Online shopping has been on a steady rise since 2010. With the global pandemic, it became a necessity. Consumers who realized the ease of use of online shopping – choosing items from the warmth of their seat and having them delivered in a couple of days – gained a new habit.
The retail giant Walmart’s e-commerce business in the US increased by 97% in Q2 of 2020.
It is now a widely expected fact that online sales will not return to pre-COVID-19 levels even after life goes back to normal. The projected e-commerce sales of Walmart are expected to increase at a higher rate in 2021.
The pandemic brought omnichannel initiatives to the top of the list of priorities for many companies. The Walmart CEO, Doug McMillon pointed out this initiative in his interview: “The stores and online merchant teams are now integrated, and we believe we’ll benefit from that change going forward.”
Before COVID-19, many companies had separate business units running e-commerce and brick-and-mortar stores, corresponding warehouses, and fulfillment centers. They managed online product availability separately from store inventory. This meant that, if they ran out of stock for an item in e-commerce, it was likely that it was not sold online until it was restocked.
Now retailers are competing on the speed of delivery, availability, and convenience for customers. These expectations brought the focus on leveraging the store network to serve online customers. Omnichannel capabilities allow them to fulfill the orders using store inventory and give customers options to pick up online orders in stores and return them at a location that is convenient for them, instead of shipping them back to e-commerce fulfillment centers. This requires integrating data and processes across the network, which requires IT investments, operational work, and additional shipping and handling costs.
Making store inventory available for omnichannel fulfillment is also an opportunity for reducing over-stocks at stores and reducing the need for markdowns.
For omnichannel fulfillment to work effectively, there are many strategic, tactical, and operational supply chain planning decisions requiring advanced analytics.
Here are some important initiatives for retailers to consider if their company is involved in omnichannel planning:
1. Capturing Sales and Fulfillment Data
There are key considerations such as the fulfillment cost, selections of stores for omnichannel fulfillment, operational requirements, and customer expectations about the delivery timeline and the number of shipments. All of these factors need to be tracked and analyzed, which means the sales transactions should include a selling location (where the order comes in), fulfillment location, expected delivery date, fulfillment date, and receiving date.
For demand forecasting and inventory optimization, order date and order location are required in addition to fulfillment date and location. Unless you have visibility to demand by order location, you are likely to keep underestimating the online demand, especially at the time of stock-outs in the e-commerce warehouse.
On the other hand, if some selling locations are not fulfillment locations due to strategic and tactical decisions, the demand should be estimated in respective fulfillment locations and not selling locations. (See Solvoyo whitepaper for more details)
2. Incorporating returns into forecasting and inventory decisions
The return rate for online sales can be as high as 30% or even more for some categories like footwear. Most of the time, the fulfillment location and return location are not the same. Since returns significantly affect the store inventory levels, this factor must be considered in inventory planning.
Keep in mind, different retailers have different ways of processing returns. Some choose to put them into the store stock on the same day, some go through a longer sanitization and QA process where returns will be available for sale after 24 hours, while some ship them back to the e-commerce fulfillment center. It is crucial to define such restriction rules and capture them in the data accordingly.
Buying behavior varies across consumers. For instance, my personal preference is to shop around and wait for deep markdowns and buy clothing items in bulk. It is normal for me to buy four pairs of shoes or eight pairs of pants at once. We do not have to wander around a shopping mall. The items are shipped to our home, and we try them on and eliminate the ones that my spouse doesn’t like 🙂 You can then return a pair of shoes and three pairs of pants to the store in the mall closest to your house.
When you buy four pairs of shoes and eight pairs of pants online, the software of the retailer should estimate the return quantity and location by using historical data. They should even personalize the data if possible, so the inventory decisions for your favorite return store would change automatically.
3. Utilizing store transfers to have a better response time and lower costs
When a fulfillment location is different from a selling location, it takes time and extra cost to fulfill orders. Some brick-and-mortar stores do not ship the products directly to customers but send them to the selling location first. The products are then shipped to customers from the selling locations.
Every time there is a demand, a low volume of products can be shipped from fulfillment stores to selling locations. However, if automatic transfers were made daily/weekly, we would benefit from the economies of scale. The waiting time for customers will be less since selling locations will have the products in stock thanks to stock transfers between stores.
4. Optimizing omnichannel order fulfillment in batches
Most companies are inclined to have business rules. These rules are applied when an order comes, so, fulfillment locations are determined for a specific order instantaneously. However, processing orders in batches allows for optimizing fulfillment decisions. Instead of deciding on the fulfillment locations every time a new order comes, they can be combined and optimization can be run “n” times a day while following the business rules. The “n” can be chosen depending on the number of orders and the required response level for customers. Optimal transportation and multi-stop routing options will be available if the company delivers the orders.
One important consideration is minimizing the number of packages customers receive for their orders. Some customers can have a strong preference for receiving a single package, then, shipping all line items of an order together will go a long way toward customer satisfaction.
The other day, we ordered ten different t-shirts for my wife and they all came in seven different deliveries in five days. We ordered two of the same t-shirt and they did not even arrive together. They probably thought this was cost-optimal but they did not consider our disappointment. We decided not to buy anything from that retailer again. They might have lost many other customers because of their policy.
5. Taking a strategic approach to determining fulfillment locations
Instead of opening all of the store networks for omnichannel fulfillment, it is worth taking time to conduct a network design analysis. The role and capacity of each location can be determined by running a network optimization model. This decision will not be made daily or weekly but it can be revised several times a year. With this work, some locations can only be selling locations to ensure optimal investments for potential fulfillment locations. That being said, the capacity for fulfillment locations could be decided on to serve offline customers better.
Omnichannel Planning Made Easy
We mentioned that there are important factors retailers should consider in their omnichannel planning efforts. Business rules must be incorporated, fulfillment locations must be considered and relevant data must be monitored for accurate predictions. All of these factors are hard to manage using spreadsheets. Most retailers transform and manage their supply chains on digital platforms such as cloud-based software.
An ideal software solution will ensure that omnichannel operations will be managed with the business rules and requirements set by the company. These can include the choices of fulfillment and selling locations which will be easily optimized during the year, based on the orders and the level of customer responsiveness.
The demand will be estimated in their respective fulfillment locations and not at their selling locations. If orders are matched one by one, this estimation can be made with business rules or the optimized answer of the software. Products can be automatically transferred between selling locations based on the prescriptive analytics provided by the software where daily replenishments and transfers are optimized based on the operational capacity.
In omnichannel planning software, all of the required data can be processed by leveraging advanced analytics and machine learning. These data will be used for predictive analytics for forecasting the demand and return rates. Then, prescriptive analytics will optimize the inventory level in fulfillment and selling locations.
A perfectly optimized omnichannel planning process will please families and make sure that they fulfill all their needs from e-commerce retailers!