How Trends Impact Quantitative Forecasting

Quantitative sales forecasting works for reasons that are sound, mathematical, and logical, and you can find plenty of examples of forecasts working in practice. If someone looks at you suspiciously when you trot out your sales forecast, you’ll want to understand the reasons that forecasts aren’t some kind of magic. One of those reasons is that many baselines — particularly baselines of sales revenues — involve trends.

A trend is easy to define, if not always easy to manage. A trend has two main characteristics:

  • It goes up (good, if you’re measuring revenues) or it goes down (not good for revenues). It may fluctuate — for example, you’ll see some temporary downturns in a baseline that’s trending up — but in the main it’s going in one direction. If you see many consecutive increases followed by many consecutive decreases, you’re probably dealing with a seasonal or cyclic baseline, and certainly not a trend.
  • The trend lasts much longer than the forecast period. Suppose you use the results of four calendar quarters to forecast a fifth quarter. You may find that the trend established for the first four is substantially different from the one you’d get when the fifth quarter’s actuals are in. But a trend figured on, say, 20 calendar quarters is unlikely to change much in the 21st quarter, unless a sudden and major sea change occurs in the market.

What causes trends? There are as many reasons for trends as you care to think up. Just a few examples:

  • Products go out of style. Smoked a cigarette, pipe, or cigar lately? If you have, you’re out of step. The society you belong to is frowning and wagging its finger at you. You’re having difficulty lighting up after a meal in a restaurant. The trend, my friend, is down: People don’t want to smell your smoke anymore (mine either, and thank heaven for the nicotine patch).
  • Inflation sets in. As this book is written, the United States has had very low inflation for several years. During the 1970s and into the 1980s (when high interest rates finally slowed it down) there was serious inflation in the U.S. economy. The inflation caused prices — and therefore revenues — to trend upward.
  • Technology improves productivity. When your parents or grandparents or even great-grandparents were small children, they might have set aside one day a week to do the wash, outdoors in a big metal drum filled with soapy water that they’d churn with a big stick. Then along came washing machines. Washing machines were initially very expensive compared to a metal drum and a stick, and the demand for washing machines was, therefore, fairly low. But as economies of scale kicked in, and unit prices came down, what was once only for the wealthy became the norm, and the increase in revenues far outweighed the decrease in unit cost, until washing machines became commodities — then revenues flattened out. But there was a significant trend for decades because of increased productivity and demand.
  • Products become more popular. There are more cars, trucks, and SUVs on the road than there were last year, and last year there were more than the year before, and so on, all the way back to the Model T and even earlier. And each year that a census has been taken, the population of the United States has increased. You get more people, you get more people wanting to drive, you get more cars and trucks and SUVs.

Trends are a principal reason that forecasting works. But if you can tell that a baseline is stationary — trending neither up nor down — you can do every bit as well as you can with a trend, if you handle things correctly.