What image does supply chain modelling conjure in your mind? Perhaps you imagine the supply chain like a map in a second world war aircraft control room, with girls moving trucks around the country using long croupier poles; perhaps it could be the scale models of trucks of various sorts that sit on the shelves of your office at home.
These days however it is likely to be a computer based model, which may well be far less interesting to look at, but can convey incredibly useful information and provide insights not available by studying model trucks! These computer models can be customised to examine various parts of the supply chain in detail, while offering broad brush views of other areas. The trick is getting the right view of the right part of the whole. For example, a superficial model produced on a single spreadsheet as a costing exercise for a preparatory business case may not be the right approach for determining the drop sequence of vehicle deliveries, but both are rightly included in the general heading of supply chain modelling.
Some of these areas have had much development effort applied to produce commercial software packages which are capable of in-depth analysis and produce well formatted outputs. Yet other areas remain the province of bespoke analytical models. It is clear, however, that there is no one solution to the question posed to us as consultants “Can you produce a model of my supply chain”?
In our consultancy work we are often required to model detailed sections of a customer’s supply chain and we regularly produce complex spreadsheet models to assist in decision making. In these cases the model is highly customised to reflect the detailed nature of our clients business. We have found that it is generally unhelpful to attempt to shoehorn a client’s business into a standard model. The foundation of these models could be built upon our standard approach, but the final model will reflect the amount of effort put into understanding the data available and how reliable a source it is derived from. Of equal importance is how it is likely to vary in the future. Collecting and validating this data is a vital part of a project to model the supply chain and the work often throws up questions that cause clients to examine more deeply some widely held assumptions about their business.
The design of the model is guided by the questions that are likely to be asked of it. Models designed to address transport costs will require different structures to models intended to address warehouse issues. However, it is usually possible to link separate model pages through some parameter in common and see the effect that changing one parameter in part of the model has on output in other areas. A common linking feature is often the desire to examine the costs of the separate parts of the operation, but optimise the whole to achieve the lowest overall cost. This sort of analysis requires a deep understanding of the business and logistics activities and is best served by building a model from the ground up rather than attempting to fit the facts to an existing template.
Building such a complex model cannot be done quickly if all aspects are to be covered, but sometimes quick wins can be achieved if the most salient features are examined at a high level. This is where experience of this form of modelling can pay dividends as a high level stripped down model can be put together relatively quickly. Again a one-size-fits-all template is not often the best solution, but separate modelling modules with proven validity can be brought together to derive an overall solution.
Building a model is often regarded as a one-off activity, constructed to answer issues at a moment in time, but, if constructed carefully, a well thought out model can be used as an on-going resource. The model can be constructed so that new data can be entered onto a consolidated front page, or a new set of data from a linked database can be read and the effect on outputs observed. A certain amount of knowledge of the operation of the model is required to avoid the effect of garbage-in/ garbage-out, but these iterations can be carried out by people other than the original modeller. In these circumstances the model may be coming close to being used as an operational tool. If that is the ultimate required objective then once the model has been validated it may be possible to simplify the model to deal with a reduced range of inputs and a consequent reduced range of output results.
In summary, we believe that modelling of supply chains, whether end-to-end or of individual components, can add considerable insight to the understanding of both the strategy and of the issues involved.