Amazon Ads in India: Every Ad Type, and When to Skip It

The same rupee, two routes: Amazon owns the checkout, your own site owns the customer.

If you already sell on Amazon.in, running ads there is not a leap. Your catalogue is live, the shopper is already in a buying session, and the checkout is not yours to fix.

That is exactly why the question gets skipped. Most brands switch Sponsored Products on because it is easy, then spend a year without knowing whether that money would have done more on their own site. This post sits inside our wider performance marketing guide, and it is about the trade-off nobody articulates: not whether to advertise on Amazon, but where you want the sale to happen.

One boundary first: listing quality, images, A+ content and organic marketplace ranking live in our Amazon marketing in India guide. This page is the ad buy itself — the ad types, the economics and the reporting.

What Amazon ads in India actually are

Amazon Ads is a self-service platform run from the advertising console, and Amazon states plainly that businesses that do not sell on Amazon can still use it — you register and select "Products and services not sold on Amazon". Most Indian advertisers, though, are sellers advertising their own listings.

On the India seller side, Amazon describes Sponsored Products, Sponsored Brands, Stores and Sponsored Display, and confirms that "Sponsored Products, Sponsored Brands, and Sponsored Display (beta) are cost-per-click ads. So, you only pay when your ad is clicked." Stores — your multi-page brand destination on Amazon.in — are free to use.

Two of those names are in transition globally. Amazon's product page now says Sponsored Display "is now a part of Amazon's wider display ads offering", and the Sponsored TV page says it "is now known as streaming TV ads" — a rename, not a replacement. Amazon's India seller page still uses the older labels, so check what your own console shows before you plan around a name.

The real question: where should the sale happen?

Answer this before you open a single campaign setting, because it decides how much Amazon advertising is worth to you.

Sell on Amazon and you get demand you did not have to create — in exchange for three things. You pay a referral fee on the order. You do not get the buyer as a customer record. And your creative, pricing and layout sit inside somebody else's template, next to a competitor's ad.

Sell on your own site and you keep the full margin and the customer data, and can build repeat purchase against both. But you manufacture the demand yourself, your checkout has to be as good as Amazon's, and first-purchase trust is earned rather than borrowed.

Neither is the right answer universally. The useful version of the question is per-product:

• Commodity, price-compared, low-consideration items. Amazon usually wins. The shopper was going to buy the category on Amazon anyway; your only real choice is whether it is your listing or someone else's.
• Products with a repeat-purchase or subscription shape. Your own site usually wins, because lifetime value is the whole business case and you cannot compute it on customers you never receive.
• New brands with no awareness. Amazon is often the cheapest place to acquire the first honest review base, then you migrate the repeat demand off-platform.

Most Indian D2C brands should run both. The mistake is running both without deciding which job each SKU is doing — then comparing an Amazon ACoS against an own-site ROAS as if the two numbers meant the same thing. They do not: one is measured after Amazon takes a cut and keeps the customer. Compare on contribution after fees instead. Our Meta ads vs Google ads for D2C comparison covers the own-site half of that decision.

The Amazon ad types, and who can actually buy each one

Eligibility matters more than most briefs admit — two of these are closed to a seller without Brand Registry, and one is a different buying motion entirely.

Sources for those rows, in Amazon's own words: Sponsored Products eligibility and placements (India is a listed marketplace, and campaigns have "no monthly or upfront fees"), Sponsored Brands formats and eligibility, and Amazon DSP. Note DSP's commercial gate: Amazon's managed service states a minimum spend of "$50,000 USD" and adds that this may vary by country. No India-specific DSP minimum is published. For most Indian sellers, DSP is a conversation with a rep or a partner, not a button.

Sponsored Brands also carries the metric most worth watching on Amazon: new-to-brand orders and sales, measured on a 12-month lookback. Amazon itself says high ROAS should not be your primary success metric for this format — an unusually honest line from a platform, and a correct one.

Ad spend is only half of what Amazon costs you

Your ACoS is not your economics. Amazon defines advertising cost of sales as "(ad spend ÷ ad revenue) x 100", and it is deliberately unhelpful about benchmarks: "There isn't a definitive number for a good Amazon ACOS." Amazon's own guidance is to work from your break-even — your ACoS needs to sit below your profit margin.

