Two search boxes, two entirely different systems
This is the confusion that wastes the most money in this category. A customer searching inside the Swiggy or Zomato app is querying that platform's index, ranked by that platform's rules. A customer searching on Google is querying the web. They are unrelated systems, and the work that moves one does nothing for the other.
| Aggregator app search | Google search | |
|---|---|---|
| What decides your position | Distance from the customer, rating, prep and delivery time, order volume, and paid placement in the app's ad auction | Relevance, links, page experience, and the entity signals Google holds about your brand |
| What you can change | Operations, menu and photo quality, packaging, rating, ad spend | Your website, your content, your profile, your brand's search footprint |
| What SEO does here | Nothing. Different index, different rules. | Everything. |
| Cost per order | Commission, ad bids, and the discounts you fund | Payment gateway plus your own delivery cost |
| Who owns the customer | The platform | You — phone number, address, order history |
You start with zero branded search, and that's the real problem
A restaurant with a shopfront gets free brand awareness every day from people walking past it. A cloud kitchen gets none. Nobody types a name they've never seen, so <your brand> order online has essentially no search volume on day one and no amount of on-page work creates it.
Branded search has to be manufactured offline first, and then captured by search. That order matters — most delivery brands try it backwards, build a website, watch it get no traffic, and conclude that SEO doesn't work for them.
- A packaging insert with a short, typeable URL and a real reason to visit — money off the next direct order, or a menu item you only sell direct.
- A QR on the lid that opens your ordering site, not your Instagram page. Instagram is discovery; the site is revenue.
- One spelling of the brand name, everywhere. Two spellings splits your branded demand in half and neither half ranks properly.
- Instagram and local creators — for a new delivery brand in an Indian metro, this is genuinely where first discovery happens, and it's what creates the search that your site then catches.
- A WhatsApp opt-in at the first direct order. The cheapest repeat-order channel available to you, and it costs a checkbox.
The Google Business Profile question, and the suspension risk
Google's local listings assume a place a customer can reach. Delivery-only kitchens can be listed as service-area businesses with the address hidden, which is legitimate. What is not legitimate, and what gets whole clusters of listings pulled at once, is the standard cloud-kitchen setup: five virtual brands sharing one kitchen, each with its own listing at the same address, several of them with keywords in the name.
That pattern is one of the easiest things for Google's systems and for a competitor to spot. When it's actioned, it's usually actioned across every listing at that address, including the one you genuinely deserved.
- One listing per genuinely distinct business at an address. If four brands run from one kitchen with one staff and one licence, four listings is the pattern that gets flagged.
- Set it up properly as a service-area business — hide the address, set a delivery radius that matches reality, and use kitchen hours rather than aspirational ones.
- The name field is your registered brand name and nothing else. 'Biryani House — Best Biryani in HSR Layout, Free Delivery' is a guideline breach a competitor can report in a single click.
- Point the ordering link at your own site, not at the aggregator. That field is free traffic to a page with no commission on it.
- Real photos of real food you currently sell. Stock photography is obvious to customers and useless to you.
Cuisine and locality pages: what wins, and what doesn't
Here's the honest version, which is not what most proposals will tell you. The head terms — biryani near me, best pizza in <city> — are held by aggregators, the map pack and review sites. A delivery brand's website is not taking those in year one, and any plan that promises it is really a plan to publish a hundred templated locality pages. That's the fastest route to a scaled-content problem that damages the whole domain, not just the thin pages. We've written about where programmatic SEO crosses the line if you want the detail.
What does work is narrower, more specific and much less impressive on a slide.
- Locality pages only for areas you actually deliver to, with the real delivery time and the honest radius. Three real ones outperform forty templated ones, and they don't put the domain at risk.
- Dish pages that answer a question. What's actually in a galauti kebab. How hot the Andhra chicken really is. Whether the biryani portion feeds one person or two. These pick up long-tail search and they're the kind of specific answer that AI assistants now quote.
- Diet and occasion pages —
jain food delivery in <area>,office lunch catering <area>,party order for 25 people. Bulk and corporate orders are the best economics in the direct channel: high ticket, planned ahead, and a query the aggregators serve badly. - An HTML menu with prices in text, per brand, per city. Not a PDF, not an image.
- A page for the thing you're actually known for. One dish, done properly, with photos and a story that's true, beats twenty cuisine pages nobody asked for.
What a direct order is actually worth
Direct is not free. It's a different cost structure, and the case for it survives being examined honestly — which is more than can be said for how it's usually pitched.
Aggregator commission in India commonly falls somewhere in the high teens to around thirty percent of order value, varying by contract, city, brand size and which visibility programmes you've bought into. We give a range rather than a number because the real figure is in your agreement and nowhere else. Take the table below, replace every figure with your own, and see whether it still holds.
| Line | Aggregator order | Direct order on your own site |
|---|---|---|
| Commission | Roughly ₹110–₹180, depending on your contract | ₹0 |
| Payment gateway | Handled inside the platform's cut | About 2%, so roughly ₹12 |
| Delivery | Included in the commission | ₹40–₹80 to a third-party fleet, or your own rider's cost |
| Discount funding | Often shared, in practice often mostly yours | Entirely yours — but you decide when and to whom |
| Visibility spend | Bid continuously to stay visible in-app | Content and SEO, paid once, compounding |
| Customer data | The platform's | Yours: phone number, address, order history |
The ordering site: speed, UPI and the second order
Ranking a delivery brand and then handing the customer a slow checkout is an expensive way to prove a point. The ordering site is the only page on your domain that has to be genuinely fast, and it's usually the heaviest one you'll ship.
- Delivery-radius check before the menu, not after the cart. Nothing burns a first-time customer like building a ₹700 order and being told you don't deliver to their block.
- Compress the food images properly. A menu is the heaviest page in the business. On a mid-range Android on 4G, that weight is the difference between an order and a back button.
- UPI intent checkout — one tap into GPay, PhonePe or Paytm. Card forms lose orders in India and always have.
- Guest checkout. Offer the account after the order is confirmed, never before it.
- Live order status on WhatsApp. It's the cheapest retention mechanism you own and it puts your brand back in their chat list, which is where the next order starts.
- A reorder link in the confirmation message. One tap to repeat the last order, at zero commission, is the entire point of everything above.
What it costs, and what we actually guarantee
One brand, one city, one ordering site sits at the ₹40,000 a month end of our SEO pricing. Multiple virtual brands with separate sites, separate menus and separate profiles is ₹75,000 and up, because it's genuinely several businesses wearing one kitchen. If your brand is also running a physical outlet, SEO for restaurants covers the map-pack side of it.
We don't promise a ranking position — nobody controls Google's index, and in a category this dependent on aggregator authority that promise would be even emptier than usual. What we do is freeze your trailing-90-day count of direct orders and enquiries from organic search on day one. If we haven't beaten that number in 90 days, we keep working free until we do. Three clients a month is what makes that a real risk we can carry.