ROAS benchmarks for D2C food and beverage brands in India

Type "good ROAS for D2C India" into any search box and you will get confident numbers with nothing behind them. No sample size, no time window, no category, no note on what the figure counts. We cannot fix the internet, but we can show our own book and let you judge it.

Where these numbers come from

This is not a market study. It is our own client list. The Shizz has published 21 client engagements as case studies, and every figure below is taken from one of those pages, linked so you can read the full context yourself.

The sample is self-selected twice over. These are brands that decided to hire a performance agency, which already sets them apart from the average Indian D2C brand. And they are the engagements we chose to write up. Both of those bias the set upward. Read what follows as evidence from one agency's book over six years, not as a picture of the Indian market.

Three published cases are excluded from every figure, range and claim on this page because their numbers are not confirmed to our satisfaction. We would rather publish a smaller honest set than a larger flattering one.

Every figure carries an n. Where a case page does not publish a number, we say that instead of estimating one.

The ROAS levels we publish

Account-level ROAS, stated on the case page. n=5.

• Pro Nature, 10 months: a steady 8× ROAS, up from 1.2× at the start. Read the case.
• Pure Whites, 8 months: 4× in the first four months, climbing past 5× three months later. Read the case.
• Vediko Origins, 8 months: consistently above 4×, with some campaigns past 5.6×. Read the case.
• Pushti Organics, 4 months: held at 3.5× to 4× the entire way. Read the case.
• Aazol, 4 months: 2.7×, described on the case page as close to doubled. Read the case.

The published band across those five is 2.7× to 8×. That is a range, and we are deliberately not turning it into an average. Five engagements at different price points, in different sub-categories, over windows of four to ten months do not share a denominator. A mean of them would be a number with no meaning attached, which is exactly the kind of figure this post exists to argue against.

Channel-specific ROAS rather than an account figure. n=2.

• Parasbaagh: 3× ROAS on Meta, at daily spends that stayed sustainable, alongside sales up 300%. Read the case.
• Lal Sweets: Meta-to-Blinkit ROAS touching 10×, with a ₹19 cost per purchase on Blinkit. This is a quick-commerce collab-ad figure for an established ₹1,500 Cr brand and it does not transfer to a startup. We take it apart properly in the quick-commerce economics post. Read the case.

Movement published, level not published. n=3.

• Kroslo: ROAS doubled in 25 days on the same budget. Read the case.
• Anuttama: ROAS doubled while revenue climbed 200%. Read the case.
• Amyra Farms: ROAS doubled in three months while CAC dropped more than 50%. Read the case.

ROAS held as spend grew, no figure published. n=3.

• My Pahadi Dukaan: ROAS held strong as spend scaled, on the way from ₹12 lakhs to ₹1.2 crores a month over eight months. Read the case.
• Zama Organics: ROAS stayed high even as significantly more spend went in, while D2C tripled. Read the case.
• Bon Fiction: ROAS targets hit consistently rather than once, with online revenue up over 6× in five months entirely on paid media. Read the case.

Now look at the shape of what is missing. Across all thirteen we publish exactly one starting ROAS: Pro Nature's 1.2×. Everything else is an endpoint or a multiple. So we cannot tell you what a typical Indian food brand's ROAS looks like before anyone intervenes, and nor can anyone else who is working from case studies, because case studies are written after the intervention worked.

For a much wider denominator: across 160+ brands and ₹150 Cr+ of managed ad spend over six years, our average is 3.8×. That spans far more engagements than the 21 written up here, including plenty never published. It is the one number we would defend if forced to pick a single figure, and it is still an average across categories that should not be averaged.

Why a single "good ROAS" number misleads

1. Contribution margin sets the ROAS you need, not the other way round

A 3× ROAS can be excellent or loss-making depending entirely on what is left after cost of goods, discounts, shipping and returns. Sunitha Viswanathan of Kae Capital made the point cleanly to Inc42: a brand can proudly show 3× on a campaign while the real profit margin underneath is 5%, once COGS, discounts and returns come out (Inc42, 21 July 2026).

