Duration
- 3 months
Inventory Efficiency Report: Turning Supply Chain Chaos into Confident Decisions
How I designed a command-center interface that reduced issue-detection time by 60% and prevented $2.1M in stockout losses for Walmart planning workflows.
Overview
My Role
Lead Product Designer
Team Role
- 2 Product Managers
- 2 BI Developers
- 2 Data Scientists
- 3 Engineers
Target Audience
- Inventory Planners
- Retail Operations Teams
Scope
Research, UX, UI, Prototyping, Testing
Tools
- Figma
- Excel
- Highcharts
- Jira
Industry
- Inventory planners
- Retail operations teams
Platform
Web (Desktop-first)
Users Impacted
300+ daily active planners
Inventory Managed
$2B+ across 150+ DCs
The Challenge
Retail supply chain planners at Walmart managed billions in inventory across 150+ distribution centers and 4,000+ stores.
Yet they struggled with a critical problem: they couldn't see the risks until it was too late.
Manual reconciliation
Risk lived in spreadsheets across Systems A–E — rebuilt every week before it hit the platform.
Manual
Judgment happened in the sheet
Planners rebuilt risk outside the weekly review — conditional fills instead of a shared exception queue.
Legacy Precima
The dashboard existed — attention still jumped across nav, KPIs, map, and tables before a clear next action.
Platform
Signal buried in density
KPIs, map, and tables competed for attention — exception paths arrived after the hunt.
I spend 4 hours a day just trying to find which SKUs need attention. By the time I figure it out, it's often too late. I need the system to tell me what matters and why it matters—not show me 10,000 rows of data.
The Core Problem
Fragmented
3 separate legacy systems + Excel reports requiring manual data reconciliation. Planners opened multiple tools just to answer one question: which SKUs need attention today?
Reactive
Issues surfaced days after they became critical. By the time a stockout appeared in reports, it had already cost the business thousands in lost sales.
Opaque
No explanation for why a SKU was flagged. Planners didn't trust automated alerts because the system never showed its reasoning — it was a black box.
Overwhelming
Thousands of rows with no prioritization logic. Everything looked equally urgent, so nothing felt urgent — more time finding problems than solving them.
The Stakes Were High
$50K–$200K cost
Stockouts
Empty shelves = lost sales
Outages cascaded across DCs before anyone owned the SKU exception.
Unowned exceptions turned local misses into lost revenue.
Warehouse gridlock
Overstock
Capital tied up + operational waste
Buffer inventory rose because planners couldn’t see true weeks of supply.
Capital sat idle while true WOS stayed invisible.
Ripple effects
Delays
Supply chain disruptions
Late insight turned local misses into multi-DC recovery work — the cost lived outside any one report.
Local lag became network recovery cost.
User Pain Points Quantified
4+
Hours
Average daily time hunting critical SKUs across 3 systems.
6 Sources
45% Adoption
24+ Hours lag
User Snapshot: The Inventory Planner
The real job: finding the right problems to solve first.
Responsibilities
- Prevent stockouts across 12+ DCs without bloating on-hand
- Cut excess where unit turns fall into the red
- Coordinate transfers when one DC is long and another is short
Metrics Tracked
Stockout RateRisk Monitor
Weeks of SupplyBalance Signal
Unit TurnsEfficiency
Aged InventoryCapital Drag
A Typical Day
Four calendar moments from Monday triage to the actions that stick overnight — framed in Inventory Efficiency report language, not a borrowed SCP module.
Success Criteria
To solve this problem, the design needed to shift the tool from reporting data to directing decisions.
We defined three success criteria.
Speed to Insight
Reduce the time it takes to identify a critical stock risk from hours to minutes.
Time-to-Detect
Decision Confidence
Enable planners to trust system insights without needing to manually validate raw data.
Trust
Workflow Adoption
Make the tool the planner's primary workspace, replacing the daily spreadsheet ritual.
Daily Active
Process Overview
Understand the System
Systems Thinking in Action
What I did
Stakeholder Mapping
Mapped planners, buyers, DC ops, and leadership to see who owned each decision handoff.
4 User Groups
Data Dependency
Traced WOS, forecast, transfers, and assortment feeds that fragmented every investigation.
6 Data Sources
User Shadowing
Sat with planners through morning triage — spreadsheet ritual, tools, and exception loops.
40+ Hours
Log Analysis
Reviewed audit trails to find where context dropped between heatmap, queue, and transfer.
Audit Logs
Define the Workflow Truth
The redesigned path
Current-state sprawl collapses into an exception-first path — triage, localize, validate, act — without leaving the platform.
WasReport X→Report Y→Report Z→Export→Excel→Decision outside4+ hrs / cycle
One exception-first path · Triage → Localize → Root cause → Validate → Act · context preserved end-to-end
What changed
Workflow friction → guided path. These are section proof points — not hero impact metrics.
Fragmented before
Reports · systems — context rebuilt on every hop
No saved state
Decision happened outside the platform
Exception-first
See risk before the table — drill instead of hunt
Closed loop, same session
Root cause → validate → act in one workflow
Report Architecture: Mapping Decision Modules
Architecture board · curated crop
Annotated distribution planning board from the Miro workshop — sticky decisions, KPI modules, and chart zones that shaped the report.
Core decisions
A
Default view = exceptions only
Open on SKUs at High/Medium risk — the full catalog waits one intentional drill away.
B
Top 10 prioritization
Surface highest revenue-impact and exposure items first so triage beats alphabetical hunting.
C
Risk scoring on the board
Explainable Red / Amber / Green classification keeps severity readable without a rebuild.
D
Progressive disclosure
Scan → understand → act stays one path — context (SKU · DC · period) never drops between views.
