Live commerce in India: what actually works here

Every live commerce pitch deck opens with China. Ignore it for a minute.

The real question is narrower. Can your brand hold a stranger's attention for 40 minutes, on a surface that lets them buy, often enough to justify the production load? For a few categories, yes. Mostly, not yet.

This page sits under our social media marketing guide, which covers the wider channel picture. Here we stay on the stream itself.

What live commerce is competing with for the viewer's time

A live stream is the most expensive thing you can ask of a viewer: fixed time, fixed place, no fast-forward. It competes with:

• Reels and Shorts — the same demo, 20 seconds, watchable at a red light.
• A creator's recorded review, which the viewer already trusts more than your brand channel.
• Quick commerce. If they want it now, a 10-minute delivery app beats your discount code.
• Cricket, YouTube and a group chat — free, and no appointment needed.

So the benchmark is not "is live commerce growing". It is "is a scheduled stream a better use of this fee than six Reels?" Often it is not. Live wins on questions — fit, shade, ingredients, assembly, authenticity — that a comment thread answers and a caption cannot.

Two doors closed, with dates.

Why the China playbook does not transfer directly

Three structural things differ. None are about culture or "readiness".

1. Platform integration. In China the stream sits inside the shopping app. In India the two biggest video audiences, Instagram and YouTube, are not marketplaces. TikTok — the one global app pairing short video with an in-app shop — has been blocked here since 29 June 2020, named first among 59 apps (Press Information Bureau, 29 Jun 2020). The most natural home for Indian live selling simply does not exist.

2. Creator economics. A live seller in China is a sales job on commission. In India most creator deals are flat fees per post, so the creator carries no downside if nobody buys. Shifting to commission changes who takes the booking — our creator economy breakdown covers those payment models.

3. Payment habits. Live selling depends on impulse: tap, pay, done. India is moving that way, not there yet. Prepaid orders reached 41% of Indian e-commerce orders in 2025, up from 32%, and only 34.5% in Tier III, per ClickPost data across 635 million orders (Social Samosa, 30 Dec 2025). A cash-on-delivery order placed in a hurry is also the likeliest to come back.

We have deliberately not quoted a China market size or an India forecast. The widely repeated numbers trace back to projections made years ago, for target years that have since passed.

Where an Indian brand can actually run live selling today

Check the surface before planning the stream. Several platforms launched live shopping, then pulled it.

Sources: YouTube Help on live product tags and Shopping on YouTube; PPC Land, 29 Jun 2025; Inc42, 1 Oct 2022; afaqs!, 3 Aug 2022; Agency Reporter, 22 Jul 2026; Firework.

The withdrawals are part of the data

Any post telling you to tag products in an Instagram Live was written before 2023. The pattern matters more than any single shutdown:

• Facebook Live Shopping closed on 1 October 2022; Instagram stopped product tagging in Live on 16 March 2023 (TechCrunch, 14 Feb 2023).
• Simsim, the Indian live social commerce app YouTube bought in mid-2021 at above $70 million, was shut down under two years later (TechCrunch, 23 Mar 2023).
• ShareChat's live commerce vertical was wound down from December 2022; the company called it a scale-down (Inc42, 15 Feb 2023).
• Amazon Inspire, its shoppable short-video feed, was discontinued in February 2025 (Tubefilter, 19 Feb 2025).

What survived: marketplace-owned live, creator-affiliate live on YouTube, and live on your own site. Build there.

The production reality of one session

A live session is a small broadcast, not a post. The work is front-loaded and barely cheaper the second time. Who must be there:

• A host on camera who can answer a sizing question without checking. A founder or trained category person beats a hired face.
• A second person on comments, feeding questions to the host. Without this the stream dies in silence.
• Someone on stock, confirming the SKU just promoted is actually available before the host promises it.
• Someone on ops for the hour after — live orders spike, with a high question rate.

Ready before you go live: a run-of-show with product order and timings, the offer and its end time, the landing page or tagged catalogue, stock reserved, promotion for the 48 hours before, and a plan for the recording — replays usually outperform live view counts.

