Companies are faced with a series of critical decisions when developing strategies to deploy a new Warehouse Management System (WMS). This is especially true for companies with multiple distribution centers. In addition to selecting the most suitable software system for the organization, these companies must determine how to best deploy these systems across their distribution centers.
Some key decisions include:
For this discussion, we will focus on the last item listed—centralized vs decentralized deployment models. This is often a decision that is relegated to the IT team and accepted without consideration to other business areas. But this decision has an enormous impact on virtually all areas of the business and should be thoroughly analyzed. Let’s take a look at both models.
Decentralized WMS Deployment Model
Decentralized WMS deployments treat each site separately. Each distribution center has its own instance of the software being deployed. Each site can be configured and/or modified independent of each other.
It can be argued that the decentralized WMS deployment model is a traditional approach for deploying a WMS across multiple DCs. This is mostly due to the fact that warehouse management systems were initially physically implemented at each warehouse. This is before the days of centrally located data centers or cloud deployment options. These ultra-on-premise deployments required a computer room at each facility with servers to run the WMS.
As is the case with all things, this has evolved over time. Now it is typical to see on-premise deployments that are centrally located in a data center often hundreds of miles from the distribution center or even cloud deployments that are hosted by the software vendor. However, this deployment model remains.
A Centralized WMS Deployment Model
The key distinction of a centralized WMS deployment model is that all DCs are deployed on a single instance. Meaning the software is implemented once at either an on-premise data center or in a cloud hosted environment. The software is then deployed to individual DCs through the use of software enabled virtual warehouses.
These virtual warehouses essentially replicate multiple installs of the software. From a warehouse user perspective there is no difference – they can only log in, view data and execute tasks for the virtual warehouse representing their warehouse.
There are several benefits and challenges associated with both approaches that will be discussed next.
Benefits and Challenges Associated with Each Model
First, let’s consider the traditional decentralized model. A few of the key benefits and challenges can be summed up as follows:
+ Customized Solution – This is especially relevant for organizations that require significantly different operational processes due to SKU handling requirements and/or available levels of automation. A decentralized model allows for supporting DC specific operational processes through distinctive configurations and customizations at each site.
+ Reduces Risk – This is the reason many companies decide on a decentralized approach. Even though redundancy and contingency plans are common for large scale deployments, software systems are still subject to unexpected downtime and outages. Unexpected downtime and outages are due to myriad of potential issues, from the unintended consequence of deploying a software patch/upgrade to an internet service provider outage. These issues can typically be contained to single facility in a decentralized model.
– WMS Consistency – Generally speaking, keeping the WMS consistent (same version, release, patches, etc.) across sites is overly complicated in a decentralized model. The result is deployed WMSs that differ from facility to facility in terms of versions and release levels.
– Troubleshooting – Troubleshooting and resolving integration issues is complicated. In a decentralized model, each location has its own set of interfaces that have to be maintained. Resolving issues often results in fixes by site instead of the entire network.
Now, let’s look into some of the key benefits and challenges for the centralized model:
+ Software Support – Deploying patches, handling configuration changes and managing upgrades is streamlined in a centralized model. IT teams make these changes and can update the entire network once. Integration issues are much easier to troubleshoot as a result of their only being one set of interfaces to maintain.
+ Network Visibility – A centralized model allows for a standard consistent deployment at all sites. This enables improved visibility/reporting across warehouses, channels and business divisions. This approach tends to be ideal for multi-site, multi-channel support of items across an entire enterprise. Other benefits can be realized as well, such as automatically generated advance shipment notifications (ASNs) for warehouse transfers.
– Additional Testing / Development – Because of the potential adverse risk exposure to a company’s network of DCs, the net testing and development effort for a centralized approach typically exceeds that of a decentralized model. This is also due to the need to standardize processes across sites to fit a single instance deployment. The typical scenario is that at least some of the DCs will have to adapt to new processes to support an agreed upon standard design.
The key to determining the preferred approach for your organization is to identify your goals for organizational focus, internal skill sets, benefits and risks. Whether it is focusing on a standardized approach to support rapid implementations with a centralized model or a highly flexible approach which typically involves a decentralized approach, the objective for your organization is the primary driver in determining the preferred deployment model.
—Howard Turner, St. Onge Company