The margin you measure against has to be the post-fee one. At the time of writing, Amazon.in charges a referral fee calculated as item price × a category percentage, ranging from 0% up to 30% depending on category and price band, plus a closing fee starting at ₹2 and, if you use Easy Ship or FBA, weight handling starting at ₹37 per item. Those figures move — check the live page before you model anything.

So a 20% ACoS on a product carrying a 15% referral fee, a closing fee and fulfilment is a very different business from the same number on your own store. Run both through one contribution model — our ROAS calculator sanity-checks the own-site side before you compare. When brands ask us to audit marketplace spend, this is usually where the answer already is, and it is the first thing we rebuild in a performance marketing engagement.

What reporting you actually get — and what you don't

This is the part buyers almost never ask about, and the part that decides how much control you really have.

What you get is genuinely good for optimisation. Campaign, ad group and keyword performance. A search term report showing the real shopper queries that triggered your ads — raw material for both your negative list and your listing copy. Placement data. New-to-brand metrics on Sponsored Brands.

What you do not get is a customer. No buyer email, no ability to build a first-party audience from an Amazon sale, no cross-channel journey that includes your own site. Amazon reports the sale; it does not hand over the person.

The partial fix is Amazon Attribution, which Amazon describes as free and which is available in India. It measures the on-Amazon impact of your non-Amazon marketing — your Google, Meta, email and display traffic that converts on Amazon — using a 14-day attribution window. Eligibility is the usual Brand Registry gate: professional sellers in Brand Registry, vendors, KDP authors and agencies.

If you are already pushing Google or Meta traffic toward Amazon listings without Attribution switched on, you are blind on the half of your spend that is hardest to justify. Set it up before you argue about budget splits.

How to launch your first Sponsored Products campaign properly

The most common failure here is not a bidding failure. It is advertising a listing that was never going to convert — paying for clicks onto a page with two photos and no reviews. Amazon publishes its own readiness bar, so use theirs rather than a guess.

1. Clear the catalogue-hygiene bar first. In Amazon's Sponsored Products best practices: be in stock and hold the featured offer, advertise products with five or more reviews at 3.5 stars or higher, use four or more zoomable images at least 1,000 pixels on a side, keep titles around 60 characters, and include at least three bullet points. A + content on top of that. If a SKU fails this, fix the listing — do not fund it.
2. Pick two or three ASINs, not the whole catalogue. Products that already sell organically and carry margin. A small budget spread across forty listings teaches you nothing about any of them.
3. Start with an automatic campaign to learn. Amazon calls automatic targeting "the easiest and quickest way to get started"; it runs close match, loose match, substitutes and complements. Its job is discovery, not efficiency.
4. Set a budget you will actually leave running. Amazon recommends starting at "$10 or equivalent" a day — call it under ₹1,000 — and taking an always-on approach with no end date. There is no published minimum. A campaign paused every third day never gathers enough data to be read.
5. Mine the search term report, then go manual. After a fortnight, promote the terms that converted into a manual campaign as exact and phrase match, keeping broad separate for continued discovery.
6. Choose a bid strategy deliberately. Amazon offers dynamic bids down-only, dynamic bids up-and-down, and fixed bids, plus placement bid adjustments of up to 900%. Start down-only while you are learning; the aggressive settings amplify a bad campaign as happily as a good one.
7. Make negatives a weekly habit. This is the discipline that separates a profitable account from a leaking one. Add negative keywords for irrelevant queries, and negative products and brands so your ads stop appearing where they cannot win. Ten minutes a week, forever.
8. Judge against break-even, not a borrowed benchmark. Amazon publishes no India ACoS benchmark, and any agency quoting you one is quoting its own account average. Your number is your post-fee margin.

When Amazon ads are the wrong answer

Three cases where the money belongs elsewhere.

• Your listing is not retail-ready. Ads accelerate whatever is already true. A weak listing with ad spend behind it just loses money faster.
• Your product needs explaining. Amazon is a comparison environment, not a persuasion one. If the purchase needs education, a landing page you control will out-convert a marketplace listing.
• Repeat purchase is the entire model. If the second and third orders are where the profit sits, paying a referral fee on every one of them — and never owning the customer — is a strategic decision, not a media one. Choose it knowingly.

Running marketplace ads alongside retail ads on your own site is fine — just treat them as different buys with different economics. Google Shopping ads put the same product in front of the same shopper, but the sale, the data and the margin land with you.