Data Science for Sustainability— Green Inventory Management
Simulate the impact of store delivery frequency on the CO2 emissions of a fashion retailer
Green inventory management can be defined as managing inventory in an environmentally sustainable way.
This involves a set of processes and rules reducing the environmental impact of order preparation and delivery.
Most retailers use inventory management systems that use a rule-based approach to replenishing stores and covering demand.
The replenishment frequency is an important parameter that can be used to optimize the distribution network.
As a Data Science Manager, can you simulate the impact of the delivery frequency on the CO2 emissions?
As major retail groups should commit to reducing their CO2 emissions by 2030, this can become your strategic tool for reaching these targets.
In this article, we will explore how to use data science to simulate the sustainable initiative of implementing green inventory management for the distribution of a fashion retail company.
Problem Statement
Scenario: Inventory Management for Retail
You are the Data Science Manager of an international clothing group that has stores all around the world.
The distribution team is managing the replenishment of stores from local warehouses.
Your colleague, the Inventory Manager, is in charge of setting the store replenishment rules in the ERP.
She has implemented a periodic review policy Order-Up-To-Level (R, S)
- The ERP is reviewing stores’ inventory levels (also called inventory on hand) every R days: IOH
- For each review, the gap between the inventory level and the target inventory S is calculated: S — IOH
- A Replenishment Order is created and transmitted to the warehouse with the quantity Q = S — IOH
The idea is to deliver the missing quantity to reach this target level.
After transmission, the order is prepared at the warehouse and delivered to your store after a certain lead time LD (days).
The target stock is defined to absorb the demand variability and the replenishment lead time to avoid stock-outs (empty shelves) at the store.
I will not enter the details of how to set these differences in this article.
However, if you need more information, you can find a detailed explanation in this article.
Among the key parameters of this rule is the time between two reviews, which will drive the frequency of replenishment in your stores.
What if we change this review period?
What is Green Inventory Management?
The review period is setting the frequency of store replenishment order creation.
- For R = 2 days: stores are replenished very frequently
You can set a lower target stock level to cover the demand during the review period. - For R = 15 days: stores are replenished less frequently
The order quantity per replenishment must be higher as your target stock level needs to absorb the demand during a longer review period.
On the left side, we have more store deliveries (with a lower quantity per shipment) for the same duration.
This will definitely impact the efficiency of your warehouse and transportation operations.
Can we estimate the impacts of these two different approaches on the CO2 emissions?
You can simulate these and help your colleague estimate the impact of her inventory rules on the consumption of cartons and CO2 emissions.
Impact on the carton usage
Items are delivered at the stores in cartons containing units picked individually.
If the store orders 5 units of the reference XXX, the operator will
- Open a box of 20 units and take 5 units;
- Take a new box and put these 5 units;
- Complete the box with other items ordered by the store;
We must use additional carton material to create these mixed cartons containing different items. (instead of shipping full cartons)
How many additional mixed cartons do we have to prepare?
You can calculate the total number of mixed cartons using the formula below.
These boxes (or mixed cartons) will require additional packing material that will impact your footprint.
With a higher replenishment frequency, the quantity per replenishment is reduced, and this situation can occur more.
Can we estimate the impact on the Transportation efficiency?
Impact on transportation emissions
As it is the main driver of CO2 emissions, you should also estimate the impact on the number of trucks used and their filling rate.
The review period impacts the number of deliveries during a certain period.
For instance, doubling the delivery frequency will
- Multiply the number of delivers for the same quantity replenished;
- Reduce the quantity per replenishment and potentially increase the space in trucks
How can we translate these insights in estimated CO2 emissions increase?
In the next section, we will translate these operational insights into a simulation model to select the optimal inventory rules.
Have you heard about Sustainable Supply Chain Optimization?
Simulate Green Inventory Management Scenarios
Build a Model with Assumptions
I will use the example of a logistic network that replenishes ten fashion retail stores in Shanghai (PRC).
In this simulation, I will consider
- 90 days of sales of 10 stores located around the warehouse
- 740 unique items (SKU) sold in these stores
- The number of units per full carton provided by the master data
- 12 pieces per mixed carton
- 1 day lead time between order creation and store delivery
Which metrics to use for the assessment of a specific inventory management rule?
Using these parameters and the formula of the previous section, I can estimate truck fill rates and the number of additional cartons needed.
How to translate that into environmental impacts?
Based on the mixed carton dimension and thickness, we can estimate the additional material quantity per carton: 0.3 kg/Carton.
CO2 emissions are estimated using the NTM (Network for Transport Measures) methodology using distances and emissions factors.
💡 Insights
- NTM methodology is adapted to our problem as it considers the impact of the truck filling on your overall emissions.
- We can also consider the filling material (in your mixed cartons) and the wrapping film on your pallets to improve the model.
I shared a detailed tutorial on how to estimate the CO2 emissions of transportation with data analytics in this article
Now that you have developed a model to estimate the impacts, you can define several scenarios with the inventory manager.
