Dark store Applying supply chain processes that meet the demands of modern grocery retailers  Joshila Makan, CEO WCS UK

Dark store Applying supply chain processes that meet the demands of modern grocery retailers Joshila Makan, CEO WCS UK

E-Commerce continues to transform the grocery industry, with increasing demand from consumers to apply the new life principles of instant gratification to their shopping habits – grocery retailers are struggling to keep up. Here we look at the overall trends, as well as insights into supply chain processes and applications that will help retailers fulfill the demands in this growing market segment. As well as what technology can do to assist across order, warehouse and labour management.

• The evolution of Grocery e-commerce
Grocery internet retailers have been split into two different groups: existing grocery retailers that have expanded into the e-Commerce space and new entrants that have a pure e-Commerce model. Existing retailers find themselves with a new challenge of choosing the best location from which to fulfil their e-Commerce orders.

It is common practice to fulfil e-Commerce orders from an existing store. Here the retailer’s employees are “picking” the e-Commerce orders alongside regular shoppers. These orders can then be picked up by the customer or delivered to the customer’s home. Re-purposed or new stores have also been commonly used as dedicated e-Commerce fulfilment centres. These are laid out in a very similar way to a traditional store but there are no shoppers in the store; it is dedicated to the picking of e-Commerce orders by the retailer’s staff. Filled orders are typically delivered directly to the consumer or picked up by the customer either at the store itself or a pick up location. For a bricks and mortar retailer the pickup location is often a retail outlet.

Like a re-purposed store, an e-fulfilment centre (or dotcom centre) is a location that is specifically designed for fulfilling customer orders directly, with orders typically being delivered directly to the customer or to a collection site for pick-up by the customer. An e-fulfilment centre, however, is laid out and functions less like a retail store and more like a distribution centre.

These operations often make use of automation such as goods-to-picker selection and sortation to reduce labour overhead. The size and cost of operating an e-fulfilment centre in the grocery industry with issues such as product freshness, special handling requirements, size of order and speed to customer, mean that e-fulfilment centres are only likely to be viable solutions close to urban areas where the throughput to meet the needs of the population will support the operation.

What is a Dark Store?
The term “dark store” has come to mean any dedicated location that is designed for fulfilling customer orders rather than picking them in an existing retail outlet. This could fall into the category of a dedicated store or an e-fulfilment/dotcom centre.The greater storage capacity of a dark store allows for a superior product assortment than can be achieved in a standard supermarket. Besides ensuring a full product range (e.g. all soup flavours), specialist ranges (e.g. premium, ethnic, gluten-free, locally-sourced), and products not easily accommodated in a grocery store (e.g. beer kegs) can be provided, encouraging customer loyalty and the flexibility to respond to consumer demand and tastes. (Rigby, 2014).

For the remainder of this article the term “dark store” is used to mean any location that is dedicated to fulfilment of e-Commerce orders, whether a dedicated store or an e-fulfilment centre.

• The customer experience

The types of devices used for placing e-Commerce orders is broadening. However, the design of the site and the shopping experience is only part of the story.

Whether the user interacts with the retailer via a desktop website, mobile website, mobile application, virtual store, smart television or other means, the experience will become more personalised and immersive. The order fulfilment process behind the shopping experience is critical. Regardless of how pretty and personal applications are, if an order does not arrive with the items required, in an acceptable condition, and on time, then the customer is unlikely to revisit the process or that company altogether.

Customer Delivery
The last mile of getting the product to the customer is the most expensive in the supply chain, and has been an historic issue with customers adopting e-Commerce for shopping. For grocery retailers there are four delivery options available:

Store collection: The e-Commerce order can be delivered to an existing store in the retailer’s network. The customer then collects the order from the store.

Dark Store: Although customers do not enter a dark store, it can be used as a pick-up point.

Selected pick-up points: Dedicated pick-up points which allow storage of the orders awaiting collection can also be used. This is gaining popularity with dedicated e-Commerce retailers to attract consumers who have difficulty scheduling a home delivery, and can be particularly attractive when placed along popular commuting routes in urban areas.

Home Delivery: a delivery direct to the customer’s home.

Order Processes: Small Top-up Orders versus the Big Shop
Even in areas where grocery e-Commerce is growing there is limited adoption for using it as a mechanism for small top-up orders. The typical use of e-Commerce ordering is for the large weekly shop. An ability to provide an affordable “top-up shop” option will increase the share e-Commerce orders have of the total grocery market.

