Shelf Life
A redistribution layer that moves near-expiry medicine from pharmacy shelves to patients who can't afford it — before it becomes waste.
Inspiration
I was standing in a small neighbourhood pharmacy when the owner pulled a tray of boxes off the shelf and dropped it into a bin behind the counter. Antibiotics, a blood-pressure strip, two inhalers. All sealed. All unexpired — with about six weeks left.
He explained it plainly: nobody buys a box with a short date on it. So at a certain point he stops trying, writes it off, and disposes of it. Meanwhile, three streets away, a community clinic rations the same molecules.
Two problems sitting next to each other, neither able to see the other. That gap is the whole project.
The problem
Independent pharmacies carry inventory risk alone. A box that doesn't sell is a pure loss, and the loss compounds: capital is destroyed, the medicine is destroyed, and disposal itself costs money and carries an environmental cost.
At the same time, patients at the bottom of the income curve don't fail to get medicine because it doesn't exist — they fail because the price at the counter is higher than what's in their pocket on that particular day.
Why existing approaches don't close the gap:
- Donation drives are episodic and run on goodwill. They collect whatever people happen to bring, not what clinics actually need, and they typically collect after expiry is already a problem.
- Discounting near-expiry stock is the obvious move, but pharmacists resist it — visible discounting trains customers to wait for markdowns and signals "old stock," damaging the shelf next to it.
- Manufacturer return-and-credit schemes exist but are slow, paperwork-heavy, and generally only reachable by chains with the administrative capacity to file claims.
- Pure charity models ask pharmacies to absorb a 100% loss and call it generosity. That is not a system. It is a favour, and favours don't scale.
Every existing option fails on the same axis: it treats this as a moral problem when it is a timing and routing problem.
Proposed solution
Shelf Life is not an app, a charity, or a discount scheme. It's a three-part operating system for near-expiry stock.
1. The Amber Band — a physical, visual expiry protocol
A single adhesive band applied to a box when it crosses a defined time threshold. Three states:
| Band | Remaining shelf life | Status |
|---|---|---|
| Green | > 90 days | Normal retail |
| Amber | 30–90 days | Eligible for redistribution |
| Red | < 30 days | Withdrawn from the network |
The band does something a spreadsheet can't: it makes time visible at the point of handling. A pharmacy assistant restocking a shelf becomes an inventory auditor without being trained as one. The Red threshold is deliberately conservative — the network never handles anything that could arrive at a patient with a date problem.
Critically, the band reframes the discount. It isn't "this pharmacy sells old stock." It's "this pharmacy participates in the Amber network." Same economics, inverted signal.
2. The Amber Ledger — matched routing, not collection
Participating pharmacies log Amber-banded items against a shared list of what partner clinics and community health workers actually need that month.
The direction of flow matters enormously. Donation drives push whatever is available. Shelf Life pulls against declared demand. A clinic that needs paediatric antibiotics receives paediatric antibiotics — not eleven boxes of antacid.
Anything Amber that no clinic has requested simply stays on the retail shelf and sells or doesn't. The network never takes custody of something it can't place.
3. The Third-Loss Model — why a pharmacy says yes
This is the mechanism the whole thing rests on.
Today, a pharmacy that fails to sell a near-expiry box loses 100% of its cost. Under Shelf Life, the loss is split three ways:
- The pharmacy absorbs one third as a write-down
- The patient or clinic pays one third — a real price, so the medicine has value and is not over-collected
- The community fund (local CSR sponsorship, municipal health budget, or a partner foundation) covers one third
Let \( C \) be the pharmacy's cost for a unit and \( p \) the probability it sells at full retail before the Red threshold. The expected loss today is:
$$ L_{\text{today}} = (1 - p) \cdot C $$
Under the Third-Loss Model, the unsold unit recovers two thirds of its cost:
$$ L_{\text{Shelf Life}} = (1 - p) \cdot \frac{C}{3} $$
The pharmacy's expected loss on at-risk inventory falls by roughly two thirds — and the improvement is largest exactly where the pain is worst, on slow-moving stock with low \( p \).
(These are illustrative model assumptions, not measured results. The split is a design parameter — the model holds at any ratio where all three parties are better off than under total write-off.)
