OPEX Allocation Tool Redesign

Services:

Systems thinking | UX Research | UI Design

Enterprise Financial Operations

Amazon

Redesigning the quarterly allocation review workflow for financial managers - replacing a legacy reporting tool with a verification-first dashboard that surfaces all decision-critical context on a single screen

Review time before ~ 9.7 Hrs to 33 Mins

Review Time After ~ 33 Mins ie. 94% reduction

Time reclaimed per cycle ~ 9 Hrs

The Problem Statement

Managers were spending most of their time finding data, not reviewing it

Financial managers processed thousands of allocation change requests each quarter using Cognos - a tool designed for data storage, not decision-making.

Every line item required a manual ID lookup with high latency, forcing managers to maintain multiple tabs and cross-reference spreadsheets to surface the context needed for a single approval.

The system treated expert users as investigators. They were doing data retrieval work, not the high-value judgment work they were hired for.

The previous design

Research

Contextual Inquiry and Interview Questionnaires

To understand the real workflow, I used interview questions and contextual inquiry - shadowing financial managers at their desks during peak quarterly review periods cross-checking PM’s reported descriptions.

This method captured the actual pain: load times, tab-switching patterns, and the informal workarounds managers had quietly developed over time.

01
Contextual inquiry

Watched managers navigate Cognos in real conditions to identify where the workflow broke down - not where they said it broke down.

During peak quarterly review periods to watch how they physically interacted with the system.

02
Friction Mapping

Logged every point where managers paused, switched tabs, or opened a spreadsheet.

These were the moments the interface failed them.

03
Decision trigger interviews

Asked stakeholders: what specific data points are required to approve or deny a request?

This defined the minimum viable information surface

The key insight:

managers weren't struggling because they lacked skill. The system was hiding the context they needed to do their job.

The Design Thinking

From exploratory to verification-first

The core shift was reframing the user's task. The old system required managers to search and assemble context.

The new design already knows what they're looking for - and surfaces everything needed to confirm or reject a request in one view.


01
Before

  1. Manual ID lookup per line item

  2. Cryptic codes, no natural language

  3. Exclusions buried or invisible

  4. No ownership context on screen

  5. Multi-tab, high cognitive load

02
After

  1. All metadata pre-computed, surface-level

  2. IDs replaced with plain language labels

  3. Exclusions surfaced proactively

  4. Owner and client visible per line item

  5. Single-surface, scan-and-decide

UI x Information Architecture

01
Information Architecture

  1. Reorganized the data hierarchy so Channel, Exclusions, Ownership, and Product type appear together

  2. Eliminated the need to toggle between views for a single decision

02
Natural Language Normalization

  1. Replaced cryptic numeric IDs with normalized labels.

  2. Reduced the translation work managers performed mentally on every row.

03
Proactive Exclusion Surfacing

  1. Exclusion flags appear inline during review rather than requiring a separate lookup.

  2. Turned a reactive error-catch into a built-in quality check.

04
Audit Trail By Design

  1. Owner and client displayed on every line item as a structural feature - not an add-on.

  2. Compliance readiness built into the default view

Time Reclaimed~ 9.7 Hrs to 33 Mins

The Outcomes

Measurable impact on efficiency and risk

94%

Time reduction

Per-cycle review time dropped from 9.7 hrs to 33 min for 1,000 IDs.

Same Headcount

Higher throughput

Team handled increased request volume without added staff or overtime.

Fewer Errors

Decision confidence

Reduced misallocation rate; managers reported lower decision fatigue.

Built-in

compliance guardrail

Exclusion visibility reduced the risk of approving invalid allocations.

The Results and Impact

a. Qualitative impact

  • Trust in Data:

    • By surfacing exclusions and owner names, created a sense of security.

    • Users no longer fear that they are approving an "invisible error."

  • Flow State:

    • Removed the "toggle tax," allowing managers to stay focused on the analysis rather than managing the interface.

    • This leads to higher job satisfaction and lower burnout.

b. Quantitative impact

  • Resource Reclamation:

    • Reclaiming ~9 hours of senior-level financial time per request

      means that

    • Senior staff can spend that time on high-value strategic planning instead of data maintenance.

c. Business Value

  • Risk Mitigation:

    • The prominent surfacing of "Product/Channel Exclusions" serves as a built-in compliance guardrail.

    • Reduced the likelihood of costly misallocations.

  • Audit Readiness:

    • This design effectively created an on-the-fly audit trail, which makes quarterly or annual financial reporting significantly faster and less prone to manual errors.

Areas of improvement

  • Scalability of Data Visualization:

    • While I solved the "drill-down" issue, the next step is moving toward predictive insights.

    • I could improve by exploring how to highlight anomalies automatically (e.g., "This request deviates by 20% from typical channel spending") rather than just presenting raw data.

  • Iterative Testing Loops:

    • Could improve by implementing more formal usability testing sessions with the Financial Managers during the design phase (e.g., A/B testing two different dashboard layouts) to quantify the "time-to-decision" before the final build.

  • Change Management Advocacy:

    • I solved the UI problem, but future iterations could focus on the human side of the transition—creating on-boarding or "bridge" documentation to help legacy users shift from Cognos to your new platform without anxiety.

Key Takeaways

A. Opaque data is the real barrier

  • Users weren't struggling from lack of skill - the system was hiding the "why" behind the "what."

  • Surfacing context transformed the experience without changing the underlying data.

B. Cognitive Load as a Mechanical Constraint

  • Bridging industrial design thinking with digital UI: friction in a system - whether physical or cognitive - can be measured and engineered out.

The Reflection

  1. A system is only as strong as its weakest connection

  2. By applying that same rigor to UX—identifying the "pipes" of financial data and cleaning up the flow—I was able to deliver a solution that feels less like a software update and more like a tool for precision decision-making

C. Verification-first over exploratory design

  • Expert users don't need guided exploration. They need complete context delivered at once so they can make rapid, confident decisions.

Next

Enterprise Metrics UI