Scenarios of Different Review Periods
Considering the stores' storage capacities and transportation resources, she estimated that the maximum review period could be 10 days.
Therefore, you decided to simulate the overall emissions and carton material usage with a review period going from 2 to 10 days.
For each scenario, the model can provide
- The percentage of mixed cartons prepared (%)
- The percentage of partially filled trucks used to deliver the stores (%)
- The total quantity of carton material used (kg)
- The total CO2 emissions of road transportation (kg CO2eq)
I added two first outputs to give an overall vision of the efficiency of your distribution network that can be shared with the logistics teams.
In the next section, I share the results of a simple simulation I conducted using 90 days of sales transactions.
Results of a Green Inventory Management Simulation
Transportation Emissions
The initial assumption was that a lower delivery frequency would improve the filling rate of trucks and reduce emissions.
The simulation confirms this assumption.
💡 Insights
- The minimum number of trips is reached for a review period of 7 days.
- -27% of emissions between Scenario 1 (R =2) and Scenario 6 (R = 7).
- -51% of trips between Scenario 1 (R =2) and Scenario 6 (R = 7).
As we can see that emissions increase for R >7, the optimal rule matches the demand's weekly seasonality.
Carton Material Usage
With a lower order frequency, we assumed that an increasing quantity per order would reduce the percentage of items picked by piece.
The results of the simulation confirm this assumption.
💡 Insights
- The total number of cartons prepared remains stable as the total demand during the simulation period is the same for all scenarios.
- The percentage of mixed cartons is dropping from 27% (Scenario 1: R = 2 days) to 9% (Scenario 9: R = 10 days)
- -65% of carton usage between Scenario 1 (R= 2) and Scenario 9 (R= 10)
The trend shows that R=7 days is the optimal frequency to maximize the footprint reduction.
The positive effects do not only include carton usage and emission reductions.
Productivity & Social Impact
In a Distribution Center (DC), walking time from one location to another during the picking route can account for 60% to 70% of the operator’s working time.
This productivity is measured by the number of cartons picked per hour paid; operators receive bonuses added if they achieve their targets.
The number of cartons picked at each location is a major parameter influencing your operators’ productivity.
For example, with a target of 200 boxes/hour, the operator will need less effort to reach if he takes 4 cartons/stop than 2 cartons/stop.
If you want to explore this topic, I published a series of articles showing solutions to reduce the walking time in a warehouse.
What is the impact of the review period on the number for cartons per order line (stop)?
As expected, the reduction in delivery frequency is increasing the number of cartons picked per line.
💡 Insights
- +65% full cartons per replenishment order line
- Operators will prepare 2.07 boxes more for the same walking distance.
This will reduce human resources variable costs and help operators reach their targets with less physical effort.
Drawbacks: Average Store Inventory Level
As nothing is perfect, there are some drawbacks to increasing the review period.
When you have a lower replenishment frequency, you need to increase the stock coverage in your stores.
💡 Insights
- +108% of average inventory for all stores between Scenario 9 (R = 10) and Scenario 1 (R = 1)
That means you will need additional space for storage in your stores.
Decision-making: Find the Right Balance
You can share these results with the project team in charge of the green transformation of your distribution network.
The discussion can be animated by the sustainability that will set the targets.
A compromise needs to be reached between store managers and logistics teams.
- Store managers need high replenishment frequencies to minimize storage space and avoid stock-outs.
- Logistics teams would advocate for low replenishment frequencies to optimize the distribution flow.
Your role is to provide the data and insights to make the right decisions and support the teams during implementation.
Conclusion
A balanced approach is needed
Like everything in Supply Chain Management, it is a matter of balance.
Depending on the cost per sqm in your store locations, you can allocate more or less space for storage.
However, this additional cost should be considered in relation to the potential savings of warehousing and transportation.
Improve the simulation
If you need additional savings to convince your management, you can improve the model by bringing additional savings calculations.
- Goods handling processes in the warehouse: picking locations replenishment, truck loading, carton packing;
- Receiving at the store: truck unloading and stock management;
- Packing Material: filling material, labels, pallet wrapping;
Most sustainable initiatives involve heavy investments in modern machinery, e-trucks or renewable energy.
Here, you can use your model to compare the ratio of emissions reductions by euro invested with these expensive initiatives to promote your approach.
The idea is to focus on process optimization (supported by Data Science) instead of heavy investment to meet emissions cut targets.
For more examples of sustainable initiatives supported by data analytics, check this article I published a few months ago.
About Me
Let’s connect on Linkedin and Twitter. I am a supply chain engineer who uses data analytics to improve logistics operations and reduce costs.
💌 New articles straight in your inbox for free: Newsletter
📘 Your complete guide for Supply Chain Analytics: Analytics Cheat Sheet
If you are interested in Data Analytics and Supply Chain, look at my website.
Reference
- Network for Transport Measures, Emissions Factors Definitions
- Network for Transport Measures, Fuel Consumption Estimation
- Sustainable Supply Chains, Yann Bouchery, Charles J. Corbett, Jan C. Fransoo, Tarkan Tan
