Supply Chain Responsiveness for a large retailer AGENDA

Size: px
Start display at page:

Download "Supply Chain Responsiveness for a large retailer AGENDA"

Transcription

1 Supply Chain Responsiveness for a large retailer Author: Sunil Anand, Xiaobei Song Advisor: Dr. Chris Caplice MIT SCM ResearchFest May 24 25, 2011 Executive Summary Research Question Background Impact Analysis Methodology Simulation results Conclusion Q&A AGENDA May 24 25, 2011 MIT SCM ResearchFest 2 1

2 Executive Summary The supply chain network moves from being a reactive entity to a responsive entity by Sharing channel information and coordinating stores demands Reallocate stock among stores using real-time demand signals The new network increases channel responsiveness without incurring higher inventory in stores Introduces risk sharing across the stores to counter demand variability May 24 25, 2011 MIT SCM ResearchFest Research Question What are the trade offs between review time (R) and lead time (L) on the inventory buckets such as pipeline inventory, in store inv and safety stock? What is the impact of lead time and review time on a) Vendor to Store (DTS) b) Vendor to RDC to Store (without re allocation) c) Vendor to RDC to Store (with re allocation) May 24 25, 2011 MIT SCM ResearchFest 4 2

3 Background A large U.S. based retailer is currently undergoing a supply chain network transformation. Original State: Supplier-Stores relationship was a Directto-Store format End State: Supplier-Store relationship is in a Centralized Distribution Center format May 24 25, 2011 MIT SCM ResearchFest 5 Background Direct to Store Supplier Responsiveness HIGH (subject to MOQ constraints) Logistics Cost HIGH What do the stores receive ORDERED QTY Direct to Store May 24 25, 2011 MIT SCM ResearchFest 6 3

4 Background Regional Distribution Center Supplier Responsiveness HIGH + 1 day delay (MOQ constraint not applicable) Logistics Cost LOW What do the stores receive ORDERED qty or REQUIRED qty Regional Distribution Center May 24 25, 2011 MIT SCM ResearchFest 7 Impact Analysis Safety Stock LEAD TIME IN WEEKS REVIEW TIME IN WEEKS May 24 25, 2011 MIT SCM ResearchFest 8 4

5 Impact Analysis Pipeline Inventory Lead time = 10 weeks Lead time = 10 weeks Lead time = 7 weeks Review time = 7 weeks Review time = 7 weeks Review time = 4 weeks 6 In Transit Inv Avg. Order per PO Pipeline INV Lead Time Review Time MODEL 1 L=7, R=7 MODEL 2 MODEL 3 0 DTS L=10, R=7RDC L=10, RDC ReAlloc R=4 May 24 25, 2011 MIT SCM ResearchFest 9 Lead time + Review Time Impact Analysis Summary Lead time + Review Time SAFETY STOCK 14 WEEKS 1448 units 17 WEEKS 1596 units 14 WEEKS 1448 units LEAD TIME REVIEW TIME AVG. IN TRANSIT UNITS AVG. ORDERED UNITS 14 WEEKS 7 WEEKS 7 WEEKS 1400 units 2000 units 17 WEEKS 10 WEEKS (INC) 14 WEEKS 10 WEEKS (No Change) 7 WEEKS (No Change) 2150 units (INC) 4 WEEKS (DEC) 2350 units (INC) 1900 units (slight DEC) 1000 units (DEC) May 24 25, 2011 MIT SCM ResearchFest 10 5

6 Methodology/Data Gathering Methodology Periodic review models: DTS, RDC, RDC+Realloc Define key performance indicators and variables Sensitivity analysis Data Gathering Big Retailer actual sales data Simulate normally distributed data for 5 years Simulate low COV data to test model consistency May 24 25, 2011 MIT SCM ResearchFest 11 Key Metrics Input Variables CSL (K value) Lead time Review time Demand Minimum ordering quantity (MOQ) Output Parameters # of weeks not stockout Inventory fill rate (IFR) In Store inventory Inventory position May 24 25, 2011 MIT SCM ResearchFest 12 6

7 Units Millions $20 $18 $16 $14 $12 $10 $8 $6 $4 $2 $ Results High COV with MOQ constraint D28V % 96.65% % 96.44% 96.08% 96.03% (l,r)=(3,7) (l,r)=(4,3) (l,r)=(4,3) DTS RDC ReAlloc Type of Flow Model Total Annual INVEOH per store $$$ Total Annual Inv Position in $$$ %Week STOCKOUT per store First Pass Fill rate service per store 98.00% 97.00% 96.00% 95.00% 94.00% 93.00% 92.00% 91.00% 90.00% Percentage for Stockouts and Fill rates May 24 25, 2011 MIT SCM ResearchFest 13 Results Low COV with MOQ constraint Units Millions $20 $18 $16 $14 $12 $10 $8 $6 $4 $2 $ D28V1 $18 $ % 97.04% 97.46% $ % 96.03% 95.49% (l,r)=(3,7) (l,r)=(4,3) (l,r)=(4,3) DTS RDC ReAlloc Type of Flow Model % 99.00% 98.00% 97.00% 96.00% 95.00% 94.00% 93.00% 92.00% 91.00% 90.00% Percentage for Stockouts and Fill rates Total Annual INVEOH per store $$$ Total Annual Inv Position in $$$ %Week STOCKOUT per store First Pass Fill rate service per store May 24 25, 2011 MIT SCM ResearchFest 14 7

