Home › JustVend

Reimagining JustVend

Feb 2026
System Design UX Research

Students eat from these machines when the mess is closed, when class is about to start, or when the café is too far to walk. We treated the machines as part of that food system.

Food can sit in the machine and still be out of reach. You cannot see the stock from your room, payment needs a network that drops, and the flap is heavy.
Context
Campus Infrastructure
Team
Arnav, Jashith, Rajshekhar
Role
Flap and the stock map
Date
Feb 2026

Context and people

We interviewed people and watched the machines at the hours they get used. Morning rush, the gap between classes, late night. Students go there when the mess or the cafés are shut, or when they do not have time to walk.

Three things break the trip. You cannot tell which machine still has what you want. Payment and the map both fail when campus Wi-Fi fails. The security flap is heavy, and getting the item out is a wrestle if your hands are tired or hurt.

Method

Interviews were short and next to the flap. We also talked to people who never buy. Security sits beside machines they are not allowed to use.

What we could change was small. Campus Wi-Fi stays bad, the machines stay where they are, and nobody was funding a new mess. The job was to make the machines that already exist easier to read and use.

ConstraintInventory is real only if someone restocked. A map that lies is worse than no map.
ConstraintPayments drop when the network drops. Offline has to be a first path, not a toast.
ConstraintThe flap is a physical object. Tired or injured hands were part of the brief, not an accessibility add-on.

Framing question

How might we get students to the food that's already on campus, when they can't see the stock, the network drops, and the flap fights back?

Design moves

I owned two of these: the stock map, which is a search across machines, and the outward flap. Offline pay, and the rest of the system, belong to the team: Arnav, Jashith, and me.

Four vending machines in a row. Arcs between them are labeled 200m, 300m, and 200m.
Four machines on campus, 200m, 300m, and 200m apart.

Networked discovery

Interviews showed a buying habit. People do not browse all thirty-odd items. They look for one thing, one juice, or one flavour of protein bar, and they only buy when it is in stock. The map is a search for that item across machines.

The habit was the evidence. A menu of everything in the machine would have answered a complaint. A search answers the way people already buy.

Pay when the network is down

If Wi-Fi drops, the phone pays from an offline wallet with a balance loaded earlier. A tap confirms you are at this machine. If GPS drops, you can say which machine you are at.

Payment that needs a perfect network fails at the moment someone is hungry. The offline path has to work on the first try, and it has to be this machine, not the one two buildings over.

An outward-opening flap

I struggled with the heavy flap myself, and interviews raised it too. A Coke machine already opens outward. Side by side with the campus machine, that flap was easier. The redesign opens outward so a tired or injured hand can get the item. A spend total on the phone shows the sum. It does not nudge, and it does not turn spending into a game.

The useful comparison was my own hands on two flaps. The campus flap fought back. The Coke flap did not.
Two vending flaps. The inward flap is marked hard and heavy, with an arrow pushing in. The outward flap is marked lighter, with an arrow lifting up.
The inward flap is heavy. The outward one lifts.

One night

11:30 PM. Mess closed, project half-done. You do not walk the hostel loop hoping a machine still has something edible. You search "Protein Bars." The map names the machine that still has them.

You get there. Campus Wi-Fi is dead. The offline wallet is already on the phone, so you tap. A check asks you to confirm you are at this machine, not the one two buildings over. That was the failure we kept seeing when GPS and payment both guessed. Then you lift an outward flap, phone still in the other hand.

That is the whole job: know what is in the box before you walk, pay when the network is gone, take the food without a wrestling match.

What testing changed

The class ended with a pitch of leverage points, so these designs were not tested with users. What changed the design was the research. The one-item habit moved discovery from a menu to a search. The Coke machine comparison backed the outward flap.

Next: test search-first with people who already buy the same item, and test an outward-flap mockup with tired or injured hands.

Role

The team was Arnav, Jashith, and me. I owned the flap redesign and networked discovery, the stock map. Offline pay and the rest of the work are the team's.

Questions still open

A wider brief, one that includes security and other people who work on campus, still has these open.

Students were unhappy even on discounted items. They said they were falling for it, and they called the offers marketing tactics. A lower price did not fix a trip that felt like a trick.
Students who buy the items marked healthy are still unhappy. The packet says wellness. The machine is still mostly junk food. It is not clear that anyone is eating better.
Security guards and campus workers sit beside the machines and still cannot use them. A service that leaves out the people standing next to it is unfinished.

Learning

The best leverage came from habits, rather than from complaints. People search for one item, so the fix was a search across machines. Some of the most useful evidence was my own hands on two different flaps.

TabSplit →