As market conditions become more dynamic and supply chains become more complex, the need for faster, more intelligent planning and decision-making has never been more real. Still, many global companies express anguish at the low quality of basic data, which makes it challenging to meet the basic requirements for a good planning process (like getting an accurate inventory picture). These companies have invested in ERP systems to solve the data problem. So what gives?
The supply chains of most organizations have become increasingly outsourced as what were once third parties now handle internal operations, contract manufacturers manage manufacturing, distribution by logistics partners, etc. Procurement of parts and materials follows a complex model, where some parts of the bill of material are negotiated and sourced by your procurement organizations. In contrast, others may be managed by the contract manufacturers.
Companies in this mode experience the most significant challenges in data visibility. These challenges cannot be addressed by adding another ERP instance. To resolve this issue, these companies must have a planning system with an overlay data management layer with the capabilities listed below.
The data management layer of a next-generation planning system
1. Connected graph-network model
Because graph modeling gives business users more control than traditional relational modeling, a next-generation data management layer must leverage graph-modeling capabilities and represent supply chain nodes across the end-to-end business at the appropriate level of detail.
For example, a brand manufacturer needs to model: (1) retail and distribution partners, (2) internal supply chain operations, (3) contract manufacturers and third-party logistics providers, (4) the relationships between these nodes, and (5) the constraints that govern the operation of the supply chain.
And as changes take place in the business, the brand manufacturer needs the ability to add and change nodes, relationships, and constraints easily.
2. Transaction connectors for outsourced nodes
A modern data management layer must be able to interface with all transaction and data systems from each outsourced supply chain partner. It must have the flexibility to set up the process for getting transaction updates for orders, inventory, and other transaction data in real-time or high-frequency batch processes.
3. Computational models for creating intelligent visibility
At any given point, a planning process and system must have a reliable picture of inventory and order status (product, procurement, distribution…) across the network. Using the connected graph network model and the transaction data, the data management layer must be able to compute a reliable picture of (1) Point-of-Sale (POS) data, (2) channel inventory data, and (3) inventory and orders across the internal and outsourced supply chain nodes. The system should be smart enough to extrapolate and create a best-guess projection when transaction updates do not happen as expected. The system then feeds that information to the planning process on demand.
Picture this
Imagine receiving a transaction update from a contract manufacturer’s system, which indicates material has shipped from the factory to a 3rd-party logistics hub. In that scenario, the proposed system should utilize the connected network model to recompute the following:
- Inventory on hand at the manufacturer
- Inventory in transit; and scheduled receipt states for the hub, based on lead times
-
Finally, when the logistics hub provides a transaction about material received, the system should show if it was, or was not, obtained according to plan!
4. Smart alerts based on machine learning
There are millions of nodes and transactions across the network, and the people managing the data need support to ensure the data are consistent with reality. The ideal data management layer can uncover trends and show planners when the reality is not reflected in the model before issues arise.
For example, let’s say the model assumed the lead time between a specific factory and a hub was five days, but the sum of transaction data over the last month shows that the lead time is seven days for 99% of the month. The planning system of record should disseminate smart alerts—that is, alerts based on machine learning—to inform planners of the exception, enabling them to respond appropriately and change the master data.
5. Unstructured data and planning policy data need a system of record
In every company, too much master and planning policy data live in spreadsheets, meeting notes, emails, and the planners’ minds. For example, planners know which products are cannibalized when a particular product is promoted, but this knowledge is not logged, shared, or maintained in a way that others can use it. Unstructured data captured in meeting minutes and emails about prior discussions/negotiations with a supplier are not maintained and archived, resulting in a loss of knowledge when a planner, or procurement officer, moves on from their job.
Like transaction data, unstructured data and planning policy data require a system of record. Next-generation data management layers must address this need.
6. Collaborative data management
Different roles across the extended enterprise are responsible for managing and updating data. Appropriate master data updates must be provided by outsourced manufacturers, suppliers, and logistics providers. The data management layer should access internal and external roles, a simple Excel or web interface to update master and plan updates, and a solid data governance framework to ensure compliance.
The impact of o9 Solutions
At o9 Solutions, we considered these needs first principles when solving the data challenges for following generation planning problems. Our next-generation platform addresses these challenges. You may request a demonstration online or contact us directly for more information.