That's the insight I kept coming back to: nobody has to be generous for this to work. The pharmacy is better off. The patient is better off. The funder buys measurable health outcomes at a third of the retail price. Generosity is a bonus, not a dependency.
What I learned
Waste problems are almost never storage problems. They're information problems. The medicine wasn't wasted because there was nowhere for it to go. It was wasted because the pharmacy and the clinic had no shared clock.
"Free" is a design flaw. My first version gave the medicine away. The more I thought it through, the worse it looked: free goods get over-requested, under-valued, and quietly resold. A small real price is what makes the demand signal honest.
A good system makes the right action the lazy action. I originally designed an elaborate audit workflow. Nobody in a busy pharmacy would ever run it. The sticker does 80% of that work for 2% of the effort, because it rides on something staff already do — touching the box.
Regulatory constraints are design inputs, not obstacles. Rules about pharmaceutical custody, licensed handling, and cold-chain integrity aren't the enemy of this idea — they're what forced it into a licensed-pharmacy-to-licensed-clinic shape, which is also the only shape that's actually safe.
How I built it
Entirely a design and systems process — no code.
- Problem framing. Mapped the near-expiry lifecycle as a timeline and marked every point where value could still be recovered but wasn't.
- Stakeholder mapping. Wrote out pharmacy owner, pharmacy assistant, clinic coordinator, patient, and funder as separate actors — then wrote down what each one loses under the current system and what each one must gain to participate. Any version where one actor only loses was discarded.
- The band system. Chose thresholds (90/30 days) by working backwards from worst-case dispensing lag rather than by picking round numbers. Designed for recognition under pressure: colour first, text second.
- Economic modelling. Built the Third-Loss split and stress-tested it at different ratios to find the range where all three parties stay positive.
- Flow design. Drew the pull-based routing loop and traced the failure cases — what happens when nothing is requested, when a box reaches Red in transit, when a clinic over-requests.
- Safety pass. Deliberately narrowed scope: no controlled substances, no cold-chain items, no opened packaging, licensed endpoints only.
Challenges I faced
Making it not-charity. This was the hard one. Every early draft collapsed into "pharmacies should donate." That version dies the first quarter a pharmacy has a bad month. Rebuilding it so the pharmacy's self-interest points the same direction as the social outcome took the most iterations by far.
Trust and liability. A near-expiry medicine that reaches a patient badly is far worse than one that reaches a bin. The Red threshold, the licensed-endpoints-only rule, and the exclusion of cold-chain and controlled items are all responses to this, and each one costs coverage. Accepting a smaller, safer scope was the right trade and not a comfortable one.
Resisting the app. My instinct was to design a platform with dashboards. But a system that requires every small pharmacy to adopt new software has already failed at pharmacy number four. Forcing the core mechanism down to a sticker and a shared list was a subtraction problem, and subtraction is harder than addition.
Avoiding market cannibalisation. If Amber pricing is too attractive, paying customers migrate to it and the pharmacy loses full-price sales. That's why eligibility is defined by patient need and clinic request, not by anyone who walks in and asks for a discount.
Implementation, users and outcomes
Users: independent pharmacies (supply side), community clinics and health workers (demand side), low-income patients (beneficiaries), local funders and municipal health bodies (the third share).
Resources required: printed band rolls, a shared monthly needs list, a local coordinator, and a signed protocol covering custody and eligibility. No new hardware. No new software.
Rollout: one district, 8–12 pharmacies, 2–3 clinics, 90 days. Measure four things — units diverted from disposal, rupee value recovered by pharmacies, patients served, and zero-incident compliance.
Expected outcomes: less medicine destroyed, less capital destroyed, more patients medicated, and a repeatable protocol that a second district can adopt by printing stickers rather than by installing anything.
Honest failure risk: if pharmacies don't band consistently, the network starves. The entire design bets on that one habit being cheap enough to stick.
Shelf Life doesn't create new medicine. It just stops us from throwing away the medicine we already made.
Built With
- business-model
- canva
- circular-economy
- design-thinking
- figma
- financial-modeling
- google-sheets
- healthcare
- logistics
- market-research
- miro
- packaging-design
- process-design
- public-health
- reverse-logistics
- service-design
- social-impact
- stakeholder-mapping
- supply-chain
- sustainability
- systems-design
- unit-economics
- user-research
- waste-reduction
Log in or sign up for Devpost to join the conversation.