Snapp Market · Order Updates Case Study
Navid Semi · navidsemi.com · 01 / 19
Title & Overview

Reducing order cancellations with a post‑checkout redesign

An evidence-led redesign for Snapp Market — Iran's #1 online grocery platform (64% market share, 20M+ users) with 5–10 minute vendor prep times. From support tickets to a shipped, A/B-tested feature.
Role
Freelance Product Designer · Remote, UK
Timeline
2w Research → 2w UX/UI → 4w Handoff & A/B
Team
1 PM · 2 Engineers · 1 Data Analyst
−38% cancellations · −58% support inquiries · 4-week A/B test
→ arrow keys / swipe
Snapp Market · Order Updates
Navid Semi · navidsemi.com · 02 / 19
My Role · End-to-End Ownership

From a pile of support tickets to a shipped, measured result

01 · Identify
Questioned the reported 15–20% cancellation rate and initiated a review of the most repetitive support tickets.
02 · Discover
Validated ticket patterns with the Data Analyst, analysed the customer journey, ran a 100-person survey and 4 vendor interviews.
03 · Define
Root cause: customers forget items, and the order locks at checkout — leaving only bad choices.
04 · Align
Challenged the business team's assumptions with evidence and won agreement for a 15-minute, add-only window.
05 · Design
Designed the flow, screens and components — and every edge-case state, not just the happy path.
06 · Deliver
Worked closely with engineering to resolve implementation challenges and ship the feature.
07 · Measure
A 4-week, 50/50 A/B test on ~500,000 live orders.
08 · Impact
−38% cancellations · −58% support inquiries · +22% retention · +15% AOV
Team · 1 PM · 2 Engineers · 1 Data Analyst · Freelance Product Designer, remote from the UK
Snapp Market · Order Updates
Navid Semi · navidsemi.com · 03 / 19
The Problem

No good options — a problem for users & business

15–20%
Order cancellation rate
(Validated with Data Analyst)
When users forget items after checkout, three bad choices — the post-checkout journey as it shipped, rebuilt in Figma:
Old Orders screen: order locked after checkout, only Track order and Cancel order, note says orders can't be changed after checkout
00 · Starting pointOrder locks at checkout. Only Track or Cancel.
Checkout for a second order containing a single yogurt, with the delivery fee charged again
01 · Place a second orderPays double delivery fees.
Cancel this order bottom sheet: refund takes 3 to 5 business days and the delivery slot is released
02 · Cancel & re-orderLoses time, risks the delivery slot.
Rate your order screen with two stars, Missing items and Couldn't change my order selected
03 · Give up on the itemFrustration — erodes trust.
The real question — what customer problem was behind the cancellations, and how do we fix the cause, not the symptom?
Snapp Market · Order Updates
Navid Semi · navidsemi.com · 04 / 19
Discovery

Looking past the metric to the root cause

Before designing anything, I traced the cancellations back to what customers were struggling with — rather than accepting the existing assumptions.
01
Support tickets
Asked to review the most repetitive customer support tickets: what were customers struggling with again and again?
02
Validated with data
Reviewed the ticket patterns with the Data Analyst to confirm they were a real trend, not a few loud complaints.
03
Journey analysis
Mapped the post-checkout journey to see where the experience broke down — and how that led to cancellations.
04
Root cause
Customers forget items, and the order locks at checkout. Every fix they had cost money, time or trust.
Then I sized it — a survey of 100 users and on-site interviews with 4 vendors, to measure how common the problem was and what operations could support.
Snapp Market · Order Updates
Navid Semi · navidsemi.com · 05 / 19
Evidence · Author's own user research

Research & Key Insights

N = 100 · Survey of Snapp Market Users
Ever forgotten an item?
81%
19% have not. A key pain point — 4 in 5 users have done it.
What do you do about it?
Place a new order36%
Do nothing33%
Cancel & re-order16%
Buy locally11%
Other4%
Shows users are forced into bad choices
How often do you shop online?
41%
Several times a month · 32% once a month · 22% weekly. A habit for most.
Top concern shopping online?
31%
Accurate fulfilment · 30% timely delivery · 23% shipping costs · 13% service charges.
The Core Insight
4 in 5 users forget items, but the app gives them no way to fix it. This forces them to pay extra fees, waste time, or give up on the order.
Methods: Survey run on Porsline — chosen for its ease of use and reliable data collection · 100 valid participants · Internal Data Analysis · AI Text Analysis (Claude)
Snapp Market · Operational reality
Navid Semi · navidsemi.com · 06 / 19
Four vendor storefronts visited for fulfillment-timing research, Tehran
4 vendors,
interviewed on-site
Interviewed 4 vendor storefronts in Tehran on real fulfilment timing. Prep time is the hard constraint on any edit window.
Prep time
5–10min

Observed vendor prep window

Sample
4

Storefronts interviewed on-site, Tehran

Today
Yes

Vendors already add items manually via cash

Snapp Market · Order Updates
Navid Semi · navidsemi.com · 07 / 19
Stakeholder Alignment

