Forecasting on the Down Side

From Bob Houk, TPMA Executive Director

I’ve written a few times in the past on the problems presented by the currently declining economy for forecasters who are basing their forecasts on data assembled over the past few years when the economy was much better. Similar problems will be presented when the recovery kicks in, and forecasts are being based on data from the recession.

No matter how good your forecasting tool is, the data is going to cause you problems.

I was pleased to have a bit of backup brought to my attention, in the form of an article by Dr. Larry Lapide, who is Director of Demand Management at the MIT Center for Transportation & Logistics. Dr. Lapide says that in an economic downturn forecasters need to change their perspective: “Forecasters don’t have to change what they do, but what they focus on." He says that there are five components of demand variation:

• Promotions and special events. These include price cuts and new product launches.
• The business cycle and economic conditions. An example is consumers switching from luxury to value-priced goods when the economy sours.
• Seasonal variations. The traditional demand spikes during peak buying seasons that recur from year to year.
• Trends. Broad variations as sales rise and fall over the lifecycles of products.
• Unknowns. The demand changes that are difficult to explain.

In TPMA, of course, our attention is primarily focussed on the first of these. How will promotional events of various types affect demand? Dr.Lapide says that the role of promotion increases when the economy weakens:

As demand slackens, companies compete more aggressively for the shrinking pool of buyers. This happens at the expense of seasonality and overall trends, as the triggers that prompt buyers to make purchases in good times are overridden by the rigors of a tougher economic climate.

Dr. Lapide also agrees that the data we’re using needs to be looked at carefully, because “projecting from historic data becomes less reliable in highly volatile markets, and there needs to be more emphasis on understanding the impact of aggressively promoted products on sales.”

Having the right forecasting tools is more important than ever. But understanding and properly assessing the reliability of the data is just as important.