Rebuilding SNIPES Wallet Onboarding
Accrue × SNIPES • 2025–2026
I joined Accrue in April 2025 and sat with 18 users in moderated usability sessions before touching the design. What they surfaced became the brief for a full onboarding rebuild.
Jump to solution
Product
SNIPES mobile app with Accrue embed SDK
Team
1 Designer, 2 PMs, 3 Engineers, 1 QA
Timeline
Apr 2025 – Mar 2026
Role
Research, Strategy, Design & Testing
THE CLIENT
Accrue × SNIPES
Accrue builds white-label co-branded wallet and loyalty platforms for enterprise retail brands. I led design for SNIPES, their flagship merchant, redesigning how users fund and manage their wallets.
I joined Accrue in April 2025. The SNIPES wallet was being readied for launch, but the onboarding flow had been built without direct user input. Before redesigning anything, I ran a moderated usability study with 18 users to find out what was actually working and where the flow was failing.
PROJECT GOALS
- 1Find the real friction inside onboarding through direct user research
- 2Rebuild the flow around what users said, not what we assumed
- 3Absorb a mid-development business pivot (barcode scan-and-pay) without losing the redesign
- 4Ship clearer copy, honest blocking states, and a shorter flow
THE RESEARCH
I sat with 18 users before I touched the design.
Moderated usability sessions, one-on-one. Think-aloud protocol. I watched where users paused, what they misread, and where they gave up. The 18 participants covered a range of sneaker shoppers across ages, tech comfort, and wallet familiarity.
Study setup: 18 participants, 30-45 min sessions, 1:1 moderated, think-aloud protocol
Method: moderated think-aloud
Each session ran 30-45 minutes. The friction surfaced itself.
- Participants installed the app and walked through onboarding live
- Narrated their thinking out loud while moving through screens
- No leading questions from me as the moderator
- Same task set across all 18 sessions for comparability
18 participants across heavy / casual / first-time mobile-wallet usage
Participants: a real range of users
Eighteen sneaker shoppers, mixed across the dimensions that mattered for this product.
- Ages roughly 18 to 45
- Mixed payment-app comfort: heavy mobile wallet users alongside first-timers
- Mixed SNIPES familiarity: loyal customers, casual shoppers, total newcomers
- Recruited so no single profile dominated the findings
Behavioral signals: pauses, misreads, repeated taps, confused questions
What I watched for
Specific behavioral signals that flag the design working against the user:
- Pauses longer than a few seconds
- Misreads of copy or interface labels
- Repeated taps on the same element
- Confused questions to me as the moderator
THE PROBLEMS
Six themes the research surfaced.
The 18 sessions produced a long list of friction. I grouped the findings into six themes that the redesign had to answer. Every redesign decision later traces back to one of these.
KYC bottom sheets: name + birthday, then address
"Verify your Identity" sounded like a government audit
We needed a name, birthday, and address. The word 'verify' set off alarms anyway.
- Felt like a credit check or legal requirement to several participants
- Multiple users physically flinched at the phrasing
- Lost users who would have happily shared this info in a less formal context
Old 'Add backup payment' form with payment details
Backup payment had no explanation users could grab
The screen said 'backup payment' without ever explaining backup for what.
- Participants couldn't articulate why they needed it
- Didn't know what would happen with their card
- Several skipped it because they didn't trust adding card details to something unclear
Old reward screen: 5% hero card, $1.00 figure in small body text
Percentages didn't land. Dollars did.
Fourteen of eighteen participants couldn't connect '5% back' to a real dollar amount. The actual $1.00 was on the screen, just small and easy to miss.
- 5% dominated the hierarchy as a huge orange hero card
- The $1.00 figure sat in body copy below, easy to skip
- Leading with the dollar amount in testing made comprehension immediate
Bottom-sheet cascade: reward → verify identity → address
Users couldn't tell what was mandatory
The onboarding looked optional in places where it wasn't. Sheets felt swipeable, the flow had no visible end, and KYC stalls left users without next steps.
- Sheets read as notifications participants could swipe away. Hierarchy failed to signal importance.
- Sheet after sheet, with no visible end or step count
- Some wanted to explore the app first: 'I just want to see it'
- Others felt blindsided by repeats: 'I already gave you my name, email, and phone'
- Denied or under-review users got no guidance on what to do next
Users could tell me 'this is confusing' instantly, but they couldn't tell me why. The research showed me what they paused on, what they misread, what they skipped. That's where the redesign had to start.
THE REDESIGN
I rebuilt onboarding around what users said.
Every redesign move maps to a problem from the research. Mid-development, the business added a constraint: an Apple Card provisioning flow for users who wanted SNIPES Reserve in their Apple Wallet. I folded it in without breaking the existing structure.
3-step stepper flow with progress indicator
3-step stepper, end visible from step 1
Sign Up → Complete Profile → Set Up Payment.
- Each step shows why it matters in its subtitle
- Total step count visible from the first screen
- No more open-ended onboarding that surprised users in the study
Copy comparison: old vs new across 4 screens
Copy rewritten for skimmers
Every screen had to work at a glance.
- 'Verify your Identity' → 'Set Up Profile'
- 'Backup payment' → 'We'll charge this card beyond your balance. You'll earn 3% back on that amount.'
- Dollar amounts replaced percentages wherever they fit the context
KYC: bottom sheet (before) vs full-screen stepper step (after)
KYC promoted to full screen with the stepper
KYC moved out of the bottom sheet and into a full-screen step inside the stepper.
