Forecasting the Turning Point
Customer spending fluctuates -- a little up, a little down -- all the time. But, normally, these small changes cause only minor inventory or production problems. However, dramatic changes in the country's economic climate can lead to a major turning point in a company's revenue growth pattern.
Knowing ahead of time that a major change in growth is about to occur allows managers to develop plans without the pressure of having to immediately react to unforeseen changes. Forecasting these turning points requires good data and the proper statistical tools -- and a bit of luck.
Statistical tools and techniques can, under the right circumstances, point to turning points in the economy. Since every industry behaves slightly differently, these techniques need to be adjusted for each company in order to find the techniques that can forecast turns in the market.
A number of
forecasting techniques are available, ranging from asking people for their opinion to creating complex statistical models. Two of the most important characteristics of good forecasting techniques are accuracy and lead time.
No forecast will be 100 percent accurate, but a good forecast will turn out to be close enough to the actual data so that developing it will have been worthwhile. In addition, it is important that the method selected be able to predict turning points -- such as when growth will go from positive to negative. It's these turning points that help managers develop plans early.
One technique used by economists to predict economic changes is to identify a series of data that changes direction several months before another type of data -- those "leading indicators" can frequently be used by marketers to predict upcoming changes in sales revenue.
Web marketers collect a great deal of data about people who visit their Web sites. Everything from the number of pages visited to the number of customer-service contacts can be tracked and recorded and made available for analysis.
This data can be compared to
key national economic data to identify leading indicators that can give you advance warning -- before changes occur in customer-buying behavior.
By comparing known leading indicators to a company's sales and Web-site-activity data, leading indicators can be identified that can predict turning points in the company's revenue.
For years, the U.S. government has made available data for
12 leading indicators that change direction before general changes in the economy.
An easy way to get started looking for your own leading indicators is to create some simple graphs.
Starting with two or more sets of data, use Excel or another program to plot them across time. If it appears that a series of data turns up or down several months ahead of your own company data, then slide the leading-indicator data a few months to the right and see if the curves line up.
You may want to start with your own data, comparing advertising expenditures to sales revenue to try to determine the lead time between changes in advertising and changes in revenue. Then, compare your revenue data to public economic data; look for a data series that changes direction a few months ahead of your own data.
Trying to line up data visually is only the starting point in the quest for finding and using leading indicators to forecast likely changes in revenue. A number of statistical and data-mining software products make it easier than ever before to create and test forecasts.
If you decide to dig into data mining, I predict you'll see a turning point in the accuracy of your revenue forecasts.
Cliff Allen is the co-author of the book One-to-One Web Marketing; 2nd Ed., published by John Wiley & Sons, and has consulted with companies on strategic marketing for 20 years.