Swiggy
Medium
The user never opens the Swiggy app. How do you show offers, Swiggy One benefits and ads?
This is a distribution question in a design costume. If the user never opens the app, the app is no longer your surface. You have to find where their attention already is, including the AI assistants that may place the order for them.
Step 1. Clarifying Questions
- Why don't they open the app? Two very different users:
a) Dormant: they don't order at all.
b) Off-app: they still order, through an AI assistant, WhatsApp, or a reorder widget.
- Assume b. The user orders through an AI assistant or chat, and Swiggy is the backend.
- What do we want them to see? Offers (to convert), Swiggy One benefits (to retain), ads (restaurant revenue).
Step 2. Why it matters
Offers, membership and ads all assume a screen Swiggy controls. If an assistant picks the restaurant, Swiggy loses its storefront. Restaurant ad revenue and membership value both depend on getting this right.
Step 3. First principles: where does the user still pay attention?
1. The request: "order my usual biryani" in an assistant or chat.
2. The confirmation: price, ETA, savings.
3. The wait: tracking on the lock screen, the most-watched 30 minutes of the order.
4. The doorstep: the rider and the packet.
Step 4. Solutions per moment
1. Request: give assistants structured data, not just a menu. Offers, delivery fee, member price and ETA in a machine-readable form, so the agent can choose Swiggy and explain why.
2. Confirmation: "You saved ₹62 with Swiggy One on this order" in the reply the user reads. Non-members see "Members would have saved ₹62".
3. Wait: a live lock-screen tracker. One line for a relevant offer, never more.
4. Doorstep: a QR on the bag for the next offer or a Dineout deal nearby.
5. Monthly WhatsApp savings summary for members: what you saved, what you missed.
Ads in an agent world:
- Restaurants pay to be a "sponsored suggestion", always labelled.
- Charge per order, not per impression. There is no scroll, so impressions mean little.
- Never let paid placement override the user's explicit request. If the user asks for a specific restaurant, they get it.
Cross question: precision or recall for showing offers off-app? Precision. One wrong or irrelevant offer in a chat reply feels like spam and the user turns it off.
Prioritise: 1 and 2 first. Without structured data the agent cannot pick Swiggy at all. Then 3 and the ads model.
Step 5. Metrics
- North Star: off-app orders per month
- Offer redemption rate off-app, member savings seen per order
- Sponsored suggestion acceptance rate and restaurant ROI
- Guardrails: opt-outs from messages, complaints that "the assistant pushed a restaurant I didn't ask for"
Interview tip: say who owns the user's attention now. Then design for that surface instead of trying to drag the user back to the app.
Community Answers (0)
No answers yet. Be the first to share your approach!