How I Changed the Business Team's Mind

01 Why they pushed back
Their initial thought
The business team rejected the idea for three reasons:
  • Extra Revenue. "Duplicate orders bring extra delivery fees."
  • Too Complex. "Editing will break our fast logistics."
  • Low Priority. "Not urgent enough to fix right now."
02 My argument & solution
A safe solution
I pushed back with evidence — support-ticket patterns, survey data and journey analysis — and a safe solution: a 15-minute, "add-only" window.
  • Wasted Slots. 1-item orders waste trips meant for big carts.
  • Angry Users. Paying twice frustrates users and erodes trust.
  • Safe for Logistics. Adding items is simple. The 15-minute limit fits the packing time and doesn't break our workflow.
The result — they accepted my offer. Because the "add-only" rule avoided complex refunds, we moved forward.
Snapp Market · Order Updates
Navid Semi · navidsemi.com · 08 / 19
Solution Strategy

A 15-minute, add-only editing window

0 MINOrder placed
5 MINVendor prep begins
10 MINPayment settled
15 MINHard cutoff · picking starts
20 MINOut for delivery
Why 15 minutes?
Matches the 5–10 min vendor prep window plus a payment-processing buffer — a hard cutoff before the "picking started" trigger fires.
Why add-only?
Adding is operationally simple; removing means inventory reversal and repacking.
Redesigned Orders screen: countdown ring at 08:26, 'Time left to add items' label, Track order and Add item buttons
Redesigned entry point
Snapp Market · Order Updates
Navid Semi · navidsemi.com · 09 / 19
Interaction Design

Anatomy of the add-item window

Five decisions on one screen — each one there to keep a rushed customer calm and clear.
5
Window closes before picking
The truck step lights up only after the timer ends — matching vendor prep time.
2
“Time left to add items”
Ties the timer to the action — never mistaken for delivery time.
4
One-line rule + Help
States the 15-minute rule; Help explains why removal isn't possible yet.
Orders screen with numbered annotations on the progress tracker, label, countdown ring, Add item button and note 5 2 1 3 4
1
Countdown ring
Time left at a glance. A calm, depleting ring — no “1 minute left!” alarm.
3
Add item, not Edit
Add-only by design: one clear action, outlined so Track order stays primary.
Snapp Market · Order Updates
Navid Semi · navidsemi.com · 10 / 19
Edge Cases & Complex States

Handling Edge Cases

I planned for common problems — and designed the state for each one:
Out of stock: item faded, red 'Out of stock' label, grey 'Not available' button
01 · Out of stock?Button shows “Not available” and turns off.
Price change: new price 52,300 Toman with old price 48,100 struck through and a note
02 · Price changed?New price shows up with a small message.
Time's up: ring at 00:00, button changed to Order again, note says time to add items has ended
03 · After 15 min?Ring empties and the button changes to “Order again”.
Payment failed card under the order: original order still on its way, Try again button
04 · Payment fails?Original order still goes through. User can try again.
Vendor app: packing locked with a countdown until the customer's 15-minute window ends; Start packing disabled
05 · Vendor packs early?Packing is locked until the 15-minute timer ends.
My approach — high-impact, easy problems first; hard ones (like removing items) saved for V2. I worked through every state with engineering to get it shipped.
Snapp Market · Order Updates
Navid Semi · navidsemi.com · 11 / 19
User Flow

One timer, two apps, every path mapped

Before drawing screens I mapped every branch — the happy path, the four customer edge cases, and the vendor lane that keeps logistics safe.
Customer app
Order placed
Timer starts15-min add-item window opens
DecisionForgot an item?
Home or OrdersTap Add item
CheckIn stock, same price?
No new delivery feePay for added item only
Same courierAdded to the same order
03 · No time leftRing hits 00:00 → “Order again”
01 · Out of stock“Not available”, button off
02 · Price changedNew price + small note
04 · Payment failsOriginal order kept → Try again
Vendor app
Vendor receives order
05 · Vendor packs earlyPacking locked — countdown shown until the customer's window ends
At 00:00Packing starts
Out for delivery
Window / edge case solved Failure, order protected Success Vendor lane
Snapp Market · Order Updates
Navid Semi · navidsemi.com · 12 / 19
Before → After

Three bad choices, replaced by one tap

01 · Place a second order
BeforeBefore: second-order checkout with a second delivery fee
→
AfterAfter: pay only for the added item, added to the existing order
Before72,100 Toman, including a second delivery feeAfter48,100 Toman — joins the same delivery
02 · Cancel & re-order
BeforeBefore: cancel the whole order to change it
→
AfterAfter: Add item within the 15-minute window
BeforeCancel everything, wait 3–5 days for a refundAfterAdd item within 15 minutes — same order, same slot
03 · Give up on the item
BeforeBefore: a two-star review, couldn't change my order
→
AfterAfter: the active order card with Add item right on Home
BeforeItem lost, trust eroded, low ratingAfterAdd item is one tap away on Home, where users land
Snapp Market · Design System
Navid Semi · navidsemi.com · 13 / 19
30+ components: order cards, re-order card, delivery and cart summary, store header, delivery selector, product cards, minimum-order progress and payment bar
Extending the
design system
Three new components designed, one existing component refined — to keep UI consistency and speed up developer handoff.
New
3

