Customer Stories

CognitOps and the Biggest Leading Personal Care & Beauty Company Partner to Drive Speed and Efficiency

Business Profile


Consumer Goods, Personal Care and Beauty

Facilities Profile


WES: Dematic
WCS: Bastian
Other: Sterling Commerce (OMS), JDA (YMS), ProShip (Parcel Manifesting)




2 non-peak, 3 peak

Order Profiles:

Multi-line, Single-line, Hazmat, VAS



Pain Point: Matching labor staffing and equipment activation with order volume and order profile variability

Value Proposition: Reduce operating cost-per-unit

CognitOps | Customer Stories - Sephora

Opportunity: Staffing had been done using tribal knowledge, Excel spreadsheets with static inputs, and monthly forecasts.
Solution: Staffing recommendations are made using historical and predictive work rates and daily order forecast.

Opportunity: Real-time labor allocation was done using siloed views of each department’s current work volumes and labor rates.
Solution: Labor allocation uses dynamic, worker-specific work rates and available work volume and work time to allocate available labor.

Opportunity: Lack of visibility to variables that impact cycle time at the put walls, e.g. order types, work rates.
Solution: Equipment recommendations are made using composition of order pool, work cycle times by area, and real-time work rates.

Time to Achieve: 90 Days
Data Required: Orders, Staffing Levels, Work Rates
Methodology: Plug the Align application to real-time replicated data and activate workforce optimizations and time series data for order completion

CPU reduction of 35%
Order Cycle Time Reduction of 30%

Pain point: Aging orders cost the operation money and impact customer satisfaction

Value Proposition: Reduce upgrade costs by hitting SLA

Opportunity: Visibility to Aged Orders was limited to the day after an order had missed its ship date.
Exceptions are identified in real-time based on trending work cycle times by area.
Shipping Cutoff Notification alerts to which order types are at risk of missing cutoff in advance.
Advanced visibility to aging/ at risk orders by order stage.

Time to Achieve: 30 Days
Data Required: Orders, Containers, Workflow Transactions
Methodology: Time-series-based storing of order transactions data. Algorithm to alert operations when an order(s) will exceed cycle time of a specific fulfillment step

Order Cycle Time reduction of 20%
Upgrade costs reduced by 50%

Pain point: Inconsistent throughput during non-peak days

Value Proposition: Decrease order cycle time to increase available building capacity and to make throughput requirements more predictable

Opportunity: Orders with exceptions are difficult to track down.
Solution: Exceptions are identified in real-time based on trending cycle times by area.

Opportunity: Order release decisions are made based on historical, stale data.
Solution: Order release recommendation takes into context the current state of the facility and upcoming order profiles and types to help minimize order cycle time.

Opportunity: Bottlenecks and constraints aren’t surfaced until gridlock hits.
Solution: Equipment activation and order exception identification prevents gridlock.

Data Required: Orders, Workflow Transactions, Equipment Transactions, Inventory
Methodology: Machine Learning implementation of Graphical Models to optimize decisions such as size of processing batches

Complete order-release visibility from WMS release until ship
Tracking and identification of performance trends by shift over time
CPU reduction of 15%
Order Cycle Time improvement of 20%