Lambda Supply Chain Solutions

Using digital twin for inventory planning in E-commerce distribution network

1. Introduction – Service level as competitive edge

– Authored by Gaurav Mogra

E-commerce is one of the fastest growing and most dynamic industries in the world today, with no signs of slowing down. This also means, Service Levels has been one of the most critical supply chain factors, often prioritized over Cost and Efficiency. Customer acquisition and driving growth has been the focus area for most e-commerce businesses.

Accordingly to the US Department of Commerce, as of Q2 2021, E-commerce accounted for only 13.3% of total sales in the US[1]. This means E-commerce is nowhere close to maturity, and will continue to grow for the next several decades. This also means the emphasis on service levels is only going to get more aggressive.

Thanks to e-commerce giants like Amazon, today we live in an age where Customers are getting accustomed to quicker deliveries and a plethora of choices to pick from. If an item is stocked out, Customers have several other options to choose from, eventually resulting in lost sales. If an organization offers 3 day delivery, there are other options online that offer the same or similar product at same day or next day delivery.

According to a study conducted by PwC in 2018, “41% of consumers are willing to pay a charge for same-day delivery while nearly a quarter (24%) of shoppers said they would pay more to receive packages within a one-or two-hour window of their choosing” [2] This only shows that faster deliveries and stock availability has become a non-negotiable for any growing E-commerce firm, justifying prioritizing Service Levels over even Cost and Profitability in the near future

2. Supply Chain Planning Areas

There are several planning areas that impact Service Levels for the end customer. These planning areas can be looked at as ‘levers’ that can be used to improve end customer experience, while still keeping costs as low as possible.

• Sourcing: Sourcing strategy can directly impact service levels by affecting inbound lead times and product availability. Choosing to outsource manufacturing to a supplier in China could mean a 2-3 month lead time, and higher inventory levels downstream. However, picking a local supplier could mean shorter lead times, lower inventory requirements but higher product costs. If a stock out event looks likely 3-4 weeks from now, a local supplier can quickly ramp up production to replenish inventory at DC locations. Sourcing strategy also entails being prepared with alternate supply arrangements if a primary supplier has a supply shock and cannot fulfill demand

• Manufacturing / Assembly: In case of in-house manufacturing and assembly, there are several factors that can affect end customer service levels. A firm can decide to manufacture all products in a single centralized plant. This would mean a high level of dependence on a single location, as well as longer in-bound lead times to customer facing DC locations. However, a firm can decide to have multiple production plants that serve different markets and regions. This can ensure shorter lead times to the customer facing DC locations. If any given product can be manufactured in multiple plants, it also provides a cushion to unexpected events by ensuring low dependence on a single plant.

• Distribution / Inventory Placement: From our experience, we have seen that a lot of mid-size e-commerce firms tackle inventory placement decisions using a manual ‘gut-feel’ based approach. This often results in inefficient placement of inventory, resulting in products not being available at the right place at the right time. It also increases interfacility transfer moves to make up for misplaced inventory, resulting in delivery delays for the end customer.

Choosing the right distribution strategy also becomes vital in determining end customer service levels. Higher levels of centralization, with lower inventory targets maintained at customer servicing DC locations could result in higher probability of stock outs and products being unavailable. However, higher inventory levels at the customer servicing DC location results in higher costs due to de-centralization of safety stock. These are important trade offs to consider while developing distribution and inventory placement strategy forany business.

• Transportation: Most e-commerce companies use hub networks of parcel providers, such as FedEx, UPS and USPS, to support their last mile parcel delivery operations. Majority of businesses ship out parcels from their DCs to the nearest available hub. The parcel provider will then ship these parcels from the hub to other hubs closest to the end customer. However, with sufficient load buildup, the business also has the option of bypassing the nearest possible hub and sending FTL shipments to specific hubs closest to the end customers. This concept, called ‘‘Zone Skipping’, also results in quicker deliveries and savings for the business by reducing 3PL hub to hub transfer movements for last mile delivery. Optimization can help

• make more efficient hub selection decisions to minimize last mile parcel delivery costs and service levels.

• Overall Network Design: Designing a robust network is vital in ensuring adequate service levels to the end customer. Finding the optimal number of DCs for a business is a fine balance between cost and service levels. Having too few DCs could be cheap but impact service levels. Having too many DCs could improve service levels, but result in very high costs. In case the network configuration needs to be expanded, choosing the right DC locations would also impact service levels and supply chain cost. Finding the right balance here between service level and cost ensures the business runs seamlessly and efficiently.


3. How Digital Twin can improve planning decisions

Most leading companies take supply chain planning decisions based on regional demand patterns, heuristics and experience. This can potentially lead to sub-optimal fulfilment networks, hurting service levels and leading to higher costs, leaving millions of dollars of savings on the table.

Using a digital twin, on the other hand, is a data driven, scientific approach that uses state of the art optimization techniques, considering millions of flow scenarios and configurations before arriving at the optimal supply chain planning decisions.

At Lambda Supply Chain Solutions, we recently assisted a $2 billion US based e-commerce firm unlock ~15% in logistics savings and improve service levels by ~20% by using a digital twin. We helped the firm with short term, medium term and long term supply chain planning decisions and better align their network with strategic business objectives.

Our team helped the organization with the following questions and recommendations in the Short Term, Medium Term and Long Term for an efficient supply chain network:

Short Term (6 Months – 1 Year)

• Which product should be fulfilled from which distribution center at what time of the year?

• How much of zone skipping is possible with current order pattern?

• Which hubs should be utilized for zone skipping during peak and non-peak seasons?

• Where are opportunities to convert express to ground service levels in the network?

Medium Term (1 Year – 3 Years)

• When do I run out of capacity in my current network?

• Where and when do I need to add capacity in the network in the medium term?

• What should be the split between different temperature zones (in case of cold supply chain)

• Which parcel hub capacities are constraining my ability to serve customers faster? What should I be telling my parcel provider?

• What should be my product flow and inventory stocking strategies?

Long Term (3 Years – 5 Years)

• Is my network capable to handle 5 year growth?

• Are there pockets of demand which are underserved but are growing fast?

• Can I get additional supply chain synergies by combining supply chain networks across different business units?

• What are the new candidate DC locations to consider for network expansion? How large will these facilities be?

• What is the optimal location to manufacture and assemble each product?

/4. How to get started

To get started with using Digital Twins to improve your service levels and lower overall supply chain costs, any business would need the following:

1. Technology Partner:

Optimization using Digital Twin is a computationally heavy exercise and needs to be performed using robust technology solution for the entire process to be seamless.

2. Consulting Team:

Building a Digital Twin is a complex exercise, with several data inputs, nuances and design parameters. Having an experienced consulting team, either in-house or external, is vital in developing accurate optimization models and deriving actionable insights.

3. Project Management Office Team:

A supply chain optimization exercise requires inputs like warehouse cost models, transportation costs, inventory targets, capacity constraints etc. This typically requires high levels of collaboration and project management across various teams to be able to identify the right data owners, process the data and convert raw data into usable model inputs.

Have questions? Reach out to us on

Want to talk to someone ?

Digital twin technology can help you optimize your short,
medium and long-term planning decisions.

© 2021 Lambda Supply Chain Solutions Pvt. Ltd., All rights reserved.