Tax Stamp Management System

Helping Ghana's licensed manufacturers submit compliant excise orders 28% more successfully by designing for professional users under regulatory pressure

28%

Increase in Orders Success Rate

Less

Errors & Support Calls

More

Trust in Compliance

Design: Adobe XD
Product/Backlog: Atlassian, Aha
Research/Validation: user interviews, UAT pilots
Delivery: demo videos, training decks

The Problem Area

The business problem:

This government-contracted platform managed excise stamp orders for Ghana's tax authority. Each order represented a compliance obligation, and errors led compliance failures, regulatory issues, and delayed revenue collection.

The platform had a high failure rate, with users struggling through complex pricing and compliance rules without proper guidance, resulting in increased support costs, repeated errors, and diminished trust in the system.

The opportunity:

The failure rate wasn't random; it clustered around three moments: product classification, cart visibility during long orders, and registration. If we could design guardrails and clarity into those three specific moments, we could improve the success rate without adding complexity to the rest of the flow. The opportunity wasn't to redesign the whole platform. It was to fix the points of highest failure with surgical precision and prove that compliance and usability are not opposites.

Discovery

To understand real ordering behaviour, I conducted:

Reviews of failed and corrected orders
Analysis of QA reports and error logs
UAT pilots with real users
Stakeholder workshops on edge cases
Interviews with operations and fulfilment staff

Key Insights that guided the decisions:

I needed to understand two things that would limit every decision:

Regulatory reality: The design challenge was to make the mandatory feel manageable, not to negotiate with compliance.
Physical context: users accessed the platform on mobile factory floors and warehouse environments.
User type: These were professional, repeat users. A design that prioritised hand-holding over efficiency would fail with this audience.

User Pain Points

This resulted in:

Low order success rate
High rework and correction time
Increased operational cost
Heavy helpdesk dependency

Technical Constraints

Compliance flow - Before & After

Conversion funnel

Registration flow

Ordering flow

Strategic Prioritisation

I mapped every identified problem against frequency of occurrence and cost per error. Classification errors were high frequency and high cost fixing them was the highest-leverage intervention. Cart abandonment was high frequency but lower per-incident cost. Registration drop-off was lower frequency but had a long-tail business cost every abandoned registration was a manufacturer not using the platform. I sequenced the work accordingly: classification first, then cart, then registration. I explicitly deprioritised UI polish and navigation improvements until these three were resolved.

Key Decisions

1. Introduce a persistent cart with real-time totals

Insight

Users lost awareness of quantities and pricing during long orders, requiring correction and resubmission.

Decision

Designed a persistent cart with real-time totals and quick editing.

Impact

Users could orient themselves at any point in a long order without losing progress. The 28% improvement in overall order success rate was substantially driven by this change.

Tradeoff: A persistent cart panel reduces the available screen area for product selection, a real cost on mobile. Engineering pushed for a collapsible cart to recover space. I prototyped both. The collapsible version tested worse: users forgot it existed and reverted to the abandonment behaviour we were trying to fix. The cart needed to be persistently visible to change the behaviour, not just available when remembered. I kept it persistent and worked with engineering to optimise the layout for mobile rather than hiding the cart.

2. Stamp Categories Were Confusing

Insight

Domestic and imported stamps were frequently misclassified.

Decision

Introduced visual product cards with images and color-coded category tags.

Impact

Minimised classification errors and reduced compliance-related support tickets.

Tradeoff: Visual cards take more space than text lists, reducing the number of items visible per screen. Engineering was concerned about performance on low-end mobile devices with image-heavy layouts. I accepted the density cost because the classification error rate was the highest-cost failure in the system. For performance, I worked with engineering on lazy loading and image compression rather than reverting to text lists.

3. Replace the flat registration form with a guided step-based flow

Insight

Compliance-heavy onboarding discouraged new users.

Decision

Created a guided, step-based registration flow with clear progress indicators.

Impact

Improved onboarding completion and lowered early-stage drop-off.

I changed the ordering flow back to single-screen with tight inline validation and reserved the wizard for registration only. This is the core principle I use now: task frequency and user expertise level should determine the interaction pattern, not the desire for visual consistency.

Outcome

Process

How I worked?

Fail logs first
Observe + interview
Map the rules
Prioritise by cost
Prototype + UAT
Ship + train

What I would do differently?

I would enhance post-launch product monitoring and continuously track order failure rates by category and user segment to ensure sustained improvement. Involving mobile users earlier would help identify failure patterns more effectively. Additionally, I propose reintroducing a post-submission order review screen to summarize user submissions and their compliance status, closing the feedback loop left open by the current design.