The same arithmetic runs in the other direction on quick commerce, where the platform takes a larger cut before you see anything. Global Websters puts the threshold for a product with 65% gross margin at a blended 4× to 6× simply to stay contribution-positive after commission, fulfilment, storage and ad spend (Global Websters, 16 March 2026). Confetti puts quick-commerce platform commissions at 8% to 25% depending on category (Confetti).

So the honest first question is never "what ROAS should we hit". It is "what ROAS keeps us contribution-positive at our margin, on this channel". That number is yours and nobody else's.

2. Price point and category change the maths before any ad runs

Look at what sits behind the high figures in our set. Pro Nature, Pushti Organics, Pure Whites and Vediko Origins sell ghee, honey, oils and organic staples: relatively high ticket, decent margin, and genuinely replenishable. Pushti's ghee alone grew to over 65% of sales, which concentrated spend behind one proven seller instead of spreading it across a catalogue.

A clean-label chocolate brand competing on a crowded impulse shelf is playing a different game at a different ticket size. Comparing their ROAS numbers as though they were the same measurement is the core error in every unsourced benchmark table.

3. Repeat frequency quietly subsidises the number

ROAS is usually measured on first purchase, so any revenue arriving later is invisible to it. Pure Whites now takes about a third of revenue from repeat orders while adding 550+ new customers a month, which means a large slice of its economics never touches the ROAS figure at all. Pushti Organics cut CAC 28% with founder-led video while lifting repeat purchases at the same time.

Two brands with identical reported ROAS and different repeat rates are not equally healthy businesses. One of them can afford to bid more.

4. Channel mix moves it more than bid strategy does

Our highest published figure and our lowest come from different channels doing different jobs. Lal Sweets' Meta-to-Blinkit collab ads catch an intent that already exists and shorten the path to it, for a brand with decades of recall. Parasbaagh's 3× on Meta is prospecting, which has to create the demand before it can convert it. Aazol's engagement was largely about moving from Meta-heavy to a balanced Meta and Google engine, and its ROAS reads differently before and after that shift.

If you benchmark a prospecting number against a demand-capture number, you will conclude your account is broken when it is merely doing a harder job.

5. Check what the dashboard is actually counting

Two failure modes are worth naming. First, on quick commerce, Zepto and Instamart dashboards frequently calculate ROAS on MRP rather than the actual selling price after discounts. Global Websters works through an example where a ₹500 MRP product selling at ₹340 reports a ROAS roughly 47% higher than the real return. Second, attribution windows and view-through settings differ between platforms, so two accounts with the same true performance can report very different ROAS.

Inc42's reporting captures where this is heading: ROAS is becoming a diagnostic for whether a specific creative is working, rather than the number that decides whether to scale. Brands are moving to marketing efficiency ratio, blended CAC and contribution margin after marketing, and are setting aside 20% to 30% of budget for retention and brand. Renée Cosmetics, quoted in the same piece, now takes 65% of sales offline, so a website ROAS measures a small slice of the business.

How fast ROAS actually moves

The two ends of our set explain the mechanism better than any average could.

Kroslo doubled ROAS in 25 days on the same budget. Nothing new was bought. The money simply stopped going to loose targeting, tired angles and traffic that fell out at the last step. Fixes that remove waste pay immediately.

Pro Nature took ten months to go from 1.2× to 8×, because that job was a rebuild: new creative, intent-segmented audiences, a tightened path to checkout, spend concentrated on hero SKUs, and retention flows. Fixes that create new demand compound slowly and then hold.

If someone promises you Pro Nature's endpoint on Kroslo's timeline, they are selling you the waste-removal work and pricing it as the rebuild.

A ROAS number without a margin, a channel and a time window attached is a rumour with a decimal point.

What to do with this instead of copying a benchmark

• Calculate the ROAS floor your own contribution margin requires, per channel, before you look at anyone else's figure.
• Split reporting into demand capture and demand creation. Judge each against its own job.
• Recalculate any platform-reported ROAS on net transaction value, not MRP.
• Track blended numbers alongside platform ROAS, because repeat revenue never shows up in the latter.
• Judge movement over a window, not a single week. Waste removal shows up in days, rebuilds in months.

If you want this run properly rather than argued about, that is what our performance marketing work is, and most of the engagements above sit inside food and beverage and FMCG.