Impact of Sales Concentration Over Constrained Delivery Date Distribution Systems
|Gilberto Ciola da Silva - Global Services Supply Chain Specialist |
|Co-Presenter,Camilo Manfredi - Services Director - Distribution & Supply Chain at NeoGrid|
|Co-Presenter, Marcelo Zanin - Supply Chain Consultant at NeoGrid Europe |
What to change?
Considering a stable distribution network in fashion retail systems, carefully designed to promote the best results out of product flows and lead times, one challenge is raised during distribution process reviews: How to setup the best DC-to-Store replenishment days of the week, in a system with limited number of deliveries per store per week?
This situation is usually found when third-part freight contractors specify a limited number of trips on its contracted period, and additional fees can be applied when a change in the schedule is requested. Even worse, if this request contains weekend deliveries, expensive penalties can be applied.
In this scenario, it is expected that delivery schedules are not fully optimized, and there is an opportunity to reduce overall stock levels and also reduce out-of-stock risks using TOC 5-steps and run simulations to prove this hypothesis.
What to change to?
Our goal was to find one standard weekly schedule per store that performs better than current one, measured through Dynamic Buffer Management KPI's and constrained by Total number of trips considering system as a whole. This methodology is based at the principle that in complex system, inject change in one degree of freedom can affect most other links, and as a result cause a significant change.
For evaluation purposes, this experiment will be considered successful if Overall Stock levels are, at least, 15% lower than current levels and Risk of Out-of-Stock is 15% lower than previous average observation.
How to cause the change?
To prove the effect of this simulation, all other input variables remained the same: Lead times, aggregation effect, minimum lots were fixed for this experiment.
The first simulation, the quantity of trips per week was preserved for each store: only the days of the week were changed.
On the second simulation, the quantity of trips per week was also kept for each store, but simulation included weekends as a valid delivery date.
On the third, the quantity of trips per week was changed, but total number oftrips/week considering all stores was preserved.
For each simulation, was queried 8 weeks of known historical data. This data allowed some KPI calculations, such as Potential Lost Sales (Number of days of Zero stock Levels), Actual Lost Sales (Number of Days-of-Sales with Zero stock Level) and resulting Green Zone Level, which its measurements can be correlated with average stock levels over simulated period.
Lessons learned? Include successes, challenges, and obstacles and how they were overcome.
Data availability was a real obstacle to ensure that this simulation was actually successful or not. It was important to be sure about all unit costs considered in each simulation in order to compare with current scenario and normalize the results. This situation was resolved after a lot of rounds of data mining across our databases.
Other caveat was the decision over data samples. Subjects must be significant enough to give most of demand behaviour situations over simulated period. Also, they must meet criteria that was established during trial, such as high sales concentrations during weekends. As a result, samples were narrowed with items with over 25% average sales during weekends and responded for more than 20% of all sales during the past year.
Camilo Manfredi graduated in Electrical, Production Engineering from FEI University. SP, Brazil, Post-Graduated in Administration from FGV Business School. SP, Brazil and Master in Leadership and People Management from FGV Business School. SP, Brazil. Responsible for the Product Roadmap, Technology Development and implementation of complex Supply Chain projects in Brazil. Book author "Customers & Companies – Like Dogs & Cats” and Supply Chain articles writer at blog - www.supplychainmix.com.br. Currently is Services Director - Distribution & Supply Chain at NeoGrid and TOC practitioners team leader. His team implements Consulting Initiatives that include CCPM and Distribution solutions in NeoGrid Customers.
Previous Speaking Experience: Distribution Events (Market Driven), NeoGrid Customer Events (+500 people in the audience), Customer Training (hundreds of companies), and others
Published Works: Book author "Customers & Companies – Like Dogs & Cats” and Supply Chain articles writer at blog - www.neogrid.com/blog.
Gilberto Ciola da Silva Graduated in Civil Engineering from Universidade de São Paulo (POLI-USP), Post-Graduated in Distribution and Logistics from FIA Business Scholl – SP, Brazil. Responsible for implementation of complex Supply Chain projects across Europe and North America operations. Currently is SCM Specialist - Distribution & Supply Chain at NeoGrid and a TOC practitioner, and also a certified TOC Fundamentals professional.
Previous Speaking Experience: Distribution Events (Market Driven), NeoGrid Customer Events (+500 people in the audience), Customer Training (dozens of companies), and others.
Published Works: Supply Chain articles writer at blog www.neogrid.com/blog and LinkedIn only.
Marcelo Zanin is a Supply Chain Consultant at NeoGrid Europe (firstname.lastname@example.org): Graduated in Computer Engineering from PUC University – SP, Brazil and Master in Supply Chain Management and Logistics from FGV Business School – SP, Brazil. Responsible for the development and implementation of the replenishment system for customers in Europe.