Neither Water or Supply Chains need Big Tech to tell them how to Flow.

Nobody tells water how fast to move, its speed of flow responds autonomously to gravity and slope.

Similarly, materials can flow autonomously through your supply chain in response to demand without being told how by a Master Production Schedule.

In fact, when forecast-driven MPS/MRP is used the inevitable result, due to forecast inaccuracy (1), is the materials moving at the wrong speed - over forecasts cause excess speed leading to stock build-ups, and under forecasts cause service issues and prompt re-plans / schedule interruptions / expediting.

These interventions use additional high cost capacity &/or interrupt & slow down the movement of other parts, thereby increasing their lead-time, which causes yet more service threats and yet more interventions. The resulting vicious circle of service threats -> interventions -> longer lead-times -> service threats -> interventions etc also leads to excessive inventory through Little's Law: Inventory = Lead-time (days) x Throughput (daily), to see this demonstrated go to Are all Supply Planners driving red cars?

A simple way to quantify the inventory cost of this forecast-driven MPS methodology is to compare your actual inventory ￡ value with its theoretical, calculated by aggregating, for all your parts, their safety stocks and 1/2 ave. order quantities.

You will find a significant excess of inventory ￡ and that can be added to your unplanned over-time, increased lead-times & service misses as the costs of you attempting to tell your materials how to flow with inaccurate forecasts and an MPS using ERP / APS.

What should companies with such an under-performing supply chain do to address the problem?

The solution offered by Big Tech providers is more tech - AI / real-time control / digitalisation etc with its associated significant IT expenditure. But they all use the same underlying forecast-driven MPS/MRP methodology, albeit with faster, more frequent and perhaps automated (performance destroying) interventions. But they are destined to fail because the interventions generate the opposite of smooth flow - variability - and its inevitable cost generating buffers: capacity, inventory & time (2).

The alternative is to allow the materials in your supply chain to flow autonomously, the way water flows in a stream.

This requires a replenishment mechanism that mimics the way water flows - perfectly smooth material flow would involve all parts passing through the various processing activities, and on to customers, one by one exactly in line with demand; as each part moves forward and is consumed it is immediately replaced by the one behind and so on all the way through the supply chain.

In reality parts usually have to move in batches and sometimes processing machines break down or some parts are rejected for quality reasons. These are the problems that Lean can address using, for instance, SMED, TPM and TQM etc.

The element of Lean that enable parts, even if they are in batches, to start moving in line with demand is Pull. In the factory this is often delivered via kan-ban, but Enterprise-wide Pull (from the distribution network, through the factory and back to raw material acquisition) is rarely practiced.

Enterprise-wide Pull (also known as Demand Driven MRP) involves the natural stock locations (eg. raw materials, sub-assembly/bulk-break positions, finished goods) being decoupled from each other and each replenished independently using their own re-order point/cycle mechanism. The result is that customer demand drives end-to-end replenishment, the de-coupled inventory positions absorb natural process variability and the smaller you can make the batches the better the movement of materials will mimic that of flowing water by responding quickly to changes in customer demand.

Supply Chains are, of course, very complex and no-one is suggesting you should implement Enterprise-wide Pull with spreadsheets, despite their ubiquity for the management of ERP/APS supply chains!

Most ERP/APS systems don't have the functionality to implement Enterprise-wide Pull (3), and Big Tech doesn't want it because the technology requirements are not particularly advanced - the effectiveness of Pull is derived not from technology but from delivering flow.

The software systems with the necessary functionality are mostly cloud / SaaS which easily interface with your ERP transaction system (so a low cost pilot / proof of concept can be quickly undertaken) and they can be found at Demand Driven Institute - compliant software.

The minimal requirements are:

1. Monthly or weekly multi-echelon re-order point / cycle target sizing/re-sizing including a properly calculated service level/statistical safety stock
2. Identification & management of significantly abnormal demand events (eg. TV promotions, tender wins, significant seasonality with capacity constraints)
3. Generation of re-order recommendations using Pull logic
4. Support for new product / change of part / obsolete-runout events and management of planned factory/supplier shut-downs, and
5. A digital twin that enables capacity / S&OP / 'what-if' scenario planning using the ROP/ROC parameters, the planned supply sequences/cycles and the forecast.

What are the benefits of implementing enterprise-wide Pull?

The typical performance improvements of moving towards Flow are:

1. Achievement of planned service levels, from
2. Up to 40% less average?aggregate inventory, with
3. Manufacturing lead-time reductions of?30% to 50% (because of reduced queue times due to schedule intervention elimination), and
4. Significant improvements in OEE/COGs (for the same reason and consequent factory stability)
5. Without the costs of expediting, firefighting, cross shipping or the need for impossible to achieve highly accurate item level forecasts.

See DDI - case studies

See also Forget the Master Production Schedule and for more detail go to The Replenishment Problem (and how to solve it)

Why not investigate how Enterprise-wide Pull across your company's supply chain could transform its performance?

(1) World class forecast mix accuracy across a portfolio is 80% but the item level forecast accuracy achieved by c80% of the sku's (those with lower volumes and higher variabilities) will be found to be less than 60% - even with heavy investment in big data / predictive analytics / neural networks / AI etc.

(2) See To Pull or not to Pull (Hopp & Spearman 2004) for more on how supply chain flow variability, such as that caused by schedule interventions, causes cost generating buffers - time, capacity & inventory. See also SC Variability: what it is, why it's bad and how it can be minimised

(3) Specifically, most ERP/APS can't support the use of multiple, de-coupled/independent 'pull' echelons (eg. finished products and sub-assemblies and/or materials), nor do they have the functionality to support the inventory target (re)sizing or concurrent forecasting through the BOMs & routings for capacity planning.