Customer Relationship Management
Customer purchase history from all channels, online and in-store, should be combined in a single customer history experience. This information can be combined in the customer order website to make recommendations to the customer, and can increase the loyalty of the customer to the brand. If the customer receives recommendations about items (or alternative items) they typically buy in a top-up shop, there is less chance of losing this business to the competition.
• Supply chain network integration

There are considerable supply chain benefits from locating a dark store at or near an existing warehouse.

Location of Dark Store Including in Existing Warehouse
It is common, especially in Europe, for retailers with a mixture of convenience stores and hypermarkets to have two picking lines. The first has unit picking locations and holds the subset of the product range that is stocked
in the convenience stores. The second is a full case pick line for servicing the hypermarkets and picking case replenishments for the single unit pick line. This same approach can be used as a retailer’s first pass at establishing a dark store operation.

Demand Forecasting: Locating the dark store at or near a current distribution centre will allow aggregate forecasting for dark store and distribution centre, reducing the overall safety stock requirement.

Replenishment: Warehouse inventory can replenish the dark store directly.

Transportation: The delivery of pickup orders to the store can be combined with a regular warehouse delivery to the store.

Dedicated to Web Customers
A dark store is typically dedicated to the fulfilment of individual consumer orders. However, there are other opportunities for a retailer that has a large store pick-up business if the dark store is considered as an extension of the store’s stock room.

Slow moving items: Store demand that is provided from a national or slow moving distribution centre can be delivered to the dark store and cross docked to go onto the same vehicle as the individual customer orders that are being delivered for pickup at the store. This same xxxxxxxxx of using the dark store as a hub can be used to supply high value items to the retailer’s bricks and mortar stores.

Store out of stocks: If required, dark store stock may also be used for occasionals tore replenishment to stores within its delivery radius. This can be restricted to promotional items or items that have had an exceptional demand. Voumes should be restrited if needed, replenishing for example a single pallet or roll cage of only the most important items.

Although benefits can be attained by adding some store replenishment to the dark store’s throughput, care should be taken not to interfere with its primary role of meeting direct consumer demand.
• Order Management

For many years grocery retailers have used purpose built (packaged, custom or in-house) order management systems. This is due to the decision rules surrounding the sourcing of an order being more complex to what is offered in a typical ERP system.

Creation of Order
There are numerous entry points for customer orders. Orders can come from a website, a mobile application, a virtual store, placed in-store or from a regular recurring order. There may be multiple orders from these different sources placed to become a single final customer order. Combining these orders for efficient fulfilment is key, along with retaining separate order details for clear reporting to the customer and for tracking demand by different channels.

Selection of Appropriate Dark Store
Traditional order management systems are very good at sourcing orders based on store placing the order and the channel (for example from a regional warehouse, national warehouse or direct from the supplier). Some of these systems also consider inventory available at these different points when making the decision.

However, these rules should be extended to allow sourcing by postal code and order size. The checking of inventory can also be extended to cover the facility’s capacity and current volumes.

Fulfil Order From Store
There may be situations where fulfilment of e-Commerce orders is most appropriate from an existing store. This can be used as an initial low cost approach by a retailer to enter the e-Commerce channel. For more rural areas with lower throughput, this approach may be ideal, allowing the retailer to meet consumer demand without the overhead of supporting an additional operation. An interesting application for this may be to introduce customers to using e-Commerce for the small top-up shops. If an order is below a certain size it can be sent directly to the pick-up store for processing. The use of stores may also be considered as part of an agile order fulfilment approach. If the order management system determines that there is insufficient inventory at the appropriate dark store or the centre does not have the capacity to supply an order, part or all of the order can be directed to the retail outlet.

When there is limited availability of the ordered product, the ideal process is to inform the customer when the order is being placed and to offer them a choice of alternate products. The order management system should be able to provide this information to the ordering application. In addition, capturing the original requested item is key to managing the actual demand for future planning.

Inventory Visibility and Allocation
The management of e-Commerce orders requires order management systems to have visibility of both dark store inventory and retail outlet store inventory (on display and in the stock room).

A critical requirement for customer service in the e-Commerce environment is that the order be fulfilled as requested by the customer. To ensure that accurate order confirmation is sent to the customer, inventory should be immediately allocated to the customer order as soon as the order is sourced.
of supporting an additional operation. An interesting application for this may be to introduce customers to using e-Commerce for the small top-up shops. If an order is below a certain size it can be sent directly to the pick-up store for processing.

• Application of material handling technology

The importance of automated material handling equipment to dark store operations and the future of e-Commerce order fulfilment should not be underestimated.