Concept Exploration
Catalogue → table & chart report
From the reports catalogue into a chart-and-table report — fast scan from a familiar export mental model.
Geo map ↔ heatmap investigation
Same report spine with geo map or heatmap mid-board so planners navigate risk by place and time.
Catalogue → KPI strip report
Catalogue opens into a KPI-led report so planners orient on health before opening inventory risk.
Convergence
We combined the high-level scannability of the Heatmap
with the drill-down power of the Table.
Map vs heatmap — hybrid model
We combined the high-level scannability of a heatmap with the analytical power of detailed investigation views. Inventory issues are rarely spatial.
Layer 01
Heatmap risk surface
See concentration of High/Medium risk across store–DC cells before opening any row.
Layer 02
At-risk queue
Top SKUs ranked by exposure — triage list replaces spreadsheet hunting.
Layer 03
SKU detail table
Full inventory evidence waits one drill down — available when validating, not when scanning.
Core Framework
3CP — Probability · Explainability · Adaptability
Hover each pillar to play its proof — exposure fill, why-trace, or role-depth path.
Probability
Rank exceptions by likelihood and business exposure so planners start on the risks most worth their time.
Explainability
Show the drivers behind every flag — velocity, DC concentration, forecast drift — without a raw-data rebuild.
Adaptability
Progressive disclosure lets roles scan, investigate, and act without losing product, DC, or period context.
Intelligence Layer Design
Input: DC/Store network + weekly signals
Core mechanic: Exception-first heatmap + explainability
Output: Prioritized actions + audit-ready validation
The workflow begins with an exception-first heatmap that surfaces inventory anomalies across distribution centers and time periods.
Instead of scanning dense tables, planners can quickly identify unusual patterns, inspect individual cells via hover, and move directly into deeper investigation when something looks off.
This allows users to shift from manual data hunting to rapid anomaly detection.
01
Scan
Heatmap
02
Investigate
Tooltip + Drilldown
03
Explain
Root Cause
04
Act
Alerts + Actions
05
Validate
Table
Understand Why It’s Happening
Once an anomaly is identified, planners can open the root-cause panel without leaving the view.
The panel aggregates multiple signals — including OTIF performance, forecast variance, sales velocity, and inventory levels — to explain the drivers behind the risk.
This removes the need to switch between multiple reporting tools and allows planners to diagnose problems directly within the workflow.
Take Action Without Losing Context
Once the planner understands the problem, the system supports action directly within the same workflow.
Rather than exporting data or switching tools, planners can evaluate item performance, adjust decisions, and prioritize corrective actions while maintaining the full analytical context.
This reduces workflow fragmentation and allows planners to move from analysis to action more quickly.
Validation
Method → findings → change log
Five moderated sessions on the Figma hi-fi. Task: identify and fix three inventory risks. Evidence drove the labels, filters, and approval path — not chrome.
Sessions
5 moderated
Prototype
Figma hi-fi
Task
Identify & Fix 3 Risks
Task success
100%
Time on task
−60%
FindingEvidenceChange made
Terminology
4/5 planners stalled on “EOP” and “velocity anomaly” before scanning the risk queue.
Renamed labels to plain language + inline glossary tooltips on first hover.
Filter Blindness
3/5 scrolled the full inventory table without applying Region or Category filters.
Pinned filter bar above the heatmap and added empty-state guidance when no lens is set.
Action Anxiety
Users hesitated on transfer approve — “what if I’m wrong?” blocked completion.
Added impact preview + confidence state before confirm; soft default to Review.
Impact
The pilot launched in Q4 and immediately transformed how the inventory team operated.
I used to dread Monday mornings. Now I can clear my risk queue before my first coffee. It's not just faster; it's less stressful.
Issue detection
48 hrs< 1 hr
98% faster
Stockout risk
$3.5M proj$1.4M
$2.1M saved
Tool sprawl
4–5 tools1 tool
Consolidated
Trust score
2.1 / 54.8 / 5
+128% lift
* Pilot sample: 12 users across 2 regions over 6 weeks.
Workflow decisions
Exception → investigate → approve → resolve
The inventory command center follows one decision path — from exception signal through investigation, transfer approval, and resolution tracking.
Scan exceptions
DecisionStart on the heatmap and top-10 risk queue — surface only SKUs at High/Medium risk before opening the full inventory table.
Investigate root cause
DecisionDrill into item detail — velocity trends, DC concentration, and explainable risk scoring reveal why the SKU flagged.
Review transfer options
DecisionCompare DC transfer paths and exposure — keep human-in-the-loop approval so planners validate before execution.
Approve & monitor
DecisionApprove the transfer and track resolution — batched updates refresh status within 15 minutes without destabilizing legacy ERP.
Reflection
What this work changed in my practice
Redesigning inventory efficiency reporting taught me how operational trust, legacy data constraints, and planner workflows shape every design decision — from exception surfacing to transfer approval.
Involve engineers early
Early feasibility checks with engineers prevented late rework on data pipelines, hierarchy logic, and the 15-minute batched update model.
Test with real inventory data
Using real data instead of Lorem Ipsum surfaced edge cases — DC concentration, velocity anomalies, and legacy cleanup — that shaped the exception-first heatmap.
Stay close to planners
Weekly office hours with planners kept the decision workflow grounded in how teams actually investigate stockout risk and approve transfers.
Define the baseline earlier
I spent too long on geographic view concepts and underestimated legacy data cleanup — defining metrics baselines earlier would have accelerated alignment.
Next Case Studies
Supply Chain Performance Report
Configurable enterprise report with hierarchy-aware drill-down — helping teams move from KPI scan to child-level explanation across product, store, and DC views.
