Home » Roles » Management Consultant » Document #341

[341] Why all the Supply Chain focus on S&OP ?    « Back to Category
Author: LCCMod1, Created on: May 24, 2016 7:06 AM
Keywords: Lean, S&OP, supply chain
Categories: Management Consultant
Language: English
Rating: Not Rated

Great article explains theory behind why things work or not


Why all the Supply Chain focus on S&OP ?
Aug 23, 20142,839 views59 Likes28 CommentsShare on LinkedInShare on FacebookShare on Twitter
Since the acronym first appeared in the early 1980s, S&OP has cost companies multiple $ millions in consulting effort to get it up and running and more in management time to actually operate the process. Over the years S&OP has been re-incarnated a number of times and is now often known as IBP. Whether its S&OP or IBP (or Core Commercial Cycle which is another term I came across recently) the process is justified by "evidence" that its implementation improves corporate performance by somehow aligning / synchronising / unifying the activities of the commercial, supply and innovation functions. And to this, nowadays, can be added, with the advent of new IT capabilities, the potential for companies to use S&OP to evaluate the financial "best' use of resources using 'what if' scenario exploration across a global network.

According to Gartner however, the vast majority of companies, on an S&OP maturity scale of 1 to 5, fail to develop beyond the Stage 2 version.

Why is this?

A key reason is that companies have been focussing on S&OP far too early on their road to supply chain excellence. S&OP obviously cannot be of value to companies that are totally incapable of delivering a Plan. And most companies, due to the inadequacy of their supply chain management processes, find that all the effort they put into Planning proves so divorced from the reality of what actually happens that S&OP becomes discredited and little more than a mid level management review process instead of the C suite strategic forum it is sold as by the many Consultants / Gurus etc who advocate its implementation.

Which is not to suggest that S&OP is of no potential value. Its value can be immense but first companies need to implement supply chain management processes that can be Planned, and Delivered, with some degree of accuracy. Unfortunately, most companies use Replenishment processes that are so ineffective that they are, in effect, out of control and no amount of S&OP / IBP / CCC is going to be of any benefit.

How Forecasts Degrade Supply Chain Performance

These companies are using forecasts for both Planning and Replenishment and, while the former is obvious, the latter is plain wrong and leads to totally inadequate supply chain performance that is always different to that which had been Planned - in terms of service, inventory and capacity utilisation.

As forecasts have been driving supply chains for many years in most companies, and their use for this purpose appears pretty intuitive, the above paragraph might be considered something of an exaggeration. It's truth, however, is borne out by evidence and very robust scientific principles (1).................

It is well known that all forecasts are wrong and 80% portfolio mix accuracy (ie. 20% wrong) is considered ‘world class’. Due to the 20:80 rule, such performance means that most medium and low volume sku’s (usually the majority) actually achieve accuracies that are somewhat worse, not least because with lower volumes, variability tends to be higher and so too is the level of forecast inaccuracy. Manufacturing or purchasing schedules based upon inaccurate sku forecasts lead to the production of unbalanced stocks with potential service issues, and expediting inevitably follows as Planners respond to exception messages.

Service saving production schedule expedites are a major source of flow variability and cause costly unplanned machine changeovers, schedule congestion and increased lead times with knock on effects upon other schedules up and down the factory routings. In consequence, average lead times increase and become volatile (contrary to the DRP/MRP assumption of fixed lead times, therefore causing a further service risk) and stock becomes both excessive and unbalanced with service issues often continuing to occur. The further up the supply chain, and away from end customer demand one goes, these problems are amplified by MRP's dependent demand, batching and latency (ie. response delays and misinterpreted demand signals).

A typical reaction to this set of problems is for SC management to increase time buffers with ERP system parameters such as lead time and time fences. Unfortunately these just make matters worse by leading to more work being released to the factory floor which puts more pressure upon capacity and significantly increasing both lead time and stock while reducing responsiveness. In this way companies end up with the classic combination of unplanned and unsatisfactory service, excessive stocks and higher than planned use of over time.

Another common response, of course, is even greater expenditure upon forecasting software (e.g. demand sensing) followed by inventory and schedule 'optimisation' technology. Both routes are actually misguided because they are attempting to improve a replenishment model which is entirely inappropriate for what is, in effect, a non-linear system.