Dark Store Layout
As mentioned earlier, the merchandise layout of a dark store is not subject to the same constraints as a grocery distribution centre. Typically, a grocery distribution centre is laid out in a similar way to the store, thus reducing
the amount of labour required at the store to put the product on display. Although it will still be subject to the constraints of temperature and contamination, a dark store is not subject to the same constraints in the bulk of the ambient area and can instead focus on a layout which optimises productivity.

Design is also driven by the amount of automation employed. Order fulfilment can be accomplished by staff pushing carts or trolleys along aisles, picking several customer orders together. With conveyor belts and sortation systems, a customer’s order can be picked into a tote which is brought to the picker, who adds product from within their assigned aisle or area. Picking systems typically employed for unit picking include Pick to Light, Goods to Man, Voice and RF-directed picking. Perhaps the most valuable piece of material handling equipment for a dark store operation is a storage and retrieval system for the storage of picked merchandise prior to loading onto the delivery vehicle.

The importance of automated material handling equipment to the future of e-Commerce order fulfilment should not be underestimated; evidence to support
this includes the Amazon acquisition of Kiva Systems (Bloomberg, 2012).

Material Handling Equipment
Many online retailers are making extensive use of automated material handling equipment in the dark store. A significant benefit of this technology is a reduction in labour required to process the orders. This approach has been very successful for a number of dedicated web retailers. (Nguyen, Ocado: Online food is not our bread and butter, 2014). Typical types of material handling equipment used in a dark store include conveyors with pick to light, ASRS, and goods to man.

However, automated material handling equipment and the time before ROI is realised can introduce their own constraints on the operation and future growth. It requires ample planning and a detailed analysis of the merchandise characteristics and projected demand.

Tesco is one of the few profitable online grocery retailers to take a gradual approach in the implementation of material handling technology. Starting with delivery from existing stores and gradually moving to dedicated facilities with increased levels of automation. (Wulfraat, 2014)

Other Technologies
Rather than being reliant on fully automated solutions a retailer may want to look at technology that will enhance a more manual warehouse operation. Some of these technologies are described below.

RF Devices
Although common in other industries the grocery industry has not been a large adopter of the use of RF devices for the picking process. Despite the improvement in accuracy, the impact on performance has always been a deterrent to their extended use. However, the smaller physical characteristics of consumer units as opposed to full cases combined with modern device design such as wearable units may make RF technology worthy of consideration in a dark store setting.

Voice Technology
Voice technology is very mature and has been applied to grocery distribution operations for many years. It is ideally suited to the dark store picking environment.

Augmented Reality
Augmented reality systems for “pick by vision optimised picking” are currently being developed by a number of different companies. (DHL, 2014). This technology could be applied to a dark store operation; with the potential for layout and procedural changes based on the season and current market demand, virtual reality may significantly reduce the time to train employees in addition to providing productivity and accuracy improvements. The combination of virtual reality with a technology like RFID, where the exact physical location of the product is known, could be a powerful tool in managing the dark store operation.

A Flexible Warehouse Management System
Different types of technology and material handling equipment will be more appropriate for the handling of different types of product and the use of this equipment may evolve over time. Any warehouse management system used in a dark store requires the ability to manage the releasing of work in different areas of the warehouse at different times to support the varying volumes and picking rates in those different areas.

Integration of the WMS with material handling equipment and different consumer applications and channels is a requirement. The provision of standard APIs to aid this integration, the use of different integration methods (web services, message queues etc.) and standard integrations to common e-Commerce platforms are all features that an OmniChannel WMS must provide.

• Operational and application design

The ability to partially release customer orders to different areas of the warehouse to accommodate variable pick times and demand provides an invaluable aid to resource management and meeting operational constraints.

Work Planning and Order Processing
A WMS that supports a line level status on the order rather than having to wave or process an entire order at a time provides significant benefits in work generation and management in a dark store operation.

The addition of late orders, changing transportation demands (due to updated route scheduling) and other variables require the ability to dynamically change the scheduling of work in real time. Likewise, systems must ensure that all supporting activities (location replenishments, etc.) are automatically adjusted.

Pick Assignment Creation
The typical output of the processing of a wave of orders is work. The work is created based on how it will be performed, using predetermined rules. The WMS routes the work to the relevant areas of the warehouse that are managed by automated material handling equipment. This work can then be optimised based on the appropriate algorithms for the equipment used.

For areas of the warehouse that are not automated, assignment consolidation is key. The smaller quantities inherent in e-Commerce orders mean that consolidation of multiple orders into a single picking assignment is critical for efficient processing. The WMS not only needs to support the picking of consolidated orders but needs to suggest the best work consolidation process to use. The use of labour management industrial engineeredstandards and optimisation techniques will provide a very powerful tool in the grouping of order picks into consolidated assignments.

