alphaspirit - Fotolia
How do I choose an effective supply chain analytics system?
Supply chain analytics promises rich insight into operations and better decision making. Here are a host of considerations when thinking about what you need.
Supply chains today are going through a radical transformation powered not just by cloud (of which supply chain applications were an early beneficiary of) but grid computing, more powerful servers, elastic cloud and new "math" -- code that is being written to add more smarts to the analysis. So, if you're researching supply chain analytics, you're likely facing plenty of questions: What are the main components of an effective supply chain analytics system? Should I get a dedicated tool from a supply chain provider, or should I get a generic analytics or business intelligence (BI) tool? What are some of key elements today? What might be the criteria of one choice versus another?
To answer those questions, we need to think about supply chain analytics in terms of two domains -- operational (daily and possibly real time) and strategic (trend analysis, discovery and reporting).
Operational analytics can sit on top of only one module, such as demand planning, manufacturing scheduling, transportation and so on. But the more interesting play today leverages analytics that have data from multiple nodes, such as work centers, warehouse locations, and in-motion inventory on a truck or ship. Products available today are working towards more cross-functional and inter-enterprise views, with the goal of overcoming siloed views to give a true picture of the supply chain as a cross-functional endeavor.
So, what might operational analytics look like? Here are some examples:
Inventory and logistics decision making. Leveraging grid computing and multinode data across a supply chain, users can evaluate and make inventory stocking decisions based on demand, sometimes called a demand signal repository. Transportation managers analyze routing and rating options across nodes or carriers and evaluate spend. What's new here is the capturing of constant updates of data from the nodes as conditions change rather than accessing static data. In this way, data becomes operationally useful far sooner. A prominent provider creating systems for this arena is One Network, which has rich multinode, multienterprise networks across complex supply chains.
Understanding high-level events. Complex event processing (CEP) is finally being used in supply chain operations. Vendors producing interesting products in this area include Savi Technologies, TransVoyant, GT Nexus (which embeds TransVoyant) and, of course, SAP. Oracle has a CEP product, but I have not seen cases yet where this has been applied to the supply chain. CEP looks at a series of events and influences. It then tries to understand causes and predict outcomes, in contrast to traditional alert systems, which just tell you something happened, but not why, how or what to do about the event. In this way, CEPs help catalyze better outcomes.
Risk management. There are several impressive risk management products that play a role in the real-time domain to spot risks, analyze causes and provide potential alternative solutions. Some interesting companies who target this space are Resilinc, Exostar and IDV.
From a strategic and reporting perspective there are two paths forward:
Analytics within a more large-scale management system. Much of the analytics within a function, such as supply chain, is based on known sets of data that are within the supply chain management system. So, when thinking of the supply chain analytics system, it often makes sense to go with the same vendor you're using (or are going to use) for supply chain management. Such products have the use case libraries and often have embedded a commercial BI product into the suite. Supply chain is a large market so there are a number of choices, such as suite providers like JDA, Logility, or cloud providers like One Network, GT Nexus/Infor. There are also more functional approaches, such as transportation management systems like Descartes or Mercury Gate or manufacturing-centric systems such as PTC, Siemens, GE, Dassault and Autodesk. SAP, of course, is a consideration, but it might better fit in the BI provider category, since SAP's goal is to sell HANA on top of various modules of ERP or supply chain apps.
One interesting development here that I like is JDA's partnership with Google, which aims to provide richer search and analytics, which will in turn look at behavior and trends to help plot new ways to look at markets.
A business intelligence-focused product. A BI provider might make sense for environments that use a multitude of management systems -- a best-of-breed approach. With this choice, users must take responsibility for building data warehouses and enabling integration, although analytics providers have built tools to help manage the pathways between systems. Many enterprises turn to a BI product that has presence in the supply chain, such as SAS, Tableau, IBM and SAP's portfolio , as well as more elastic options such as Microsoft Azure for big data analytics.
Counsel and caution
Supply chain analytics is an area that is changing fast. Older analytics were largely about accumulating data in warehouses and are user dependent. That is, users stare at mountains of data and build their own queries and reports. Today, we have libraries of advanced queries for both strategic and operation applications.
I am really bullish on the new analytics in the operational space with their focus on real-time data and smarter optimization and analytics. Users should take some time to really understand the power of these and how they will use them. New supply chain analytics tools are designed to give companies more predictive, more real-time response, so operations can get the jump to solve problems more quickly. Even more important, the analytics today are not just about speed, they also supply richer insights and learn over time what the best options are. They access and use new types of data that are fed into the supply chain applications -- from weather, traffic, sensors and so on -- and blend this with more traditional data sources. That type of complex and more immediate analysis and insight is where modern systems are going. Often the business is procuring these systems on their own without IT's involvement. So, IT should partner with the business user and understand how they fit in the overall portfolio of tools.
