Blinkit vs Zepto vs Instamart: the quick commerce maths for FMCG brands

Quick commerce stopped being a delivery story a while ago. For an FMCG brand it is now a shelf you rent, a media auction you bid in, and an operations commitment you either meet or get demoted for missing. The upside is real and so is the way it punishes a thin margin.

This is the arithmetic, platform by platform, with the caveat that fee structures on these platforms move fast and vary by category. Every number below is attributed to its source and its date. Confirm current terms with the platform before you budget against them.

Why the money is moving

Storyboard18 reported on 2 June 2026 that a Datum Intelligence projection has Blinkit, Zepto and Instamart alone generating nearly ₹4,900 crore in advertising revenue during the current calendar year, with Datum's Satish Meena putting total quick-commerce advertising across categories at ₹5,000 crore to ₹6,000 crore annually. An executive told the same publication anonymously that for many FMCG and impulse categories, between 10% and 25% of digital performance marketing budgets are already shifting to quick commerce.

The same report carries the demand-side numbers from earnings calls: Dabur described quick commerce as the primary driver of its online business in Q4 with 54% growth, Tata Consumer put quick commerce and e-commerce at 21% of its domestic business growing 62% year on year, and Britannia said nearly 70% of its e-commerce business now comes from quick commerce.

None of that means the channel works for you. It means the auction you are entering has well-funded competition in it.

The three platforms are not the same product

The differences that matter to a brand are the commercial model, the data you get and the money you have to commit before you learn anything. The figures below are as reported by Global Websters in a practitioner guide dated 16 March 2026.

Blinkit

Self-serve through Brand Central, with SKU-level and city-level reporting through the Seller Hub. The listing fee is reported at ₹25,000 per SKU per state, returned in full as ad credits, which makes it closer to a media buy than a shelf rental. Ad delivery is inventory-led: your sponsored product only appears in pin codes where the nearest dark store physically holds stock, which is the usual explanation for ads not serving despite wallet balance.

Zepto

A vendor and purchase-order model with no self-serve portal, so listing and advertising need direct coordination with category teams and a longer runway, reported at four to eight weeks. The bundled entry package is reported at ₹5 lakh to ₹6 lakh, with an optional Atom analytics subscription at ₹30,000 a month for share-of-voice and competitor tracking. Zepto's Swap and Save format, which offers your product as an alternative when a shopper adds a competitor's item to cart, has no direct equivalent elsewhere.

Swiggy Instamart

Sold as a quarterly brand pack reported at ₹8 lakh to ₹10 lakh, bundling display, push notifications and featured category placement, with the structural advantage that a large share of shoppers arrive from the Swiggy food-delivery habit inside the same app. That helps brands with existing recognition in Swiggy's base and does noticeably less for brands with none.

The margin gate

Here is the number that decides everything else. The same practitioner guide puts the requirement at 60% to 70% gross margin for quick commerce to absorb platform commission, fulfilment fees, storage and advertising and still deliver a positive contribution. Confetti's eligibility guide dated 27 June 2026 reports platform commissions in a range of 8% to 25% depending on category, and describes how category managers evaluate an application: competitive pricing within the category, margin viability after commission, and MRP parity across Blinkit, Zepto, Instamart, Amazon and offline retail.

Model it on one SKU before you apply. Take your MRP, subtract the commission at the top of the reported range for your category, subtract fulfilment and storage, subtract the ad spend per unit you would need to win placement, and see what is left. If the number is negative at a realistic monthly volume, more advertising accelerates losses rather than growth. Category managers can estimate your cost structure from your MRP and category benchmarks, and they will not onboard a brand set up to lose money.

The ROAS trap

Worth knowing before you scale anything. Global Websters reports that both Zepto and Swiggy Instamart frequently calculate dashboard ROAS using maximum retail price rather than the actual selling price after platform discounts, which can overstate the return substantially. Recalculate manually on net realisation before you make any decision to increase spend. A number that flatters by tens of percent will happily fund a losing SKU for a quarter.

