Your Amazon ACoS is lying to you.

Walk into any seller's dashboard in India and the first number on the screen is ACoS. Founders quote it on calls. Agencies pitch it in decks. People win and lose budget over a tenth of a percent. And almost none of it survives contact with reality.

We've run paid media for 20+ Indian brands across Amazon, Flipkart, Quick Commerce and Meta. The pattern is the same every time: the brands obsessing over ACoS are the ones leaving the most money on the table. Not because the metric is wrong — but because it's partial. Watching only ACoS is like flying a plane with just the airspeed indicator. Useful, but you'll fly into a mountain.

Here's what's actually broken, and what to track instead.

What ACoS measures (and what it quietly hides)

ACoS = ad spend ÷ ad-attributed sales. That's it. A 20% ACoS means you spent ₹20 to make ₹100. Lower is "better."

The trap: "ad-attributed sales" is a fence Amazon draws around clicks within a 7-day window. Anything outside that fence — the customer who saw your ad, didn't click, but bought your listing organically two days later; the search traffic that came in because you climbed to page one; the Diwali shopper who searched your brand name because they remembered the Sponsored Brand video — none of it is in your ACoS. So your ACoS can be "great" while your business is shrinking, or "terrible" while it's exploding.

We've seen both. A skincare brand we audited last quarter had a 14% ACoS — beautiful number — and total revenue was down 22% YoY. Their ads were efficient at harvesting demand they already had. Nothing was generating new demand. Meanwhile a Quick Heal SKU we run sat at a 38% ACoS for six weeks while organic search rank climbed from page 4 to position 3 and total category revenue tripled. The "ugly" ACoS was the smartest money the brand had spent that quarter.

Metric 1: TACoS — the only number that has to go down

TACoS = total ad spend ÷ total revenue (ads + organic). It's the number that tells you whether your ads are leverage or a crutch.

If your ACoS is 25% and your TACoS is 6%, you have a healthy business: ads are a small slice of total revenue and organic is doing the heavy lifting. If your ACoS is 25% and your TACoS is also 24%, you're a business that exists only because the ad button is on. Turn it off and the lights go out.

Healthy TACoS bands we benchmark against in India:

• Year 1, building category presence: 10–18% TACoS is fine
• Year 2–3, scaling: 6–10% TACoS is the goal
• Established brand with strong search demand: 3–6% TACoS

If TACoS is flat or rising while you're scaling spend, you're not buying growth — you're renting it.

Metric 2: Brand vs non-brand ACoS — the lie within the lie

Your reported ACoS is a weighted average of two completely different businesses: people searching your brand name and people searching the category.

Brand search ACoS in India almost always sits between 4% and 12%. These are people who already know you and were going to buy anyway — the ad just shows up in front of intent that already exists. Counting it as "ad-driven sales" inflates your performance.

Non-brand ACoS — people typing "wireless mouse" or "neem face wash" — is the only number that tells you if you're winning new customers. It almost always runs 25–60% in competitive categories.

If you split the two, the picture flips. The brand we mentioned with the 14% blended ACoS? Brand search was 6%. Non-brand was 41%, with shrinking volume. They weren't growing — they were re-buying their own existing audience and calling it marketing.

Metric 3: Organic halo — what the ad bought after it stopped clicking

Amazon's algorithm rewards velocity. Run ads on a listing that converts well, and your organic rank moves up. Move up, and you start getting free traffic that doesn't appear in your ACoS at all because it isn't ad-attributed.

That's the halo, and it's the entire point of running ads on a marketplace. If you only judge ads by ACoS, you'll cut the campaigns that are doing the most for your organic position and keep the ones that are just intercepting people who'd have found you anyway.

How we actually measure halo: we tag a control set of SKUs that we don't run ads on, track organic BSR weekly, and compare to ad-supported SKUs of similar profile. If ad-supported SKUs are gaining BSR rank 2–3x faster than control, the halo is real and the ACoS number alone is undervaluing the spend by 30–50%.

Metric 4: Contribution margin per order — the only profit signal that matters

ACoS is a percentage. Percentages don't pay salaries. Two SKUs at 30% ACoS can have wildly different contribution margins once you stack referral fee, closing fee, weight handling, GST and storage on top. One pays for the office. The other quietly bleeds.

We model every SKU's net contribution per order — payout from Amazon minus COGS minus shipping minus ad spend allocated by units sold. If contribution per order is positive and rising as we scale, ACoS can drift up and we don't care. If contribution is shrinking, even a "great" 18% ACoS means we're scaling a loss.

There's no shortcut here. You need a per-SKU P&L that updates weekly. We built our break-even ACoS calculator and the Amazon India fee calculator for exactly this reason — start there if you don't have one.

The four-metric stack we run for every client

That's it. Four numbers. Look at them in that order and you'll never make a stupid spend decision based on a single ACoS line again.

What to do tomorrow morning

1. Pull your last 90 days of ad spend and total marketplace revenue from Amazon and Flipkart. Calculate TACoS week-on-week. Is it rising, flat, or falling?
2. Split your campaigns into brand and non-brand reports. If you can't, your campaign structure is the problem — fix that first.
3. Pick three SKUs you don't run ads on. Track their BSR for two weeks. Compare to three SKUs you do run ads on. The delta is your halo.
4. Build per-SKU contribution. Net payout − COGS − ad spend ÷ units. Sort descending. Cut spend on anything below ₹0.

None of this is hard. It just requires willingness to stop staring at a single number that makes you feel either great or terrible about a business that's much more complicated than that.

If you'd rather not build it yourself — fair, that's our job — drop us a line. We'll send back a one-page diagnostic on your account before you commit to anything.