Back of house · Swiggy Dineout

Building a unified partner system

One app where restaurants on Dineout handle bookings, the digital menu, bills and payouts.

Role
Product Designer
Team
PM, engineers, with restaurant partners, staff, city heads and account managers
Platform
App & web, for restaurants
Reservations: the day's bookings with offers and guest notes
Transactions: consolidated net sales and AOV above recent transactions
Payouts: the consolidated payout and past payouts

In short

Restaurants shouldn't have to ask where their money is. So we put it where they already look.

The problem, on both sides

For restaurants
Bookings and the digital menu were stuck on a desktop, away from the floor. Bills paid through Dineout weren't visible, and payouts came as emails. Servers had nothing to match their bills against.
For Swiggy
With nowhere to check, restaurants sent every money question to support or their account manager. Answering those in the app means less load on both.
Before thisEvery money question meant digging through emails.

One restaurant, three very different jobs

One screen for all three would suit none of them. So the app shows each role what it needs.

What I held on to

Money, always visible
Restaurants know what they earned and when it arrives.
Fast enough for service
Anything used on the floor reads in seconds.
Each role, its own view
Owners see totals. Staff see just what they need.
One system, everywhere
Same patterns on phone and web. Learn one, know both.

A table's day, in four tabs

The four tabs follow one table, from booking to payout.

Reservations · built from scratch, for the phone

There was no reservations experience on the phone at all. Bookings lived only in a wide desktop table: guest, time, party size, offer, phone, status, and more hidden in each row. I built the mobile version from scratch, bringing over everything that mattered on web, and restructured it, because that table doesn't shrink to fit a phone.

What comes first
Out of every column, a manager mid-service needs four things: time, party size, offer and any special request. Those lead each card; everything else is a tap away.
Grouped by meal
A flat list of the day becomes breakfast, lunch and dinner tabs, each with a guest count, because that's how the floor is planned.
Nothing missed
Requests like a birthday cake are highlighted on the card, not buried in an expanded row.
Bookings, by meal
Bookings, by mealGrouped by meal, with the offer, party size and special requests up front.
Find a booking fast
Find a booking fastSearch and filters, for finding one guest mid-service.
The same bookings, as a desk view
The same bookings, as a desk viewA filterable table, each booking opening into its full details.

Transactions · built from scratch

For the first time, restaurants could see every bill paid through Dineout. Servers scan thousands of these, so each row had to say as little as possible.

The one thing I didn't compromise onThe last four digits of the transaction ID, up front. That's what servers match against their own system.

M1 · launch lean

We shipped a small first version with what we already had, so restaurants got value quickly.

A list built for matching
A list built for matchingBill amount and the last four digits first, so a bill is found at a glance.
The full bill, one tap away
The full bill, one tap awayThe split opens in place, keeping the list fast.

After M1, partners asked for totals too, but only for some people on the team.

M2 · shaped by who's looking

So M2 splits the app by role. Restaurants choose who sees what from the dashboard.

For owners: the totals first
For owners: the totals firstNet sales and AOV against last week, above the list.
For staff: nothing but the list
For staff: nothing but the listScan-first, with failed and refunded payments standing out.
Filters that match how they ask
Filters that match how they askDate, outlet, status and offers, in one sheet.

What a row says

On web, a full ledger
On web, a full ledgerEvery charge per transaction, with one-click reports.

Payouts · from inbox to dashboard

Payouts used to arrive as emails. They come every week or two, so the app shows each cycle clearly, not just one amount.

Every rupee, explained
Every rupee, explainedThe latest payout and its UTR, broken down to the net amount.
Any past cycle, in full
Any past cycle, in fullEach past payout opens with its breakdown and reports beside it.

On the phone

The answer, at a glance
The answer, at a glanceThe period's payout on top, past cycles as cards.
Every cycle, one list
Every cycle, one listPayout history, out of the inbox.
The breakdown, sized for a phone
The breakdown, sized for a phoneThe full split on one scroll, with the UTR.

Many outlets, one owner

Many owners run several outlets, sometimes under one PAN, and want to see money both ways.

By outlet, or by PAN
By outlet, or by PANOne filter for a single branch or the whole business.
A lighter switcher, for five or fewer
A lighter switcher, for five or fewerPitched: outlets as chips, one tap to switch.

One system

Four tabs, two platforms, one set of building blocks. Learn it on web, and you already know it on the phone.

Key decisions

The last four digits come first
Servers match bills by them, so they stay up front.
Launch lean, then split by role
M1 shipped fast. M2 used partner feedback to give owners and staff different views.
From inbox to dashboard
Payouts belong where restaurants can check them anytime, not in email.
Context before the guest arrives
The offer, party size and special requests come first, so the team can prepare.
Less in every row
Rows show only what servers scan for. Full details are a tap away, and on web. Less in the row, a faster list.

Impact · since launch

Restaurants can now see every bill, follow every payout, and run bookings and their digital menu, all in one app. Transactions, Payouts and Reservations on the phone are all new, and the Digital Menu lives in one tab.

What I learned

Designing for people at work is different. A server mid-rush doesn't explore, they match and move on. The most useful thing in a row was four digits.

Shipping small and then listening beat designing everything up front. M2 is better because M1 reached partners first.

And when money is involved, clarity is the kindest thing you can design.

Thanks for reading (or probably just scrolling).