And the thing most brands get wrong: one session is not a test. Your first stream measures your promotion, not your format — nobody has a habit yet. Treat session one as the cost of learning your run-of-show.

On scale: Business Today reported an Amazon India live creator, Dhananjay Bhosale, streaming once or twice a week to between 100 and 20,000 viewers per session, while Amazon India's own shows and livestreams reached about 4 million customers in the preceding year. The same piece called live commerce "still nascent in India" (Business Today, 4 Jun 2024). A brand channel starting cold should expect the low end.

Which categories suit live selling, and which do not

The pattern is consistent: live works when watching answers a question reading cannot.

Works well:

• Visual and demonstrable — cookware, appliances, tools, cleaning, craft supplies. You can show it working.
• Variant-heavy — sarees, kurtas, jewellery, lipstick and foundation shades. A host holds up 30 options in 30 minutes; a product grid cannot.
• Trust-sensitive — handloom, gemstones, collectibles, refurbished goods. Unedited video is the proof.
• Impulse-priced — the range a viewer buys without consulting anyone else.
• Drops and limited stock — scarcity is the only honest reason to watch live instead of the replay.

Works badly:

• High-ticket considered purchases with long research cycles. The buyer will leave to compare, and should.
• Regulated or claims-heavy categories — health, supplements, financial products. Unscripted live is a compliance risk.
• Single-SKU commodities. There is nothing to show for 30 minutes.
• B2B, where the buyer is a committee and nobody attends a stream at 8pm.

Not in the first list? Put the budget into recorded video and micro-influencer demos — most of the persuasion, none of the scheduling risk.

The minimum viable test

Directional, not a rate card. The shape matters more than the spend.

Measure these, not concurrent viewers:

• Revenue per session, plus revenue in the 72 hours after — live orders trail the stream.
• Average watch time. This tells you whether the format works at all.
• Comments per 100 viewers. A silent stream never converts.
• Add-to-cart rate from the stream versus your normal product-page rate.
• Return rate on SKUs sold live versus the same SKUs sold normally. More sales with more returns is not a gain.
• Replay views and orders. Often where the value sits.

Kill condition, stated upfront: if by session six revenue per session has not risen across at least three sessions, watch time is flat or falling, and returns on live-sold SKUs are worse than normal — stop. Keep the recordings as shoppable video and move the budget back into your e-commerce social media plan.

How we would scope a first live test — DH's own view

Our own position, from scoping video and commerce work for D2C brands. No client results implied.

For a first test we would deliberately under-build it. Six sessions, one surface, the founder or a category expert hosting, a second person on comments. No studio — a phone on a tripod, one soft light, a clean wall, and a connection tested at the actual time of day. We would use SKUs that already sell, because live does not fix a product nobody wants, and reserve stock so the host never promises something that is out.

Working through where the money goes: nearly all of it is people-hours and the creator fee, not kit. So we would not add a creator until session four — paying one to appear in a stream you have not figured out is the commonest way this budget is wasted. Social media management in India runs roughly ₹25,000–₹1,00,000 a month by scope; a live programme sits inside it, with creator fees and paid promotion separate.

When we would advise against it entirely. If any one of these is true, we say no and put the money into recorded shoppable video:

• Nobody internally can host credibly, and the budget only covers a hired face reading a script.
• Your price point needs a comparison decision, so the viewer has to leave to buy properly.
• Fulfilment is already strained. A live spike makes an ops problem public, in the comments.
• You cannot commit to six sessions. One stream is a content shoot with extra stress.
• Your category is regulated and claims need legal sign-off. Live is the wrong medium.

Where live selling fits in your next year

Treat live as a format you earn into, not a channel you launch. Build the audience on recorded video first, where a miss costs one post. Once one question keeps appearing in your comments — fit, shade, authenticity, how it works — that question is your first live stream, and you already know people want the answer.

That sequencing also keeps it affordable. The same host, SKUs and run-of-show feed your D2C marketing strategy, your festive calendar and your wider e-commerce marketing — whether or not live survives session six.