- Hierarchy signaled importance without relying on copy
- Re-testing showed users treating it as a real step, not a notification
- Stepper progress kept users oriented through the heavier step
KYC blocking states triptych: blocked / under review / docs needed
Honest blocking states with a way forward
Three blocking states, each with a clear next step:
- Blocked: the dashboard explains why and what happens next
- Under review: a realistic time expectation without false promises
- Documents needed: clear instructions and a single CTA
STEP 1 OF 3
Select payment method
Apple Card provisioning flow: optional path off payment step
Apple Card provisioning, folded in mid-flight
The business added a requirement mid-development: SNIPES Reserve had to provision as an Apple Card.
- Designed as an optional path branching off the payment step
- Users who wanted it opted in with one tap
- Users who didn't never saw the additional friction
BETA + PIVOT
Beta surfaced gaps. The business shifted the model.
I soft-launched the redesign with 50+ SNIPES employees as beta testers. The in-store context exposed problems the office study had missed. Around the same time, the business shifted the checkout model from tap-to-pay to barcode scan-and-pay. The redesign absorbed both.
Beta findings summary: speed, payment surface visibility, card-linking friction
Beta with 50+ SNIPES employees
Employees used the wallet at the register, in the store, during real shifts. Their feedback exposed gaps the office testing had missed:
- Speed at the counter became a real concern, not just a theoretical one
- Clerks needed a visible payment surface they could verify on the customer's screen
- Card-linking was hard for users to complete during a busy shift
Scan-and-pay barcode screen with balance toggle
From tap-to-pay to barcode scan (in development)
Store associates couldn't see balances with tap-to-pay, which broke the customer-clerk interaction. Leadership shifted to barcode-based scan-and-pay.
- The barcode screen becomes the destination of onboarding
- 'Complete your profile to get your barcode' replaces 'verify your identity'
- Ships next sprint, alongside OCR scanning and the choice screen
Pick how to pay — Tap to explore
Choice screen: 'Load cash' vs 'Add a card' with value props
Load cash or add a card, the user picks
After onboarding, users hit an honest choice point with both options shown side by side:
- Load cash up front: pre-fund the wallet, earn 5% back
- Add a card: wallet charges automatically beyond balance, earn 3% back
- Skip is always available, no dark patterns
STEP 1 OF 3
Select payment method
OCR card scanning flow: viewfinder → auto-detect → confirm
OCR card scanning for speed at the register (in development)
Camera-based card entry via on-device SDK. Critical for the in-store context where speed matters most.
- Removes manual entry that slowed users on a busy shift
- Recommendation: BlinkCard or Apple Vision after reviewing 18 competitor implementations
- On-device processing keeps card data off the network
FINAL DESIGNS
The complete onboarding experience.
From sign-up to barcode. Every screen redesigned for clarity, speed, and trust, with a real user quote behind each decision. The stepper, copy, and blocking states are live; the barcode hero and OCR scanning ship next sprint.
Final designs: full-width hero composition of the complete onboarding flow
The full stepper flow: checklist overview → profile → payment → barcode
Barcode screen with balance toggle
Barcode as the destination (in dev), with Reserve balance toggle
OCR card scanning viewfinder
Camera-based card scanning (in dev)
Choice screen: load cash vs add a card
Honest path-picking with skip always available (in dev)
Mixpanel before/after funnel overlay across the redesigned flow
Every change traces back to a specific user quote from the study
THE IMPACT
55.9% → 45%
KYC drop-off
10 percentage points reclaimed at the step that lost more than half of new wallets.
41.2% → 65.86%
Profile completion
+24 points over the old flow. The stepper and rewritten copy did the work.
12 → 3
Onboarding steps
Compressed from a multi-screen checklist to a 3-step stepper. Reduced perceived effort before users start.
18K
Wallets / month
New wallets created in the last 30 days on the redesigned flow.
91–99%
Mid-funnel retention
Between steps inside the stepper. The remaining cliff is the funding moment, not navigation.
The shipped pieces of the redesign — the 3-step stepper, rewritten copy, and honest blocking states — moved every funnel number they touched. The barcode hero, OCR card scanning, and contextual card-linking nudge are designed, approved, and in development; they ship next sprint to target the load cash step, the steepest remaining cliff.
LOOKING BACK
Research first, design second.
Sitting with 18 users for a few days reordered my entire redesign brief. The biggest problems were not the ones I would have prioritized without their input.
Copy is design.
Renaming 'Verify your Identity' to 'Set Up Profile' changed how users perceived the step. '$1.50 on this purchase' landed when '5% back' bounced. Copy did more than layout in this redesign.
Users don't read.
Most participants skimmed at best. If a screen relied on the user reading a sentence to understand it, that screen failed. The redesign had to communicate to skimmers, not readers.
A clear foundation survives the pivot.
The business shifted from tap-to-pay to barcode scan-and-pay mid-development. Because the redesign was anchored in real user problems, the pivot extended the work instead of restarting it.
WHAT I'D DO DIFFERENTLY
I'd add a round of in-store usability testing before public launch. The office sessions caught most of the language and structural problems, but the in-store speed and clerk-visibility gaps only surfaced once the employee beta hit the register. Testing in context earlier would have shaped the barcode pivot before the business asked for it.
LOOKING AHEAD
Ship the barcode hero, OCR card scanning, and contextual card-linking nudge. Target the load cash step, the steepest remaining cliff. Instrument the new flow more granularly so we can iterate without another full research round.