Predictive analytics: a simple approach to dealing with price shocks in raw materials

,

Simon BittlestoneSimon Bittlestone, Managing Director, writes on markets and economics

As an organisation, the International Grains Council (IGC) may not have the public profile of say, the ISO, or the World Bank, but its analyses of worldwide cereals production offer intriguing insights into the very top of the supply chain for food processors and retailers.

The IGC’s latest update outlines what appears to be a sated marketplace: record harvests in some areas, coupled with a 29-year peak in inventories. In contrast with some extractive industries, depressed emerging-market currencies are serving to support foodstuffs production in many countries. For some products and territories, however, prices have now fallen well below the cost of production. With global population rising even as global agricultural yield rises too, some sort of adjustment seems inevitable in the medium term, and a wide range of scenarios are possible.

For downstream businesses, gauging the likely consequences is not easy. But, if processors and retailers are to minimise the negative impact, they must be able to model the effect of all likely eventualities. Knowing the effect of a given price movement in advance allows business leaders to put together an action plan than can be implemented if that scenario transpires. Furthermore, the process used to understand the impact of cost changes on the business can, if it is accurate enough, also assess the likelihood of a given situation occurring.

Conventional BI and CPM software, which is usually built on a rigid data architecture, simply cannot do this. Neither can so-called ‘big data’ applications, which do not take account of the structure or circumstances of an organisation. The only way to gain reliable insight is to visually map the drivers of value that affect each outcome of the business, arrange the data relevant to each driver in a hierarchy that reflects that structure, and then apply the calculations that govern the effect of each input (for example gross sales value = volume x unit price).

That creates a very robust mathematical model of the business, and simple probability calculations can then be applied to likely movements in each input. The impact on the outputs (i.e. value in the business) is then easily assessed. It is also possible to represent these outcomes as a visual range (such as in a fan chart), which allows business leaders to quickly evaluate the likelihood of a given scenario taking place.  That may influence medium-term financial decisions, such as future derivative purchases, and should also inform actions to preserve revenue and margin in the short term.

These might include negotiating new pricing structures with retailers or suppliers, looking at cushioning the impact of price increases on customers, or changing the product mix to reflect the likely impact of price changes on market demand. None of these actions can happen overnight, but an awareness of which will be necessary, and in what circumstances, means that the process can begin early; vital for competitive advantage in industries where all players are affected by commodity cost changes.

Leave a Reply