revenue rather than on placed orders. Flat fees for a story slot are fine for launch noise, but they don't compound.
Two operational notes from doing this repeatedly. First, discount codes leak into coupon sites within days, so cap usage or use a landing page with an auto-applied cart. Second, pay on delivered revenue in a COD-heavy market or you'll be paying commission on parcels that come back. Write that into the agreement up front; renegotiating it after month one goes badly.
Does social commerce pay? Run the arithmetic before you scale
Here's the calculation we do in a spreadsheet with founders, using round numbers you can substitute your own into.
Say you spend ₹90,000 on Meta at a ₹9 CPC. That's 10,000 sessions. At a 1.8% conversion rate you get 180 orders. AOV ₹1,450, so ₹2,61,000 in placed revenue. ROAS looks like 2.9 and everyone is pleased.
Now apply reality. 65% of those orders are COD, so 117 parcels. RTO on COD runs 25% for a lot of apparel and accessory brands, so 29 orders never land. You're left with 151 delivered orders, or ₹2,18,950 in delivered revenue. Real ROAS: 2.43.
Then costs. Gross margin 45% gives ₹98,527 of product margin on delivered orders. Shipping and packing at ₹85 per delivered order is ₹12,835. Return legs on the 29 RTOs at ₹120 each is ₹3,480. Payment gateway on the 63 prepaid orders, roughly 2% of ₹91,350, is ₹1,827. Subtract those and the ad spend: 98,527 − 12,835 − 3,480 − 1,827 − 90,000 = ₹−9,615.
Negative. On a campaign that reported 2.9 ROAS. Cut RTO from 25% to 15% with a WhatsApp COD confirmation and prepaid nudges, and delivered orders go to 163, product margin to ₹1,06,357, and the same campaign clears roughly ₹−1,000 to break-even before any AOV work. That's the whole argument for building the chat layer before buying more clicks.
Fix the landing experience before you buy more social traffic
Social traffic is almost entirely mobile, largely on 4G, and arriving with about four seconds of patience. A product page that takes 3.2 seconds to paint its main image is throwing away the top of your funnel before any of your merchandising gets a chance. We routinely see stores where getting LCP from around 3s to under 1.8s moves add-to-cart rate by a fifth, and that is a cheaper win than any bid adjustment.
The usual culprits are the same every time: five apps injecting scripts on every page, a hero image served at 2400px to a 390px viewport, and a review widget that blocks render. Our Shopify speed optimization work starts by auditing what each installed app actually costs you in milliseconds, and a fair number get removed. If you want to see where you stand without engaging anyone, SwiftStore scans the store, fixes what it safely can and tracks the score so you notice when a new app quietly costs you 400ms.
Measurement: what to believe
Post-consent-changes attribution is soft everywhere. Meta will claim orders that Shopify assigns to direct or email. Both are partly right. Rather than arguing with the platform, we use two numbers: blended MER (total revenue ÷ total ad spend) as the trend line, and a periodic geo or spend-holdout test to check whether a channel is actually generating demand or harvesting it.
For social specifically, track delivered revenue per creator and per campaign, not placed revenue. Track COD share by campaign too, because some audiences skew heavily COD and are materially less profitable at the same reported ROAS. If your reporting doesn't separate those, it's telling you something reassuring and wrong.
Where to start if you're building this in the next 30 days
In order, because the sequence matters:
- Audit the Shopify-to-Meta catalogue sync. Fix variant attributes, stock status and titles.
- Stand up WhatsApp: order updates, COD confirmation, abandoned cart. Utility messages first, marketing later.
- Get PDP LCP under 2 seconds on a mid-range Android, on 4G, not on your office wifi.
- Move three creators from flat fee to commission on delivered revenue and compare over six weeks.
- Only then increase spend.
If you're selling into the Gulf, the same order applies but the COD arithmetic softens and prepaid share is higher, which changes what a viable ROAS looks like; our team in Dubai and the wider GCC works those numbers differently.
Pull your last 90 days of Shopify orders, split them by COD versus prepaid, and calculate delivered revenue per channel. If that number surprises you, book a free store audit and we'll go through the feed and the chat flows with you.