Simple Queuing Theory, Lean and Flow

The relevant underpinning of Supply Chain Management and Operations is Queuing Theory which recognizes that any process involving at least one conversion stage, be it a supermarket checkout counter, a factory’s production line, a distribution channel or supplier, has finite throughput capacity (ie. is a constraint) and always suffers from queues. If demand exceeds capacity the queue will grow indefinitely. Even if demand is below 100% capacity, however, there is still a queue. This is because variability in the rate of arrivals at a work station and / or its processing time will cause both lost capacities and demand spikes and the development of a queue with finite average length (NB. if there is plenty of spare capacity, however, no queue will actually be observable for most of the time).

The average waiting time in one of these queues, for a given level of capacity utilisation, is directly proportional to the flow variability. On the other hand, the relationship between the queue length and the level of capacity utilisation is exponential, and very noticeably so when variability or utilisation is high. Both relationships are demonstrated s follows:

The relationship between queues, variability and capacity utilisation can also be modelled using the VUT equation, a simplified version of which is

Average wait in queue = Variability x 1/(1-Utilisation) x processing Time

We know that queue time is a major component of lead time and through Little’s Law we also know that lead time is directly related to inventory level

System Lead Time = System Inventory / System Throughput

Little’s Law and the VUT equation are at the heart of Hopp & Spearman’s ‘Factory Physics’ and are used to explain the underlying rationale for why the tools of Lean are so effective.

Conceptually we can understand that if supply was unconstrained and totally flexible we could meet any demand (ie. volume and mix) and achieve perfect continuous FLOW without holding any static stock, either in process or as finished goods. In the real world, any inability by the supply conversion process to respond instantaneously to demand changes implies the existence of a constraint (eg. a processing machine) that inevitably has flow variability which causes the creation of cost generating buffer. The default buffer response to variability is an immediate increase in lead time due to queuing, ameliorated by use of spare capacity (if any); management may respond with an increase in capacity (eg. overtime) and finished goods inventory up-sizing often follows to prevent service issues.

Lean activities such as TPM, TQM, SMED, 5S, Standard Work, DFM etc eradicate not only wasted effort and cost, they also reduce flow variability, including that caused by excessively big batches (thereby supporting flexibility), which means less cost generating buffer is created or needed. And this is particularly valuable when working at high levels of capacity utilisation. In consequence, the Lean supply chain uses less unplanned capacity and can operate at higher levels of capacity utilization (ie. lower costs), with shorter queue / lead times (ie. greater responsiveness) and lower levels of stock. Reducing flow variability is actually synonymous with minimising cost and inventory while increasing responsive flexibility, and is why Hopp & Spearman describe Lean as:

“..........fundamentally about minimising the cost of buffering variability”

Lean’s full potential has rarely been fully exploited, however, because most companies, through their inaccurate forecast driven replenishment execution processes are, as has been described, still unwittingly introducing huge amounts of re-schedule and expediting variability (ie. forecast error induced) into their supply chains and operations.

Demand Driven Replenishment and Flow

In addition to the ‘shop floor’ Lean that increases flexibility and minimises factory variability, Operations and Supply Chain leaders can significantly reduce the performance destroying and Plan confounding impact of the supply chain variability that they generate themselves through their use of inaccurate forecasts to drive replenishment execution.

The Demand Driven approach uses replenishment processes that are autonomously responsive to real demand, and its variations, with minimal cost generating buffers that are of the right size and in the form that best serves both the company and its customers.

Companies that adopt these techniques sustainably achieve their desired service levels with inventory reductions of between 30% to 50%, lead-times of up to 85% and cost savings of c20% through better use of capacity. A secondary metric that also improves significantly is schedule adherence which tends to stabilise at between 95% and 99%. These companies no longer, of course, have any further concern for sku level, weekly / monthly bucket forecast accuracy.

