The Autonomous Supply Chain: Defining the New World of Planning

I like this picture. It is an artistic view of robotics at work. As we approach Supply Chain 2030, and the autonomous supply chain, technologies and people will change dramatically. Factories will be run by one to two people, perhaps in their PJs from their living rooms. New positions, like robotic supervisors, will evolve and companies will need to adapt. With the evolution of the autonomous supply chain we are witnessing the transformation of labor. The processes of the future will be more automated. To adapt, the workforce will need higher skill levels in math, computer science, and systems thinking. What we have defined in the past will not be equal to the challenge of the future. The biggest challenge will be changing the paradigm.

A Historic View

With the evolution of supply chain planning, companies added headcount. The promise of the technology was greater than the cost. As a result we have a dilemma. The promise of supply chain planning was overhyped resulting in the rise and fall of the technology market. The planners remain, the technology market consolidated; yet, only a few companies realized the overhyped value proposition.

In our research, those that realized the value of planning, in this first evolution of planing, had five characteristics:

  1. Supply Chain Planning Career Paths. The companies that underachieved used supply chain planning as a flow-through position. At these companies, supply chain planners came and went in two to three year cycles. In the more mature companies, supply chain planning team career paths were established. The role was clearly defined and centers of excellence established.
  2. Clear Definition of Supply Chain Excellence. Planning focuses on the important. Executional and transactional systems are more focused on the urgent. In the role of planning, the teams define the trade-offs between source, make and deliver. While the traditional supply chain processes focus on functional excellence, the planning teams must traverse the functional silos to define the trade-offs. For success, there needs to be a clear definition of supply chain excellence that is supported by the executive team. 
  3. Planning System Technology Implementation Methodology. Traditional projects focus on quick implementations and a well-defined project plan. In contrast, planning projects take time. The planning engines require fine-tuning. Inputs need cleaning and models must be tested and outputs validated. Teams need time to adapt to learn new skills. A mistake that companies make is not taking time to test and validate the models.  
  4. Governance. The first planning systems were defined for regional supply chain teams. A global deployment is very different one than a regional one. Within a regional supply chain planning there is more focus on operational and executional planning. In a global planning system, there is more focus on tactical and strategic planning. Within global planning, there is a greater need for role clarity of global planning, and the roles of the regional and business planning systems. They need to work together. As companies become larger through M&A, role definition becomes more important. In a regional driven planning system, like J&J, each region operates autonomously. In the case of a company like Colgate where there are cross-company shipments between regions, there is a greater need for a global and regional planning. In a larger company, like BASF, there is a greater need for corporate planning, especially in the area of logistics and procurement. In all cases there is a need for well-defined time horizons. 
  5. Well-defined Time Horizons. A successful planning system has well-defined time horizons. In Figure 1, I share a sample definition from a client. In each company, there is a need for a short-term plan, a medium-term plan, and a long-term plan. The short-term plan is focused on replenishment. A medium-term plan helps to define asset strategies and material sourcing plans, and a long-term plan defines the network (the roles of each factory and distribution center in fulfilling the supply chain requirements). Executional planning is rationalization of orders to inventory within the order cycle and slush period. Operational planning is the short-term coordination between factories and distribution centers to coordinate transportation routing plans, Distribution Planning requirements (DRP) and Material Requirements (MRP) with production scheduling. In contrast, tactical planning places manufacturing load (aggregate demand) on factories and distribution centers and helps to align the forecast to manufacturing and procurement strategies. In well-defined planning systems, there are clear definitions of the planning horizons with clear hand-offs between global, regional and business teams. No two companies are alike. What drives the difference in successful planning is clear definition. (The colors in Figure 1 are levels of maturity. We have never worked with a company that is mature in all areas of planning.)

Figure 1. Planning Time Horizons

In my research, I find that no two companies' supply chain planning deployments are alike. In the early deployments, team-after-team brainstormed "as is" and "to be" processes. Because supply chain planning was new, it was hard to conceive a "to be" process. As a result, and unfortunately, in the first evolution of technology deployment, companies largely made their current processes more efficient, but failed to actualize the promise. Most companies today bought supply chain planning systems, but do not use the optimizers. Instead, they plan by spreadsheets. We have many Excel ghettos where we should have supply chain planning systems.

At the top of the hype cycle in 2000-2007, company resources were thin and many consultants implemented the technologies badly. Today there is a maturing of the second generation of planning technologies and the evolution of boutique consultants with deep expertise. However, the drag on labor productivity continued. (As shown in Figure 2.)

Figure 2. A Range of Planner Productivity by Revenue

A Future Look

We are entering an age of virtual assistants. Whether it is Siri on our iPhones, Cortana on our Microsoft devices, or Alexa on Amazon's echodot, the use of voice activated assistance is becoming mainstream. It has the potential to replace clunky supply chain planning user interfaces. During this month, I witnessed demo(s) by technologists featuring the use of a digital assistant to interface with cognitive computing engines to redefine planning.

