As businesses continue their march toward digitization, companies must deploy strategies that provide the right foundation to compete. A key component of success as a true digital champion is the ability to deliver the highest customer service levels at the least landed cost. And to do that, complete and accurate supply-chain data is essential. Data has become a currency that increases in value as it approaches real time, and the more it can be shared among trading partners.

What has become apparent is that data should be real time and only exist once. Using a federated master data management (MDM) approach, information can be current and consistent, rather than being duplicated in multiple formats and silos, where it becomes stale and latent (whether a data lake/warehouse is used or not). Modern technology approaches create a trusted representation of each’s/item/unit level and then can roll that fidelity into an overall real-time Network solution that optimizes planning and execution across all partners using agent-based prescriptive analytics. These networks support multi-party collaboration, planning and transaction execution in real time and help all participants to execute in the “next normal,” provided they take into account three critical capabilities.

Enabling real-time transactions with a network architecture. We’ve already seen how a network-based customer service and asset leverage model has been applied in adjacent industries. Uber, Facebook, Airbnb, and Alibaba have all deployed network-based architectures which provide multi-party capabilities. Trade happens across all parties in a network. And since product sourcing and delivery span multiple parties in the network, so planning and execution must be real time and also include multiple parties, to provide maximum asset leverage, least landed cost, and highest levels of customer service.

This level of collaboration combined with control tower level visibility, analytics, planning and execution, is core to future competitiveness. This is not surprising given that even the simplest of transactions involves multiple parties, such as a customer, brand manager, co-packer, supplier, carrier, 3PL, and distributor. 

Unfortunately, many enterprise-centric technology deployments were designed in a hub-and-spoke fashion, meaning they act as the center of the universe for that node in the network and treat their processes and data as such. This hub technology is designed to gather data in a point-to-point/spoke-to-hub fashion. It then decides what to do about changes in demand, capacity, or supply based on the hub variables in isolation, and then shares some of the stale or latent data, post processing, with some of their spokes, both inbound and outbound. As a result, in a typical trade network, it can create over 20 store and forward type processing actions across trading partners, upstream and downstream, wasting time, manpower, and assets. Worst of all, it affects customer service levels because parties are not aligned as a coordinated network serving the end consumer.

Establishing a single, trusted view. If data is your currency, then multiple ERP instances are similar to a federation of entities where each source acts like its own unique legal tender. Data becomes trapped in your ERP stove-pipe instances and then is typically shared in a hub-and-spoke fashion in one-to-one trade relationships with network partners. Needless to say, this is a sub-optional approach, because even if you were to export data to a data warehouse or data lake you are creating latency and staleness in the data. This devalues the currency in terms of decision making across network partners.

As a result, organizations are adopting solutions that include federated master data management. Using this approach, network trading partners can opt into the network and share both their master and operating data with other trading partners. Based on a secure permissions framework across the network, data exists only once and is federated to trading partners based on the permissions granted. Given that the data is not copied or duplicated between echelons, tiers, or nodes across the network, it is by definition real time and readily available to optimize asset leverage, customer service, and least landed cost. 

Supporting actionable and autonomous prescriptive analytics. Given the multiparty nature of network-based trade relationships, the final criterion is the ability to model the entire end-to-end supply-chain network in order to correctly analyze and take action on problem resolution and opportunity creation. Since the problems or opportunities exposed by the analytics could manifest in strategic, tactical, or operational timeframes, the foundation must be seamless across these time horizons. It also must offer services, algorithms, and analysis that run across the network representation in real time, whether used to solve problems predicted to happen in six months or during a delivery scheduled for later this afternoon. The good news is if your approach includes both criteria one and two, this foundation is already in place.

The end-to-end, real-time supply network platform enables the ability to test out new supply-chain policies, network resiliency, the feasibility of strategic or tactical plans, activate alternate parts or suppliers, modify modes of transportation, or even add additional shifts at a plant.

Since there are many ways to solve problems related to demand, supply, logistics, and fulfillment in a network, it is important that the analytics workbenches have real time access to every material variable possible. Traditional systems typically only give you one way to solve a problem, due to static lead times and stale data. In contrast, an AI-based analytic workbench is a prescriptive environment where organizations will be presented with the top three or four solutions that best meet their targets. It can also support machine learning which is just a better way to predict outcomes and is extremely valuable when recommending prescriptive actions. Not only that, machine learning improves its predictions over time as you provide more data.  

These capabilities will provide the platform and architecture necessary to enable this foundation and provide the capabilities needed to compete moving forward.

Joe Bellini is chief operating officer at One Network Enterprises, provider of AI business network software.