The Demand Driven process uses a segmented approach to replenishment technique selection based upon the item’s demand profile in terms of volume and variability. The key segments are as illustrated below:

And the key options replenishment techniques are, broadly, the following:

High Volume / Low Variability – where demand is high and relatively stable, as it often is for mature products and for upstream items before sku specific customisation takes place, supply can be leveled at a suitable fixed rate, subject to periodic review or using min-max logic. As with all ‘make to stock’ replenishment techniques, this rate-based or level schedule technique requires some inventory buffer.
Medium Volume / Medium Variability – these, usually the majority of items in a portfolio, use a form of consumption based pull whereby inventory is positioned in the supply chain to decouple processes, minimize lead times and prevent residual process variability being propagated through the supply chain.. Supply activities at each work station up and down the supply chain are scheduled according to an efficient sequence / cycle, and stock targets (buffer plus average demand over lead time) calculated so that the quantities triggered, rounded as necessary for increments / MOQs, replace what has been taken from the location immediately down stream. The consequence is that both the timing and quantity of replenishment responds autonomously in line with demand while capacity utilisation is kept high and level loaded. The sequence predictability, virtually guaranteed availability of components and lack of schedule interruptions mean that this technique brings significant cost saving stability to Operations. The technique can be used even if demand has trend or is seasonal so long as the replenishment parameters reflect future demand patterns appropriately.
High Volume / High Variability - when demand is high and genuinely volatile so that it is uneconomic to provide a stock buffered service (eg. response to tenders, significant promotions and other events), time buffered responses such as ‘make to order’ or ‘assemble to order’ are options for Event Management. Both these ‘postponement’ techniques, if supported by appropriate ‘design for manufacture’ and asset configuration, can deliver very cost effective and quick response. In these situations, use of an event forecast to drive advance stock build is also an option, but this is very different from traditional ‘forecast push’ MRP.
Low Volume / High Variability – low volume items with or without high variability can usually be serviced ex stock from high demand coverage batch volumes (eg. 2 bin systems), especially if the item is of low value. Otherwise, and depending on the demand pattern, MTO, ATO or Poisson based buffer techniques can be used.
These techniques can be used in combination at different levels of the supply chain. Level schedule might be used upstream where materials are common across many SKU’s which themselves are replenished using consumption based pull. Similarly, in a contract manufacturing business, material supply may be managed using consumption based pull to support a fast ATO response to customer orders. Companies might also add value by providing their customers with a VMI service using collaborative demand driven techniques. This also has the benefit of allowing them to respond to relatively stable end user demand, instead of lumpy customer orders, which enables them to more efficiently level load their entire supply chain using rate based or consumption based pull.

The New Planning and S&OP

The implementation of Demand Driven techniques has a significant impact upon the Planner’s role. Instead of constantly expediting and re-cutting inaccurate forecast driven replenishment schedules, the Planner can concentrate properly upon value add activities such as Supply Chain Conditioning.

‘Conditioning’ enables the supply chain to autonomously respond to demand, it is undertaken at each stock level echelon through demand profile analysis and replenishment technique selection and calculation of rate and stock targets. These activities are performed regularly (eg. technique selection annually and parameters monthly) but changes are very much by exception – only around 5% of targets may need changing at any one time.

Planners will continue to be involved with NPL, Phase Out and Event Management, albeit the latter less frequently, and as the Demand Driven principles can be applied to manage replenishment across company boundaries, Planners will also have time to work a lot more collaboratively with key suppliers and customers.

In addition, of course, forecast based S&OP can now add real value as a senior management process. Due to the eradication of forecast error induced variability, aggregate supply chain activity and performance will closely match that which had been Planned and 'firefighting' will be a thing of the past. As a result, S&OP will earn credibility in the eyes of senior management and will become a valued, reliable and accurate forum able to support real evidenced based tactical and strategic decision making.

If you would like to learn more about how Demand Driven ways of working can transform your company's supply chain performance, a great place to start is www.demanddriveninstitute.com

1 - see https://www.linkedin.com/pulse/can-scm-deductive-simon-eagle?trk=mp-reader-card

Report this
Sales Operations
Supply Chain Management
Simon Eagle
Written by
Simon EagleFollowing
LikeCommentShare on LinkedInShare on FacebookShare on Twitter59 likes28 comments
Jon Kirkegaard Add your comment

Add your comment
Jon Kirkegaard Jon Kirkegaard YOU
Owner / President at DCRA Inc. Automotive Engineering Division
Your explanation of queing theory is excellent. The Issue with applying the basic principles you explain in multi-enterprise supply chain is management is not line of site, manage by walking around as in a single facility. This is why S&OP can be so effective as it is the queing, load balancing, measurement system required in a an "invisible suppy chain". Without data and data un corrupted by local optimas ... local optimas win

Document Rating
Rating: Not Rated

No comments.