Conceptual case study · Product design
A recurring order for the things you run out of fastest, with a reminder the evening before so you stay in control.
*Independent concept. Not affiliated with or endorsed by Blinkit.
People who live alone or have just moved to a city have no neighbourhood supply chain. Milk, bread and eggs run out fast, and every refill is a trip to the shop or a fresh order with a delivery fee on a ₹25 item.
Let people set up a recurring order for what they buy most. The evening before each delivery, send a reminder with Deliver, Skip and Pause. If they do nothing, it arrives.
A 21-screen interactive prototype across four flows: set up, stay in control, handle exceptions, and let the system learn. Every decision traces back to something survey respondents said.
01 · Problem
I started from my own observation of people moving to a new city. Everyone had a workaround, and every workaround leaked.
None of these is a planning tool.
I wanted something in between: a recurring order that adapts to how people actually live.
02 · Research
A short online survey (June to July 2026) covering how often people buy essentials, what happens when they run out, whether they have used recurring delivery, and what would make them comfortable or uncomfortable with it. I clustered the open-ended answers into an affinity map.
Multiple choice, so percentages overlap.
02 · Research · Affinity map
Mostly get from shops if not so urgent. If urgent, then use Blinkit
Lives with family
I bought it the next day
Lives alone
InsightPeople are not forgetting to order. They are choosing not to plan. A recurring plan has to feel worth the effort.
There isn't a fixed duration of time when a particular thing runs out for us.
Lives with family
Recurring deliveries will create extra stock. And for perishable items it's not good.
Lives alone
InsightThe strongest objection is to a fixed schedule, not to the idea. Flexibility is load-bearing.
I'd like to review at least the items that are going to be delivered before money is debited from my account.
Lives with family
I would like to have control over what gets deliver to me rather than a recurring service.
Lives with family
InsightAutomation is acceptable only with an escape hatch. Skip and pause must be as easy as setup.
Autopay of literally ₹1000 in 2 days! We had only spent ₹400 on items on day 1.
Lives alone
Paying 50rs for a 25rs milk packet adds up.
Lives alone
InsightSurprise charges are the biggest trust-breaker. Show the amount before it is taken and never hide fees.
Online milk often gets curdled due to product being delivered near to expiration.
Lives with family
I don't trust quick commerce with things with low shelf life.
Lives with family
InsightMilk is the hero item and also the riskiest. Freshness has to be signalled, not assumed.
Sometimes I live at another location. Can't guarantee when I'll be home to receive the order.
Lives alone
I won't be able to add items I might need if it's pre-ordered.
Lives alone
InsightA delivery you cannot receive is worse than none. The reminder is essential, not optional.
02 · Research · Findings
I went in with three assumptions. The data pushed back on all of them.
I expectedThis is a problem for people living alone.
I found35 of 48 live with family, and they described the same frustrations.
I expectedPeople would welcome a recurring service.
I foundAbout 15 said their consumption is too variable for a fixed schedule.
I expectedFew had tried it because it is new.
I found2 of the 5 who had used one edited it by hand whenever their needs changed.
Only 13 respondents lived alone or with flatmates, so my target group is thin. I read the data as design constraints, not proof of demand. Routine became a flexible baseline with an easy exit, not a rigid subscription.
Sometimes the banana will get over before Wednesday, then I had to go to the app and set my next upcoming order from Thursday to Wednesday. And again I will change it to Thursday after the next delivery.
Survey respondent, lives alone, used a milk subscription
03 · Reframing
People did not reject automation. They rejected automation they could not see, adjust or stop. That turned the brief into one question and three principles.
How might we make a recurring order feel like a flexible baseline instead of a commitment?
Defaults do the work. Every default can be changed for one delivery or for good, in one tap.
In the product: pre-filled setup, one-tap editsDo not make people manage a subscription. Ask them the evening before, when the delivery is real.
In the product: the reminder with Skip and PauseNo surprise charges, items or timings. Show the amount, the items and the cutoff every time.
In the product: amount and 10 PM cutoff in every message04 · Solution
The key screens are below, in the order a customer meets them. Each screen sits beside the thinking behind it, with a line to the exact element. Choose Try this screen to use the real prototype.
Flow 1 · Discover and automate · Start
Home · the entry point
“Previously bought” is where the prompt gets its evidence.
