Case Study 01

Pharma
Inventory
Optimization

Unified data visibility enabling $20M+ projected annual savings for a global manufacturer.

Role

BI Consultant

Duration

6 months

Tech Stack

SAP • Snowflake • Microsoft Fabric • Power BI

01 The Challenge

“We are throwing away millions in expired product, but nobody can tell us exactly how much, where, or why.”

A global pharmaceutical manufacturer faced a critical visibility gap. Data was scattered across SAP, custom ERPs, and spreadsheets. Manufacturing operated disconnected from demand planning, leading to a costly cycle of overproduction and expiration.

1.2B

Rows of Data

4

Disconnected Systems

0

Unified Reports

02 Core Indicators

Days Inventory On Hand

DIO

The number of days current inventory would last at the current sales rate.

A DIO of 90 means you're holding roughly three months of stock. In pharma, DIO must be evaluated against shelf life.

Discard Rate

DR

The percentage of manufactured units that are destroyed before sale.

A 5% discard rate might be acceptable; 30%+ signals systemic overproduction or demand forecasting failure.

Turnover Rate

TR

How many times inventory fully cycles through in a year.

Higher turnover generally indicates healthier demand-supply alignment. Low turnover combined with short shelf life is a red flag.

Sell-Through Rate

STR

The percentage of received inventory that actually sells within a defined period.

Useful for identifying products that move immediately (potential underproduction) versus those that sit (overproduction risk).

Shelf Life

SL

The window from manufacturing date to expiration.

The gap between DIO and remaining shelf life determines discard risk. Products expiring before they can be sold are a guaranteed loss.

03 The Solution

Architecting the Truth

We built a unified data pipeline that ingests, cleans, and harmonizes inventory data from all global sources.

Layer 01Source Systems
SAP S/4HANA
SAP S/4HANA
Legacy ERPs
Legacy ERPs
Excel Planning
Excel Planning
Layer 02Data Integration
Python Scripts
Python Scripts
Dataflows
Dataflows
Snowflake
Snowflake
Layer 03Semantic Layer
Microsoft Fabric
Microsoft Fabric
Layer 04Presentation
Power BI
Power BI

Core Metrics Output

Days Inventory On Hand
Sell-Through Rate
Discard Volume & Value
Margin by SKU
Turnover Analysis
01

Inconsistent IDs

Product codes varied across acquisition-heavy legacy systems.

02

Granularity Mismatch

Daily actuals vs monthly forecasts required complex interpolation.

03

Data Gaps

Missing historical fields necessitated probabilistic backfilling.

Implementation Timeline

Key Activities

  • Stakeholder interviews
  • Source system inventory
  • Data availability assessment
  • Define core metrics (DIO, turnover, discards)

Deliverables

  • Data source map
  • Metric definitions
  • Integration requirements

Main Challenge

No single source of truth existed - each team had their own version of 'inventory health'

Key Activities

  • Extract from SAP, ERPs, Excel
  • Standardize product codes
  • Reconcile time granularity
  • Build transformation pipelines

Deliverables

  • Unified data model
  • ETL pipelines
  • Snowflake staging layer

Main Challenge

Product identifiers varied across systems - matching required fuzzy logic and manual validation

Key Activities

  • Build semantic model in Fabric
  • Develop Power BI report
  • Validate metrics with finance
  • Iterate on visualizations

Deliverables

  • Power BI dashboard
  • Quadrant framework
  • SKU-level drill-through

Main Challenge

Balancing detail (SKU-level) with executive summary views required multiple design iterations

Key Activities

  • Portfolio segmentation
  • Root cause analysis
  • Savings quantification
  • Recommendation development

Deliverables

  • $20M+ savings projection
  • Optimization shortlist
  • Divestiture validation

Main Challenge

Connecting operational data to financial impact required close partnership with finance team

Key Activities

  • Executive presentation
  • Planning team enablement
  • Monthly review cadence
  • Threshold alerting (planned)

Deliverables

  • C-level adoption
  • Strategic decisions enabled
  • Continuous monitoring

Main Challenge

Shifting from one-time analysis to operational tool required change management

04 Portfolio Analysis

The Efficiency Matrix

We developed a quadrant model to segment the entire product portfolio based on inventory health and discard risk.

Filter By:
Days Inventory On Hand (DIO) →0300600Discard Rate (%) →02550ScaleMonitorWatchOptimize

05 Impact

$20M+

Projected Annual Savings

By identifying divestiture candidates and optimizing production schedules for high-risk products.

From Reduced Discards~80%
From Recovered Market Share~20%

Ready to optimize your operations?

Start a conversation