What it takes to get listed at all

Before any commercial conversation there is a compliance gate, and it is the same across the three platforms. Per Confetti's June 2026 guide:

• GST registration, unconditional, because the platforms collect 1% tax at source under Section 52 of the CGST Act. Selling as a private individual is not accepted.
• An FSSAI licence for food and related products, at the tier matching your turnover, with the 14-digit licence number printed on the primary consumer unit. Dark store inward teams check this physically. Platforms are also reported to require six to twelve months of remaining validity at application.
• A GS1-registered EAN-13 or UPC-A barcode on every unit, placed on a panel that is visible when the product is shelf-stacked.
• Shelf life at inwarding: Blinkit is reported to require 90 to 120 days remaining, with Zepto and Instamart at roughly 60% of total shelf life.
• Replenishment capability. Fast-moving categories need restocking two to three times a week in high-velocity dark stores, and platforms are reported to expect 95% to 98% in-stock availability as a performance target, with algorithmic demotion for brands that miss it.

That last point is the one D2C founders underestimate. Quick commerce is an operations commitment dressed as a marketing channel.

When it works: the Meta to Blinkit bridge

Lal Sweets is a ₹1,500 Cr business built almost entirely offline, a household name in sweets with the distribution and recall that comes with it, and almost no view of D2C performance as a channel. Paid media amounted to basic catalogue ads and seasonal offers. The work ran across Meta, Amazon, Blinkit and the brand's own website, with the content stack modernised, the funnel rebuilt on a narrative-led AIDA approach, the lineup segmented by regional demand, and Meta-Blinkit collab ads introduced, a channel the brand had never touched.

Cost per purchase on Blinkit dropped to ₹19, with Meta-Blinkit ROAS touching 10×.

Read that honestly rather than as a benchmark. A cost per purchase that low means the advertising was not doing all the persuading alone. It was catching intent that already existed and shortening the path to it, for a brand with recall built over decades and distribution already in place. The mechanism is the transferable part: the ad creates the craving and the platform behind it satisfies it in the same session, closing the gap where most food advertising loses the sale. A new brand with no recall does not get that discount, and should not budget as though it will.

When it does not work

• Thin margin. If you cannot clear the commission plus fulfilment plus advertising and still contribute, the channel is not a growth lever, it is a faster way to lose money.
• No hero SKU. The channel rewards concentration. Pushti Organics grew 570% in four months once ghee was made the hero and grew to over 65% of sales, and on Soothys two hero products came to drive over 60% of sales. A wide catalogue with no proven seller is the wrong shape for a dark store and for the ad auction.
• No replenishment capacity. A six-week manufacturing lead time cannot support weekly dark store restocking, and missed fill rate is punished algorithmically.
• Low purchase frequency. Quick commerce curates around products bought often and decided on quickly. A considered, occasional purchase does not fit the shelf.
• No recall at all. Bundled visibility packages reward brands consumers already recognise. If nobody knows you, a placement is an expensive way to find that out.

A sane sequence

The common operational mistake is launching all three platforms at once, which guarantees all three are sub-optimal. The sequence in the March 2026 guide is Blinkit first, because self-serve gives full data visibility from day one, with one to three hero SKUs in one metro city, and the listing-fee credits deployed as a performance campaign on launch day rather than saved for later. Only once fill rate and contribution margin hold does Zepto make sense, carrying keyword, city and SKU learnings across. Instamart comes third, and only where the category genuinely overlaps with a food-delivery audience.

Whatever the platform, review ROAS at SKU level in individual cities weekly rather than by campaign or category average. Problems compound quickly on this channel, and a campaign that looks profitable in aggregate routinely hides SKUs losing money in specific cities.

Offline scale does not transfer online by itself. A ₹1,500 Cr brand still needed funnel thinking, regional targeting and modern creative before its recall converted.

If you want this modelled properly against your own numbers, that is performance marketing and strategy working together. More of the category sits under FMCG and food and beverage.