Why the Forecast for Supply Chain Planning has been Only Partly Cloudy
Cloud-based supply chain planning is gaining traction. Each year more users opt for cloud-based systems and there is a lot of support in the analyst community. But cloud-based supply chain planning has not proliferated nearly as fast as many have predicted and “public cloud” (versus “private cloud”) implementations are still a small minority, leaving a lot of head scratching about why cloud hasn’t been as widely adopted as some expected.
I have a theory as to why we are seeing some market skepticism about supply chain planning in the cloud, and it hinges on the distinction between Systems of Record versus Systems of Engagement.
Systems of Record (SORs) are operational systems designed to manage business processes such as finance, HR, order processing and product life cycle management. They primarily handle internal data. Supply chain management is a system of record. These are core mission-critical applications, and companies are reluctant to part ways with their on-premises operation—executing and managing the processes, vaulting the master data, and capturing and archiving the data generated during the transactions.
Moreover larger planning applications that deal with large amounts of data, advanced analytics and optimization place huge demands on computing, database and data extraction resources. For instance, imagine trying to run a stock exchange on a public cloud rather than on dedicated servers. As Geoffrey Moore of Crossing the Chasm fame explains, “Traditional mission-critical systems of record applications will probably remain in legacy form on premise for many years, rather as mainframe computers have remained part of every larger organization’s IT landscape since they were first forecast to be replaced thirty-five years ago when the PC and later the Unix server arrived on the scene.”
Systems of Engagement (SOEs) are systems that depend on collaboration, gathering and orchestrating information outside the enterprise. “Here enterprises are looking to empower their tactical and practical business users to make more consequential decisions during critical moments of engagement,” Moore says.
Functions such as demand and supply collaboration that pull in external data are SOEs that are naturals for the cloud. Demand or supply-side sensing require capturing signals from customers, supply chain partners or other enterprises across the value chain. So for on-premise versus cloud-based architecture, there is a natural distinction between SORs and SOEs. SOEs are particularly well suited to the cloud. SORs, not as much.
On the plus side, cloud-based software is initially cheaper and gets continuously upgraded. It offers SOEs the elasticity and flexibility to scale to a large number of users, often geographically disparate. And there may be more flexibility to alter course by business requirements and improve the user and customer experience.
Issues to be considered with a cloud-based approach can be particularly applicable to the types of larger SOR applications I described above. Gartner says users are finding that cloud solutions require similar amounts of training and cost to support business models and target workflows. “Merely moving capabilities into the cloud doesn’t translate into cost savings or business and customer benefits. In some cases, doing so actually increases cost and process complexity,” Gartner says in Lessons Learned from Cloud in Manufacturing Industries. Other issues to consider:
Data stewardship – “How will the SaaS [Software as a Service] vendor secure this data? How will it use this data? If you decide to switch to another vendor or bring the application on-premises, how much of the data will the vendor hand over and in what format?” Gartner asks.
Difficulty switching among competing cloud providers – “The problem is that when you buy cloud software, you’re going to be stuck with it for a long time, according to R “Ray” Wang, principal analyst at Constellation Research, Inc. in Cloud-Computing Promises Fall Short (The Wall Street Journal, October 2015).
Scalability and performance may be another issue for more powerful optimization and analytics (such as processing millions of SKU/Locations in a few hours) requirements. Gartner also asks, “Will a new release impact performance of the application that wasn’t anticipated or able to be tested before it was moved into production?”
Lora Cecere, founder and CEO of Supply Chain Insights said in a post at supplychainshaman.com, “Cloud solutions are being hyped today, but avoid the temptation to buy cloud for the sake of cloud.” Gartner suggests firms identify a business case for cloud that considers the costs, benefits, risks, and timelines.
And I would add, focus on SOEs before SORs.