Most supply chain professionals and planners think of finished goods when they hear attributes. But attributes are present in every part of the supply chain from suppliers to raw materials, to machinery, products and distribution centers. Attributes define the specific properties of each object. Output quality produced by a machine, a qualified supplier for a given customer, surface tension or shade of a WIP material or likelihood of late delivery from a supplier are all example of attributes of objects in the supply chain. As you can see attributes are not just in finished goods. More importantly, attributes are not static. They change as the supply chain changes. For example, when you acquire other companies into your supply chain, when you get new subcontractors, when you add or delete products, when new customers are added and delivery policies change. To this end, supply chain model needs to be very flexible and have the ability to change and grow as your organization changes and grows. Attribute are absolutely critical to every supply chain planning process. They help to maintain the supply chain in a way that is a true representation of the supply chain. They also help to define your changing business processes without having to use developers to make changes to the supply chain planning system. Below are highlights of the functions and benefits that attributes enable.
Supply chain model needs to be very flexible and have the ability to change and grow as your organization changes and grows.
As your supply chain changes so does your business rules and your priorities. Attributes have the ability to define the new environment without having to re-code the supply chain planning software. Using Boolean expressions, they help to define your updated model of the supply chain by introducing the properties of the new changes in business rules, procedures or policies. Thus, the supply chain planning application keeps molding itself into your new environment.
Attributes because of their nature act as user-defined and dynamic constraints. For example, if only a new sub-contractor is to be used or a new sourcing strategy is to be deployed, then attributes define this as a new constraint. Of course, just having a constraint is not good enough; in addition, one needs to have the architecture of the system designed in a way that takes into account attributes/constraints into account while searching for an optimal plan. As such, attribute-based planning forms an AI expert system that is constantly trained by a user and therefore improves its performance over time as needed.
Scalability and SKU Reduction
Attributes allow defining generic bills of material, generic routings, generic finished goods as well as generic intermediate goods without having to define a unique bill of material for every little variation in the product or production process. One can define a generic product and then add attributes as needed. For example, if a customer demands the use of a certain qualified supplier for the same product, then there is no need for a whole new BOM for this specific customer. The same BOM is used with an added attribute. They also help pegging the right WIP to the final order. For example, depending on the shade or texture of a fabric, the WIP lot may no longer meet the original properties of the final order and therefore it is pegged to some other order.
As a result, there is no explosion of SKUs. Thus, a lot less memory is needed to define all the different variations of BOMs and routings as well as SKUs. It also makes maintenance of the system a lot easier by simply introducing new properties to the existing ones. Deploying so much less memory and generic definition of SKU’s and other objects, the search process for optimization becomes faster by orders of magnitude!
Digitalization is the process of representing your supply chain by a Digital Mirror©, or as Gartner calls it a Digital Twin. This implies very accurate and detailed representation of the supply chain. Attributes do precisely that. They allow every detail of all the objects in the supply chain to be represented forming and accurate model. In the absence of an accurate model it would be impossible to mimic the supply chain behavior and make the right decisions. Think of an airplane cockpit simulator. It has all the detailed and necessary features and nuances of an actual airplane in operation. In addition, all the external forces such as weather changes, wind directions and instrument malfunctions can be accurately simulated. For a supply chain to get to that level of accuracy, attributes are a must.
Enable Deep Learning
Neural networks take attributes of objects and try to classify them according to their values. To this end, once the attributes of products (new or old) as well as policies are given as input to the network, then they can be trained to make inferences as to what policy is suitable for what product for best forecasts. Likewise, they are used for clustering groups of suppliers, customers, processes amongst others such that exceptions and abnormalities in the supply chain can be identified.