“You buy milk often” sits above the nav and can be dismissed.
Routine never replaces the one-off order. Nobody is forced into planning.
Start a Routine · one pre-filled sheet
“Bought 6 times last month” explains the suggested schedule.
Repeat, time and payment each open their own edit. Defaults, never locks.
The button states the exact amount: ₹60 per delivery.
The stepper is on the sheet, so the most common edit costs nothing.
Free delivery and no minimum order remove the fee objection.
Flow 1 · Discover and automate · Schedule
Change schedule
Every day, specific days, weekly, every X days, monthly, each with an example.
“Delivers: Mon · Wed · Fri” restates the choice before Continue.
Mon, Wed, Fri is a tap on a chip, with a sensible default pre-selected.
Routine created
Item, schedule, time, address and payment as a final review.
“We'll message you the evening before each delivery” says what happens next.
A success state first, then the full summary so nothing is assumed.
Free delivery and no minimum order appear again at the moment of commitment.
Flow 2 · Stay in control · Manage
Routine detail
“Active” and the next date answer the only two questions people have.
There are no auto-replace or price rules to set. If something changes, the app asks you then.
Schedule, time and payment each open one focused edit. Nothing is buried.
Pause is one tap and opens vacation dates too. Cancel is clearly marked in red.
Flow 2 · Stay in control · The reminder
Day-before reminder · lock screen
“Reply by 10 PM to change, otherwise we deliver as scheduled.”
“Running low on atta?” is a one-time add with no extra delivery fee.
Deliver, Skip, Pause on the notification. No app launch needed to say no.
Total, free delivery and payment, plus Add or remove items.
Flow 2 · Stay in control · Edit one delivery
Tomorrow's delivery
“Edit until 10 PM” sits next to the time, so there is no guessing.
“Changes apply to tomorrow only. Your Routine stays the same.”
“No reply by 10 PM? We'll deliver as scheduled.”
The same three choices as the notification, as large cards.
Change quantity or remove an item for tomorrow only.
Add items for tomorrow
“Added once to tomorrow's delivery. Your regular Routine stays the same.”
The normal product grid, with a running total on Done adding.
Flow 2 · Stay in control · Pause
Pause and vacation mode
Time away is the scenario respondents described, so it leads.
1 week, 2 weeks, or until I resume for open-ended pauses.
Leaving and back-on dates define the pause. Invalid ranges are blocked.
“No deliveries 15 Oct - 22 Oct. Restarts automatically when you're back.”
Flow 3 · Handle uncertainty · Recovery
Out of stock
A clear banner states the problem and the two ways out.
The item is tagged “Out of stock” so nobody has to hunt for the cause.
Four alternatives with prices, each one tap away from a confirmation.
Search other milk, or skip it for tomorrow. Nothing is replaced without a tap.
Price change · pop-up
The item in tomorrow's delivery is tagged “Price up”, so the cause is obvious.
Keep at ₹36, or choose another milk.
₹30 to ₹36, and ₹12 more on this delivery, so the impact is concrete.
Skip milk for tomorrow, in plain text.
Payment failed
“Tomorrow's delivery is still on. Fix it to keep it.”
Fix it by 10 PM tonight, stated in one line.
A Failed tag sits next to the payment method that did not work.
Retry UPI or pay on delivery. A link changes the method.
Flow 4 · Blinkit learns · Adapt
Adaptive suggestion
“You skipped 3 of your last 5 milk deliveries.” The reason is their own behaviour.
Update is the main button. Keep current sits right below it.
The days you get now beside the suggested days, with deliveries per week.
The schedule only changes when the customer taps Update.
05 · Build and QA
I did the research, analysis and product decisions. I then used ChatGPT and Claude to build the prototype from my survey findings and Blinkit screenshots, so the Routine feature reads as native to the app.
Building was a loop: I tested every flow, found logic and flow errors, described them precisely, and had them fixed. The AI wrote code; the judgement about what was wrong came from using the product.
06 · Reflection
Surveys rarely say “build this”. They tell you where an idea will fail. The most useful data was the objections, and treating them as design constraints shaped the whole product.
I also learned to treat AI output as a draft. The product thinking came from research, and the quality came from testing and being specific about what was wrong.