WCS pioneered the use of industrial engineered standards as a use for labour optimisation with the RF Interleaving product to ensure the efficient interleaving of forklift work in the warehouse. These same techniques should be applied to the determination of pick work assignments in the dark store environment.

Pick Sequence Optimisation
Many WMS applications use a pick sequence based on the optimal method to pick the product in the warehouse. Recognising the variable size and different types of stores serviced by modern retail distribution centres, some WMS solutions offer multiple pick sequences that are assigned to different categories of customer. However, the variable nature of the order profile for individual customer orders may require a more dynamic approach.

Again the use of optimisation and labour management standards can determine the optimum pick method for a group of consolidated assignments. Combining RTLS technology with the picking application, a solution can be
developed that tells the warehouse operator exactly where to stop and the sequence in which to pick the product. This process can be made even more efficient with the use of electric pallet jacks that do not require the operator to
be mounted on the device to move them. The time spent getting on and off the picking vehicle is not insignificant.

Goods to Man Pick Operations
It is increasingly common to use goods to man picking systems (for example Kiva in the case of Amazon). These systems have very powerful optimisation engines that will ensure the product is delivered to the picker and orders are picked in the appropriate sequence.

Bulk Picking
In a manual operation there may be types of product for which a bulk picking process is the most appropriate. Here the demand of several products over multiple customers is combined into a single pick assignment and the picked container is then out-sorted to the individual orders. This may provide some of the benefits of a goods to man type operation but without the capital expense of automation. This can also be applied to products with physical or other characteristics which make them unsuitable for automated material handling equipment.

Stockless Operations
Stockless operations may be difficult to manage in a dark store operation and there is the risk that allocating inventory that is not currently available to consumer orders may lead to disappointed customers. However, some fresh products that have reliable supply or even high value merchandise that may be stored in a national distribution centre may be suitable for a pick by line type operation.

The ability to direct late deliveries to a bulk pick area to meet immediate order needs will offer a distinct operational advantage.

Fast Mover Replenishment
Fast moving product may have dedicated locations with pallet quantities of the item. The timing of de-trashing of the product is an important point to consider. Should the product be de-trashed at point of replenishment? Clearly it will be ideal to de-trash the product as early as possible in the process. However, consideration needs to be given to product stack-ability and the stability of the de-trashed product.

Slow Mover Replenishment
Slow moving product is ideally suited to a goods to man automated pick operation. When a manual pick method is used there are a number of ways to optimise the replenishment process. It is important to ensure that the pick locations are sized appropriately for product demand and that replenishment points are set accordingly.

In areas where space is at a premium a multilevel order picker can be used to “pick” the replenishment product out of the reserve locations. The resulting multi item pallet can then be used to replenish the pick locations of the products.

Deli Item Replenishment
Deli items can be prepared prior to the generation of the wave processing. The WMS should be able to identify the requirement for these items for a specific wave, shift or day and create work for the generation of these items.
Standard units can be used. For example, ham can be prepared in advance in half-pound packets and a customer ordering one pound of ham will actually receive 2 half-pound packets.

A similar approach can be taken with custom orders such as catering or celebration cakes. The details of the custom order are provided to the appropriate dark store department where it is produced, the work must then be scheduled to meet the demand of the pick operation. Capacity planning for these areas is key, with visibility of the plan provided to the customer so the dark store has the ability to produce the product within the time constraints of the customer and the available delivery slots.

Labour Management Standards
It is now common practice in grocery distribution centres to use labour management systems combined with industrial engineered labour standards, often coupled with an incentive scheme. These systems drive productivity in the warehouse and provide management information on operational performance from a total facility level down to the individual employee level.

Successful use of labour standards is often accompanied with the implementation of an employee incentive scheme. For an incentive scheme to be effective it needs to also include input from the quality side of the business Any incentive paid to an order picker in the dark store environment should be directly related to customer satisfaction with the orders picked.

• Shipping to Customer

The management of a large number of picked orders prior to loading onto the delivery vehicle is a complex problem, and one of the first areas to consider with the use of automated material handling equipment.

Staging and Loading Product
Without the use of material handling equipment a WMS must track the different parts of orders stored in different staging locations and manage the generation of work to bring them to the loading area. In addition to container consolidation and repacking, other work activities such as QA may be managed at this time (although any auditing may be most appropriate directly after the pick operation).

RFID technology can also be a benefit in ensuring the real time location of all of the picked containers is tracked.

