Fertilizer demand is governed by a variety of factors with complex interrelationships and does not lend itself to accurate prediction. This is not to imply that we can do without forecasts if they cannot be accurate. It is better to have a forecast with a tolerable margin of error than to operate in a vacuum without a forecast or with subjective, irrational guesses about the future.
Lack of preparation to meet demand leads to loss of sales and loss of profit to marketing organisations as well as loss of food production. High inventories to meet any demand eventuality result in high interest and storage costs, product loss and waste of foreign exchange. Heavy discounting and uneconomical prices caused by oversupply impose heavy strains on wholesalers and marketing organisations. If both underestimation and overestimation of demand produce serious consequences and if accurate forecasts are rarely obtained, we are in a dilemma. The way to resolve this is to select a practical methodology, periodically review the forecast and make adjustments to the methodology specific to each situation, so that the margin of error is minimised. A periodic comparison of actuals with the forecast is very useful to ascertain where assumptions have gone wrong, to what extent and why and to use the knowledge for making better forecasts in the future.
Beware of straight line projections when forecasting demand over the long term. The straight lines mean that demand grows at a fixed percentage, which does not happen in reality. The rate of growth or decline tends to diminish with time.
When a trend is projected into the future, the sensitivity of demand to various factors should be examined. It is preferable to arrive at a range of forecasts based on different scenarios. When alternative forecasts are presented to management of a fertilizer company or to the higher levels of government, judgement and experience should be used to select the most likely trend. This underlines the importance of interaction of scientific method and expert judgement to arrive at a forecast that makes sense.
Although only one forecast is normally used for planning, it is advisable to select the maximum and minimum forecasts to impart a degree of flexibility to marketing strategies. This also recognises the fact that no forecast can be one hundred percent accurate.