In 2014 we predicted the third-act of technology for supply chain planning. The forecast was the redefinition of supply chain planning by 2019. In parallel, the prediction was that companies with a focus on IT standardization would lag the market. It is happening. In the last week I witnessed two demos of Alexa-like interface for cognitive-computing based planning. Today's planning systems are built using descriptive and predictive analytics. Cognitive computing along with prescriptive analytics drive greater insights.

Figure 3. Continuum of Analytic Engines for Planning

What could planning processes look like? Consider the shift in the Table in Figure 4:

Figure 4. Rise of Autonomous Supply Chains and the Redesign of Supply Chain Planning

I know it will not be this easy or dramatic, but I think that the future of planning will be less stressful and more productive for supply chain planning if we are open to the outcomes and don't limit our thinking by traditional thinking. Let's face it. The job is tough today, and the work can be more productive and easier in the future.

What do you think? I say, "Let's welcome Alexa to the supply chain planning team! And if Siri or Cortana perform better, I vote to welcome them as well." We need to redefine planning through new analytics to touch data less and drive greater insights.

Rating: 
0
No votes yet

Share this post

Comments (4)

  • I really like this article.  It articulates a lot of what I have been thinking.  I would be very interested to learn what companies are working on this.

    Apr 06, 2017
  • thanks Ed. Most companies are investing in the present. We are hoping to jump start the industry. Any others have anything to share.

    Apr 06, 2017
  • Lora

    A nice piece of writing... sort of a start on a future use case.    However, much easier to invision -  then to actually make feasible.   In past 12 months there seems to be a massive PR trend in technology providers of all types to predict machine learning combined with big data will replace humans for everything from driving a car to managing a supply chain.

    If you study history of IT and algorithms and decision support this same prediction happens every few years and results are like watching for the big 4th of July fireworks show but all you get is a tiny smoke bomb.    

    I would put my money on a focus on well orchestrated, well engineered process, hardware, software solutions that will save people time not replace them.   Some examples I would love to see

    - simple tools that use rules (AI) to cleanse execution data that feeds planning systems.   Takes in multple feeds for same knowledge (e.g. intransits) screens to select one that is most likely to be correct

    - Webservices, Jdocs, integration sockets that easily let participants in supply chain network reveal inventory, planning bills, mfgr. capacity, transport lead times to private and public participants.  

    - Simulation technology for S&OP that helps business leaders understand weakness of GAAP accounting and the physics of flow of finance and how S&OP projects and simulates future income statement / Balance sheet vs. working on purely historical data

    - Algorithms that leverage simple S&OP data relationship to enable roll-up of components to finished product / projects

    - Transportation lead time clearing houses that let inventory managers (yes inventory managers) substitute lead time using transport for inventory buffers.     Lets business see through freight forwarding / international transport for lead time management  of inventory not pure transport

    Maybe these efforts can benefit from "Alexa" style conversational user interfaces but I don't think that is the goal.   

    Keep up the great work

    Jon Kirkegaard

    www.sopbook.com

       

    Apr 10, 2017
  • Lora

    A nice piece of writing... sort of a start on a future use case.    However, much easier to invision -  then to actually make feasible.   In past 12 months there seems to be a massive PR trend in technology providers of all types to predict machine learning combined with big data will replace humans for everything from driving a car to managing a supply chain.

    If you study history of IT and algorithms and decision support this same prediction happens every few years and results are like watching for the big 4th of July fireworks show but all you get is a tiny smoke bomb.    

    I would put my money on a focus on well orchestrated, well engineered process, hardware, software solutions that will save people time not replace them.   Some examples I would love to see

    - simple tools that use rules (AI) to cleanse execution data that feeds planning systems.   Takes in multple feeds for same knowledge (e.g. intransits) screens to select one that is most likely to be correct

    - Webservices, Jdocs, integration sockets that easily let participants in supply chain network reveal inventory, planning bills, mfgr. capacity, transport lead times to private and public participants.  

    - Simulation technology for S&OP that helps business leaders understand weakness of GAAP accounting and the physics of flow of finance and how S&OP projects and simulates future income statement / Balance sheet vs. working on purely historical data

    - Algorithms that leverage simple S&OP data relationship to enable roll-up of components to finished product / projects

    - Transportation lead time clearing houses that let inventory managers (yes inventory managers) substitute lead time using transport for inventory buffers.     Lets business see through freight forwarding / international transport for lead time management  of inventory not pure transport

    Maybe these efforts can benefit from "Alexa" style conversational user interfaces but I don't think that is the goal.   

    Keep up the great work

    Jon Kirkegaard

    www.sopbook.com

       

    Apr 10, 2017