Tracking of Delivery and Proof of Delivery
The location tracking capabilities of the dark store need to be extended to the order once it has left the facility. The use of RTLS technology to track the progress of the delivery combined with a mechanism for alerting the customer to any delivery delays is critical. Peapod, a US-based online grocery retailer that is owned by international food retailing group Royal Ahold, provides customers with frequent updates on the progress of the order including updates when the driver arrives at the stop prior to the customer and when the driver arrives at the customer location.

The ability to capture a proof of delivery from the customer, with the opportunity to comment on the delivery service will not only help record actual delivery timing information but enable the collection of feedback on customer satisfaction.

Reusable Delivery Assets
With increasing awareness of environmental issues, the use of returnable delivery assets is increasing throughout the supply chain. There may be an opportunity for their use in the home delivery and e-Commerce markets. Significant labour savings can be realised by using a single container from the point of picking in the warehouse, putting on the conveyor and delivering all the way to the customer location with no repackaging or bagging. The use of these containers may also open up additional delivery windows. Merchandise will be protected by the container allowing delivery to the customer overnight with the customer waking up to their online grocery delivery on their doorstep. This may be appropriate for more rural markets or others where the risk of theft is low.

There are a number of home delivery grocery vendors that already use this approach. Chicago-based Oberweis have a regular home delivery service for dairy products that makes use of reusable coolers and milk bottles. (Oberweis, n.d.)

Any proof of delivery application used in this environment will need to track these reusable containers and the inventory a customer currently has. A tracking method will be required as part of the proof of delivery mechanism but buy-in from the customer will be required for any billing mechanism introduced for non-returned containers.

• Demand Management

The management of demand should be related to order management profiles and history tracked at the individual customer level.

Segregation of Demand and Sales Forecasting
Demand should be stored at the lowest possible level. However, forecasting algorithms should use total demand for all customers serviced out of the dark store to generate a single forecast. Typically, demand used for forecasting has been stored at an aggregate level: either at the warehouse level or store group level. Storing sales history at the individual consumer level and using the same rules as the order management system for order sourcing will facilitate a more dynamic and accurate model.

Information Sharing
The increased use of regular orders will reduce the reliance on inaccurate sales forecasts, adding a more definite component to the demand pattern. Sharing this information with suppliers will aid joint planning and may lead to reduced inventory positions if suppliers time deliveries closer to the demand from the facility.

Promotional and other Event Tracking
Because experimentation with new user functionality and promotional techniques is common, the customer experience will be evolving constantly as new features are implemented. The ability to closely track customer behaviour and to assess the resultant impact is critical. Data collected from the consumer UI should include navigational statistics as well as demand details. The ability to perform rapid affinity analysis of the customer basket will help measure success and predict impact on demand as changes are rolled out to other customers and channels.

Demographic Information
The analysis of e-Commerce demand may uncover demographic and other information for a specific market or region that was not known before. The use of this information in ranging stores in the demographic may prove to be an example of how the data gained from the e-Commerce operation can be used in the design of the retailer’s other channels.

• Business Intelligence

The tools companies are using today to provide business intelligence and data analytics capabilities are changing.

The introduction of products such as Tableau (www. and Spotfire ( is removing the need for significant IT involvement and investment in the development and customisation of reports and dashboards. A dark store is the ideal application of these tools. The ability to provide required data to a business analyst or industrial engineer dedicated to the performance improvement of the facility and for them to have the tools required to interpret this data quickly, producing recommendations for improvement, is going to be key to the continued success of the dark store operation.

As mentioned earlier, rapid affinity analysis of the impact of changes to the customer experience should be performed to measure the success of changes and to predict the impact on demand as they are rolled out.

Cost to Serve Analysis
A key analysis for the operation of the dark store is the calculation of a cost to serve. This is another area when the use of a labour management system can be invaluable. The data that is provided by the LMS will be valuable input into this analysis.

To achieve the benefit from any operational analysis, the WMS used must be easily and quickly configurable to meet the needs of the changing operation without requiring the need for expensive consultants and technical resource

The pace at which the grocery sector is changing is being matched by technology vendors, who are innovating and investing in key technologies that will help retailers grow and remain competitive. As retailers begin to continue to ride the wave of consumer demands, a need for a predictive and more proactive approach to technology strategies comes to the fore – investing in technology for the future is what will be the real game-changer for tomorrow’s successful grocery businesses.
Bloomberg. (2012, March 19). Amazon Acquires Kiva Systems In Second Biggest Takeover. Retrieved from com/news/articles/2012-03-19/amazon-acquires-kiva-systems-in-second-biggest-takeover

DHL. (2014). Augmented Reality Report. Retrieved from http://www. augmented_reality_report_290414.pdf

Leave a Comment