How we count dark stores in India

Updated October 2026

What we count

We count a location when a quick delivery app uses it to pack orders. For Blinkit, Zepto, Swiggy Instamart, BigBasket BB Now, Flipkart Minutes and Amazon Now, these are dark stores: small stores stocked for app orders, almost always closed to walk-in shoppers. Each store counts once for each app it serves, and stores that are closed or only planned are left out.

Amazon Now counts in full. Next to its own dark stores, it uses some partner More stores as dark stores, and we count those too, because they pack Amazon Now orders like any other. Amazon Now joined our totals in October 2026, so earlier reports count five apps.

JioMart is different. It delivers from Reliance Stores, so we count them separately and show the figure next to the dark store totals, never inside them.

Where the data comes from

• The apps’ own store listings

  Our city, area and store pages start from the stores the apps show publicly. Each store gets its app, city, area and PIN code, and a street address where one is available.

• Company results and filings

  Eternal (Blinkit) and Swiggy (Instamart) report store counts each quarter, and Zepto shared its count in its IPO papers. These are the best national totals, though they come with a lag and no city split.

• Announcements and news

  For launches, closures, fee changes and new rules, we use company announcements and business news, always with the date attached.

How we clean the data

Raw store data is messy: the same store turns up twice, one locality is spelt three ways and PIN codes go missing. Before a number reaches the site, every store goes through the same checks.

1. Duplicates are merged, so one store is counted once.
2. Records that don’t add up are held back, such as a store whose name belongs to a different app, until we can confirm them.
3. Places get one name. Different spellings of a locality, and old city names like Bangalore or Gurgaon, are matched to one name.
4. PIN codes are checked against the store’s city, and we only change one when two sources agree.
5. Store names are made readable, without company suffixes or internal codes.
6. Totals are compared with company figures. If an app’s total is far from what the company says, we find out why before publishing.

Why our count differs from company figures

The two numbers are taken at different times, so either can be higher. Companies count on the last day of a quarter and publish weeks later. A new store can appear in our count first, while a store the company already counts may not yet show up publicly. On each app’s page we show both figures side by side.

How to read our numbers

• City totals add up the six dark-store apps, Amazon Now included. JioMart’s Reliance Stores are shown alongside, not added in.
• Rounded figures appear where we don’t have an exact count: “~120” means about 120, “<5” fewer than five and “5+” five or more.
• Cities per app can be higher than a company’s own city count, because we list suburbs and nearby towns separately.
• City tiers are based only on how many dark stores a city has, not the census grouping, so a city we call Tier 2 may be Tier 1 elsewhere.

How often we update

We refresh store and city counts regularly, and every data page shows the month its figures were last reviewed. Company totals are added when new results come out, and fee changes, launches and new rules are added once we confirm them. Each dark store report sums up what changed.

What our numbers can’t tell you

Store listings change every week, and we can’t visit every store. Treat our city and area counts as a strong guide rather than an audit. A store near you also doesn’t guarantee delivery at every hour, because apps pause orders when riders are busy or stock runs out.