Technologies in Supply Chains, jobs and assumptions

Roger OakdenLogistics Management, Logistics Planning, Procurement, Supply Chains & Supply Networks2 Comments

business analysis

Technologies and acceptance

Articles by commentators have forecast that technologies will result in the loss of millions of jobs within a few years, including in supply chain functions. But are these forecast accurate?

A blogpost from 2017 illustrates that forecasting job losses due to technologies is not new. Now, a recent forecast stated that jobs involved with materials management (and by extension, supply chains) would disappear by 2025; other forecasts have noted that more than 80 percent of supply chain planning tasks will be automated in the same five years.

There appears to be an assumption that artificial intelligence (AI) – better described as machine learning (ML), is at the heart of these forecasts. But how capable and accepted is this technology within supply chains? The 2020 Hype Cycle for Supply Chains written by the research firm Gartner notes that AI is at the first of five stages in its hype cycle (Innovation Trigger) and will not have general acceptance in supply chains for more than 10 years. The DHL 2020 Logistics Trends report is more hopeful, forecasting that AI will be accepted in supply chains in about 5 years.

However, both reports do not identify the areas within supply chains where the use of AI will be most common. Physical operations, such as in a factory warehouse or distribution centre, are more likely than in functions that require thinking input, such as in operations planning.

Challenges in Supply Chain Planning

The research firm Gartner issued a report in 2019 which identified the ‘4 evils of supply chain planning’ as uncertainty, bias, data and model:

  • uncertainty: due to an organisation’s supply network being a non-linear ‘complex adaptable system’ that is:
    • not controlled by any organisation,
    • influenced by differing market demands
    • affected by external and internal complexity; variability (due to random events and statistical fluctuations) and external and internal capacity constraints
  • bias: the interpretation by different parties in an organisation of supply chain data. Examples are: reasons for increases and decreases in sales and delivery and the pattern of orders and deliveries for similar SKUs following an ‘out of stock’ situation for one SKU at a warehouse or major customer
  • data: clean, accurate and timely input data is a continuing challenge for planners, including missing, not considered, inaccurate and mismatched data from various sources, which are not controlled by planning
  • model: Requires a ‘flow’ model containing a cost/product mix/volume “what-if’ trade-off capability; an emphasise ‘effectiveness’ rather than ‘efficiency’
    • Expectation that Operations Planning should be precise when the forecast inputs actually indicate the probability of demand; therefore operations should be planned within a range of probabilities
    • Requires a formal segmentation of order types, customers, suppliers and inventory
    • Lack of a supply network view of the organisation provides variable alignment between accounts, demand (sales), inventory, procurement (supply), manufacturing, engineering/technical, distribution and transport
    • Lack of visibility across a geographically complex supply network, which incorporates outsourced functions and contracted tasks.
    • High asset utilisation and increased SKU has resulted in lumpier demand and longer response time, which affects planning
    • Buffers of inventory and capacity through supply chains is typically not a recognised planning technique by management
    • Emphasis by management and IT on ‘integrated’ systems rather than ‘connected’ applications

Concurrent Planning

These fundamental supply chain design challenges influence how automated planning applications are structured. In addition, is the understanding and logic concerning elements of planning supply chains. An example is Concurrent Planning, for which there is at least three different definitions:

  1. Visibility of changes to the plan, with minimal delay, for all other planners in the team when a change is made by a planner
  2. Reduce the planning horizons of strategic, tactical and operational planning and execution scheduling
  3. Plan inventory, supply, orders & forecasts and capacity at the same time

The third definition has existed for many years, so which definition is ‘correct’? This illustrates that without commonly accepted definitions, it is difficult to comprehend the topic being discussed.

Based on the third definition, why is Concurrent Planning not the standard planning practice for supply chains? At the birth of MRP applications, there was insufficient computing power available in most computer systems to concurrently plan. Therefore, the materials plan provided a total of capacity required by work centre, based on the assumption of infinite capacity. Schedulers were then required to manipulate the orders to fit the actual (finite) capacity.

Over time, academics wrote papers about structuring finite capacity models and later about linking the materials and finite capacity manufacturing plans. Subsequently, writing the required programming code proved to be a challenge (one application took about 10 years of iterations prior to release), so commercial systems containing concurrent planning were slow to arrive.

As the concept of supply chains gained prominence, so the requirements of concurrent planning became more complex, due to the different capacities to be considered at each node and link through a supply chain. But again, definitions are not standard; for example, software suppliers often promote their capability of planning ‘end to end’ supply chains. On review, they are typically referring to the core supply chain (tier 1 customers back to tier 1 suppliers); not the extended supply chain (customer’s customers through to supplier’s suppliers).

Synchronised Planning

Concurrent Planning is now considered to be a part of Synchronised Planning. This is when a continuous flow of data comes from an organisation’s supply network to enable an accurate plan for supply, imports, production and distribution, to match actual demand. In the situation of an interconnected digital supply network (DSN), the planning information then travels across the nodes and links of the organisation’s supply network, enabling suppliers, contract manufacturers and logistics service providers to better plan and provide resources when and where they are needed.

If you think this description is for sometime in the future, it is. Organisations in supply chains and software application suppliers are not yet ready for this level of change, but it could occur in steps over the next 10 to 20 years. A competitive advantage could be gained by an enterprise being an early adopter, or they could lose a lot of money through implementing ‘flaky’ software, not understanding the fundamentals of supply networks and not preparing their organisation for change.

In closing, the final comment in the 2017 blog stated: An approach to predictions about the future of technologies and jobs should be to ‘read, hear and see, then verify’. Identify the ‘facts’ that authors and commentators use, then check their source. Typically you will find there is a large amount of ‘spin’ and ‘puff’, but few hard facts.

Share This Page

About the Author

Roger Oakden

LinkedIn Twitter Facebook

With my background as a practitioner, consultant and educator, I am uniquely qualified to provide practical learning in supply chains and logistics. I have co-authored a book on these subjects, published by McGraw-Hill. As the program Manager at RMIT University in Melbourne, Australia, I developed and presented the largest supply chain post-graduate program in the Asia Pacific region, with centres in Melbourne, Singapore and Hong Kong. Read More...

2 Comments on “Technologies in Supply Chains, jobs and assumptions”

  1. Roger,

    Another good commentary.

    With regards to synchronised planning, I recall something I heard about 25 years ago on hosiery manufacture. It involved transmitting daily retail scan sales directly to a fully integrated scheduling / CAM system operating on MTO principles and 24 hour turnaround from receipt of scan sales data to product despatch.

    On another subject. I was thinking about business metrics and specifically those that companies use for inventory (SLOB, service level, stock turns / working capital investment). As you would likely agree, there is often little understanding of the relationships between these and the influencing factors of demand and supply variability, frequency of ordering / manufacturing, shelf life, etc. This might be one for a future article and to that end I can share with you a version of a “unified inventory cost” model that I sometimes employ. It provides a heatmap graphical representation of total inventory cost versus replenishment frequency versus service level.

    1. David. Thank you for the recollection. At a similar time period I experienced a retail chain sending scan data to a trouser factory and replenishment occurring within 24 hours. The advantages of minimal inventory compared to low labour costs have been generally ignored by senior executives. This could be due to the voice of accountants being accepted against that of logisticians.
      Thank you for the offer concerning the ‘unified inventory cost’ model. It is rare that the ‘total cost of ownership’ is considered in supply chain and location decisions. A methodology that helps logisticians to make a comprehensive argument is welcome.

Leave a Reply

Your email address will not be published.