8 Results Sensitivity Analysis of High COV demand IFR kept constant ~97.2% Millions In Store Inventory $15 $14 $13 $12 $11 $10 $9 $8 $7 $6 $5 $4 $3 $2M, 19% $0.9M, 12% $ > Week IFR/In store Inventory sensitive to review time changes Individual stores subject to higher exposure under DTS model RDC model more robust to changes DTS (R ) DTS (L ) RDC (R ) RDC (L) May 24 25, 2011 MIT SCM ResearchFest 15 Management Insights Lead time directly impacts in transit units Review time directly impacts units per order RDC is increasing the channel s responsiveness without incurring higher inventory in stores More insight into value of sharing and coordinating channel information Develop more comprehensive approach to evaluate optimal decisions for the entire channel May 24 25, 2011 MIT SCM ResearchFest 16 8

9 Q & A THANK YOU May 24 25, 2011 MIT SCM ResearchFest 17 Units Millions Results High COV with MOQ constraint Daily Sales D28V % 97.31% 97.62% 96.42% $ % $19 $ % (l,r)=(3,7) (l,r)=(4,3) (l,r)=(5,3) DTS RDC ReAlloc Type of Flow Model Total Annual INVEOH per store Total Annual Inv Position in $$$ %Week STOCKOUT per store First Pass Fill rate service per store 99.00% 98.00% 97.00% 96.00% 95.00% 94.00% 93.00% 92.00% 91.00% 90.00% Percentage for Stockouts and Fill rates May 24 25, 2011 MIT SCM ResearchFest 18 9

10 Results High COV without MOQ constraint D28V1 Millions $25 $20 $15 $10 $ % 96.29% $8 $14 $ % 96.03% $6 $ % 96.38% $ % 99.00% 98.00% 97.00% 96.00% 95.00% 94.00% 93.00% 92.00% Percentage for Stockouts and Fill rates $ (l,r)=(3,7) (l,r)=(4,3) (l,r)=(5,3) DTS RDC ReAlloc Type of Flow Model 91.00% 90.00% Total Annual INVEOH per store $$$ Total Annual Inv Position in $$$ %Week STOCKOUT per store First Pass Fill rate service per store May 24 25, 2011 MIT SCM ResearchFest 19 Impact Analysis Pipeline Inventory LT=7weeks, RT= 7weeks 2500 UNITS Order O1 placed Order O1 in transit Order O1 delivered SS DEMAND ORDERED InTransit Average In transit Weeks May 24 25, 2011 MIT SCM ResearchFest 20 10

11 UNITS Impact Analysis Pipeline Inventory Longer Lead Time - LT=10weeks, RT= 7weeks Order delivered Lead time for order Order placed SS DEMAND ORDERED InTransit Average In transit Weeks May 24 25, 2011 MIT SCM ResearchFest 21 UNITS 4000 Impact Analysis Pipeline Inventory Periodic review LT=10weeks, RT= 4weeks Orders placed SS DEMAND InTransit ORDERED Average In transit Weeks May 24 25, 2011 MIT SCM ResearchFest 22 11

12 Metrics Store 1 SKU 1 Store 1 SKU 2 Store 1 SKU 3 Store 1 SKU 4 Store 1 SKU 5 %of Weeks Stocked out Percentage of weeks were stocked out of the total 52 weeks Annual STOCKOUT Weeks Total number of weeks were stocked out in 52 weeks Annual Units Stocked Out Cumulative number of units there were short First Pass Fill Rate Service Level % =(1 Unfilled Demand/Total Demand) % Average Week ending Inventory on Hand (Units) Average weekly units in store Avg Weekly INVEOH $$ Dollar value of average weekly units in store Avg Weekly INV ORDERED Units Average weekly units ordered in store Avg Weekly INV ORDERED $$ Dollar value of average weekly units ordered AVG Weekly Inv Position Units Average weekly Inventory Position in store AVG Weekly INV Position $$ Dollar value of average weekly Inventory Position in Store Avg Weekly Inv Intransit Units Average weekly units in transit Avg Weekly Inv Arrived Units Average weekly units received at RDC Avg Weekly Inv Intransit+Arrived Units Sum of intransit and arrived units Avg Weekly Inv Intransit $$ Dollar value of average inventory in transit(weekly) May 24 25, 2011 MIT SCM ResearchFest 23 Results Sensitivity Analysis Thousands $800 $700 Cost for every 0.1% IFR increase $600 $500 $400 $300 $200 $100 $ (4,3) (4,4) (4,5) (4,6) RDC (R, L) Cost for every 0.1% IFR improvement increases exponentially as RT becomes longer May 24 25, 2011 MIT SCM ResearchFest 24 12