← All works
Product Design Case Study5 min read 8 sections

Billjot: A POS That Fits in Your Hand

Designing and shipping a billing system for small Indian restaurants still running on old billing machines and notebooks: researched across 168 restaurants, redesigned for phones after the first test, and deployed in 8.

Product
Point-of-sale and billing system
Role
Research, design, development, and testing; solo on the first version
For
Family restaurants up to 3-star hotels
Status
Deployed in 8 restaurants, December 2024
  • 168

    Restaurants surveyed across five towns

  • 24

    In-depth restaurant interviews

  • 8

    Restaurants running it

  • ₹6,200

    Per unit, hardware included

Billjot: A POS That Fits in Your Hand
0108

Overview

Billjot is a point-of-sale system for small restaurants: the kind of busy, locally famous family place that still bills on an old machine and keeps tabs in a notebook. The goals were short: build a POS that fits in your hand, make it easy for anyone on staff, and let each restaurant shape it to how it works.

I did it alone, as a freelance project in the final year of my B.Tech: the research, the design, the development, and the testing in real restaurants. It started life as Own Your Bill; today it's Billjot, billing and point of sale for restaurants and shops across India.

Version 2, designed at a bakery owner's request: menu on the left, the bill on the right
Version 2, designed at a bakery owner's request: menu on the left, the bill on the right
0208

Phase 1: The Story

It started with an observation. A restaurant in the lunch rush: staff juggling orders, the kitchen shouting updates, and a cashier stuck with a frozen billing system. The owner muttered that he only wished it worked as smoothly as it should.

That scene wasn't unique, so I asked owners three questions: what are you using, where does it hurt, and why are you still using it?

What, where, and why: an old billing machine, three pain points, and three reasons nobody switches
What, where, and why: an old billing machine, three pain points, and three reasons nobody switches
QuestionWhat owners said
What do you use?An old-school billing machine, plus a notebook for tabs
Where is the problem?Issuing tokens, tallying wages and expenses, and counting orders: paper gets lost, and lost paper means stealing
Why not switch?Cost: modern systems run ₹30,000 or more. Fear: complex systems scare staff who worry about pressing the wrong button. Inertia: twenty years of doing it one way
The first round of questions
0308

Phase 2: The Sherlock Holmes Part

One restaurant isn't a market. I classified restaurant types the textbook way, about twenty of them from dhabas to tiffin centres, then grouped them by how they actually take orders and bill.

Major typeIncludesHow they bill
Traditional and regionalDhabas, ethnic restaurants, sweet shops, halwai shopsRegional food, often one bill at the end
Modern and casual diningFine dining, casual dining, multi-cuisine, buffets, pubs and barsTable service, many rounds per table
Quick and on-the-goQuick-service restaurants, food trucks, street food, tiffin servicesFast, high volume, often takeaway
Speciality and cafésCafés, bakeries, seafood places, canteensCounter orders, packaged goods
Four groups, from about twenty restaurant types

Then came the hard questions, asked the same way of each group: system reliability, ease of use, inventory, payments, role-based access, and custom features. Quick surveys covered 168 restaurants across Vijayawada, Kakinada, Rajahmundry, Tuni, and Yanam, and in-depth interviews covered two dozen restaurants over 18 days.

0408

Constraints and the Market

The big players were already there: Petpooja, RanceLab, Posiflex, Vyapar, and Gofrugal. The target was everyone they priced out: basic, locally famous family restaurants up to 3-star hotels.

I wrote down my own limits as plainly as the market's, because they shaped every decision after.

ConstraintWhat it meant
MoneyNo budget for travel, more research, or prototype hardware
PeopleThe first version was solo: research, design, build, and test
TimeFinal year of a B.Tech, so academics came first and travel was limited
MarketGoing against established products, so the audience had to be hooked from day one
The ground reality, as I wrote it down
0508

The First Design, and What Testing Changed

The first design was a desktop screen sketched on paper: table numbers down the side, items in a grid, the bill and a print button on the right.

The first sketch: table numbers, an item grid, and billing on the right, with an early phone layout below
The first sketch: table numbers, an item grid, and billing on the right, with an early phone layout below

I released it and tested it in restaurants. The feedback was clear about where it worked and where it didn't.

FeedbackWhat I did
Good for traditional restaurants that bill at the endKept that flow exactly as it was
Desktop version is goodLeft it alone
Bad on mobileMade the phone view the focus of the next round
They wanted something like SwiggyDesigned the phone flow around browsing and tapping items
First-round feedback and the new plan
0608

Key Decision: Design for the Waiter's Phone

Restaurants with table service don't bill at a counter; they take orders on the floor. So the phone flow became the main design: the waiter picks a table, adds items, and sends a temporary bill that becomes the kitchen order ticket. The counter sees every table's bill live, and prints when the table leaves.

The mobile flow: table, items, send bill, kitchen order ticket, live bills
The mobile flow: table, items, send bill, kitchen order ticket, live bills

The desktop layout was then rebuilt as an annotated wireframe for a takeaway restaurant with one user: logo and navigation on the left, categories and search on top, order cards in the middle, and a billing panel on the right. Current orders sit as cards along the bottom that you swipe through.

The annotated desktop wireframe for a single-user takeaway counter
The annotated desktop wireframe for a single-user takeaway counter

A bakery owner then asked for a version of their own, and that became Version 2: categories as tabs, product photos, and the order panel always visible. The process followed a triple-diamond, user-centred approach: research, then experiment, then deliver, with a review loop at the end.

0708

Phase 4: Deployment

By 29 December 2024 the system was running in 8 restaurants, the ones that had supported me from the start. Each deployment was software and hardware, at a making cost of ₹6,200 per unit, built on Next.js as a PWA, Node.js, and MongoDB, served from a Raspberry Pi in the restaurant.

I tracked which features each restaurant actually used. That turned out to be the most useful design input of the whole project.

Feature use across the 8 restaurants: the core is universal, the extras are not
Feature use across the 8 restaurants: the core is universal, the extras are not

The core (offline mode, printing, inventory, analytics, sync, and access control) was used everywhere. The extras split the room: barcode scanning only mattered to places selling packaged goods, and tabs only to places with regulars. Nobody used the payment gateway, because every restaurant already had its own UPI; the virtual wallet and custom website went unused too.

0808

What I Learned

LessonWhere it came from
Fear is a design problem, not a training problemStaff afraid of pressing the wrong button kept old machines running for twenty years
Test early, even on paperThe first release showed the phone view mattered more than the desktop
Customisation means turning things offMost extras were used by three restaurants or fewer
Don't rebuild what people already haveEvery restaurant had UPI, so the payment gateway went unused
Lessons from research to deployment

The full research board, with the interview tables and sketches, is on FigJam.