Countdown timer · sticky Active Order card · Add-Item bottom sheet

Documented
30+

Shipped components, every state in Figma

Per component
4

States handed off: default, loading, disabled, error

Snapp Market · Order Updates
Navid Semi · navidsemi.com · 14 / 19
Usability Testing & Iteration

The Pivot: 4x Fewer Misclicks

Version 1 put the add-item window only in the Orders tab. In moderated sessions (n = 10, task: find & edit an order), 1 in 4 users tapped the wrong place first.
So I moved the active order card onto Home — the screen people land on after checkout.
100%Task success
89%CSAT, up from 75%
67.4sAvg. time · 40% faster
Version 1Version 1: add-item window only on the Orders screen
Orders tab only24.1%misclicks · 1 in 4 users
→
Version 2Version 2: active order card with Add item moved onto the Home screen
Moved to Home6.3%misclicks · 1 in 16 users
Snapp Market · Order Updates
Navid Semi · navidsemi.com · 15 / 19
AI suggestedMy decision
Show a “Recommended items” popup in the add-item window to raise order value.
→
RejectedThe user is rushing to fix a mistake. A popup adds friction and risks the 15-minute window.
Send a push at minute 14: “Only 1 minute left to edit!” to maximise usage.
→
RemovedA sudden alarm causes panic, not action. Good UX should be calm.
Analyse the 100 survey responses automatically.
→
AcceptedClaude grouped the answers — roughly 2–3 days of manual work saved.
AI optimises for clicks and order value. I optimise for real context, emotions and the user's actual goal.
Snapp Market · Order Updates
Navid Semi · navidsemi.com · 16 / 19
A/B Testing Results & the Trade-Offs

Four weeks, ~500,000 live orders — the wins, and the reality check

−38%Order cancellations — relative drop vs. control
−58%Support inquiries — net aggregate drop over 4 weeks
+22%Customer retention — 30-day repeat purchase rate
+15%Average order value — users add, not abandon
Reality check — the "add-only" constraint frustrated users who wanted to remove items, causing a temporary spike in "how do I remove" support queries in Week 1.
The fix — an in-context tooltip explaining why removals weren't available yet. That query normalised by Week 2, so the net −58% aggregate support drop held steady.
Snapp Market · Order Updates
ROI & Business Impact

One feature.
Four teams
better off.

Every user metric from the A/B test maps to a business lever someone owns.

GrowthSupportFinanceOperations
Navid Semi · navidsemi.com · 17 / 19
User metricBusiness valueWho gains
−38%→
Revenue kept
Fewer cancelled baskets — the whole order stays, not just the forgotten item.
Growth
−58%→
Lower support cost
The most repetitive, low-complexity requests disappeared from the queue.
Support
+15%→
Revenue captured
Forgotten items added to the basket instead of being abandoned.
Finance
+22%→
Higher lifetime value
More customers came back to order again within 30 days.
Growth
Delivery slots freed up
No more one-item duplicate orders taking up couriers — delivery slots stay free for the big baskets that need them.
Operations
Vendors back in control
No more adding forgotten items by hand for cash — every addition is paid in the app, and packing waits for the timer.
Vendors
All metrics from the 4-week, 50/50 A/B test on ~500,000 live orders.
Snapp Market · Order Updates
Navid Semi · navidsemi.com · 18 / 19
What's Next & Learnings

Phase 2, and what I'd carry forward

Live for ~3 months, until Iran's internet blackout cut the long-term review short.
Phase 2 roadmap
Next
Item removal & auto-refunds
The hard case saved from V1 — users asked for it in week 1.
Then
A 30-minute window for Snapp Pro
Turn flexibility into a premium membership perk.
Later
Smart suggestions
“Forgot your weekly milk?” — remind people before the window closes.
What I'd carry forward
01
Constraints drive focus
The 5–10 min prep window forced a simple, high-impact solution.
02
AI is an assistant
It optimises for metrics; users need context.
03
Strategic rejection
Knowing when to say no to AI saves weeks of wasted effort.
04
What I'd do differently
Involve a backend engineer earlier to map payment-gateway limits.
19 / 19
Closing
Order Updates · Snapp Market

Asking why turned a cancellation number into a fix customers use.

From support tickets to a shipped, 15-minute add-only window — designed with the business, vendors and engineering.
Navid Semi · Product Designer
navidsemi.com
Takeaways
03
Home with the active order and countdown
Orders screen with Add item
Pay for the added item only
Window closed, Order again
The redesigned post-checkout flowHome → Add item → Pay → Window closes
01

Own it end to end

From the support-ticket review to design, delivery and measurement.

02

Lead with evidence

Tickets, data and journey analysis changed the business team’s mind.

03

Ship it with engineering

Every edge case worked through with engineers before launch.

→ End · Navid Semi · navidsemi.com