Sales and Operations Planning Models for Different Sectors

This paper looks at the types of planning model that can be used for Sales and Operations Planning. In our experience, even in conventional, standard-product manufacturing, where S&OP originated the S&OP model needs to be tailored to fit each individual business. Other types of business can require fundamentally different models.

S&OP processes are quite well defined for manufacturing companies where two basic models exist. However, for other environments S&OP has not yet evolved to the point where there are well established and widely accepted models.

This widely accepted model resolves the issues faced in taking a longer term strategic view of the business by addressing the issues of complexity and forecast error by grouping products into families. Sales family groupings allow forecast errors and SKU numbers to be reduced. The Group Technology view of the manufacturing process allows resources, whether material, labour or equipment to be aggregated into the few that are strategically important.

Whilst mapping the two groupings onto each other can prove difficult, it is usually possible to arrive at a good enough approximation to allow the strategic pinch points to be identified. Once this is done the ability to flex the model, providing suitable financial figures are attached, allows different scenarios to be evaluated and comparisons made.

This model is long-established and well proven but not widely known or understood. It suits project-based manufacturing businesses. In this case simplification is achieved by grouping resources, ideally where there is a reasonable degree of interchangeability, and by producing high level project plans that retain the main interdependencies but reduce the activity definitions to a small and manageable number. Providing the high level plan is supported by an equivalent detailed level plan, the resource assignments can be tested.

In this model the scenario planning is achieved by moving all the projects into a multi project environment and using resource constrained scheduling to evaluate the project outcomes. By flexing the project priorities and the resource availability it is possible to model to costs and outcomes of different options and look at the impact of changing the mix or adding in new projects.

This model addresses multi-location distribution networks. It is the subject of much interest, but also much debate. How a supply chain model should be configured depends on the predictive role it needs to fulfil. In most cases it needs to look at supply chain capacity impacts as well as the usual manufacturing capacity issues. These may include:

• space requirements;
• transportation;
• lead time;
• holding costs;
• customer service

Some of these elements can be managed through a model similar to the Standard Manufacturing Model. Grouped resources such as storage space, manufacturing capacity, strategic material supply and transport are susceptible to this treatment. However, if one of the predictive roles is the optimisation of the inventory disbursement across the supply chain the family grouping approach is unlikely to work, as it assumes a common relationship between volume and variability within each family group.

Simplifying this aspect of the model requires products to be grouped according to the nature of their demand characteristics, separately for each of the stocking locations, rather than their product families. A demand analysis matrix can be used as a basis for such a classification. For products in all but the high value erratic demand quadrant, the inventory impact may be determined using general groupings. However, the few high value erratic demand items may need to be modelled on an individual basis.

To be of value the model needs to look at the impact on the supply chain of the variation in forecast error for those items, as a test of sensitivity. This means that it also needs to define links back to the factors driving that error, so the model can be used to predict the impact of self generated variation.

The nature of the model needed depends upon how demand is driven.

Some businesses, particularly those supporting the construction industry, are very much driven by customer projects. In this case, the model needs to combine aspects of the Project Environment and Supply Chain models, generating resource profiles as in the project environment, but also modelling the allocation of inventory across locations and projects. Once again, classification by demand is important but it may also be possible to simplify in other ways by grouping by volume and value, to come up with a range of “resource categories” that can them be assigned to projects at a summary level. Modelling resource requirements, as in the project model, and inventory allocation will then provide a strategic view to inform the more important investment decisions and “bid/no bid” decisions.

In other cases, demand is not driven by specific events or projects and follows more conventional patterns. In this case, a variant of the Supply Chain Model will be appropriate.

In either case, the model needs to predict the likely receipts of items returned from the fleet in the field. This aspect of the model will be very dependent upon the particular applications of the items that are being managed.