Buyer attitudes are as sensititve to changes in the economic environment as sales volume
Editor’s note: Jim Haughey is a market research and economics consultant in Sudbury, Mass. He was formerly vice president, market research and economics at Cahners Publishing Co.
Marketing managers know that sales vary predictably with changing economic conditions so they interpret and predict sales results accordingly. They know that much of the sales growth in 1993-94 was due to the economic recovery and not their brand management actions. Business cycle variation, at the market level, accounts for all of the variation in sales, except for gains or losses to competitive markets or technologies.
But few marketing managers account for the impact of business cycle variation in their data on buyer attitudes. Variation in brand specific buyer attitude data is a combination of changes in economic conditions, market shares and individual attitude scores. The business cycle impact is difficult to identify, even with a long and consistent time series of buyer measurements.
As a practitioner of both economics and market research, I have found that "hard" and "soft" market data vary together over the business cycle. Data measuring business cycle changes can be used both to forecast sales and other dollar-denominated measures and to separate the uncontrollable impact of the economy on buyer attitudes from the controllable impact of brand management decisions. The cyclical impact on buyer attitude data can be approximated by using one of the available consumer or business confidence indexes but this is less exact than using cyclical measurements extracted from market or brand sales data.
Removing the business cycle impact from buyer attitude data shows the actual effectiveness of brand management actions. It also permits forecasting of attitude data so products and communications can be changed with buyers’ changing preferences.
The first step in doing this is understanding how economic conditions and buyer attitudes in every market are linked through the business cycle. Step two is understanding how the business cycle is identified and measured in market and brand sales data. Step three is the simple mechanics of separating changes in buyer attitude data into cyclical and brand management components.
Business cycle dynamics
A generic description of business cycles is shown in Chart 1, often called a rate-of-change chart. Plot any of your monthly or quarterly product data for eight years or more, measuring anything sensitive to economic conditions, and it will look like Chart 1. It will have peaks and troughs with the same frequency and spacing. These are turning points in the rate of change, common to all markets, so they can be used to identify and measure the business cycle element in any data.
The business cycle is defined by two measures: (1) the time between high (or low) turning points which are points of time with similar economic conditions and (2) the change on the horizontal percentage scale between the last two turning points relative to the change to the second to the last turning point. The business cycle in Chart 1 has a length of 48 months with a relative amplitude of 80 percent for the latest expansion period and 111 percent for the preceding contraction period. With the right scales labeled on the axes, Chart 1 could represent any measure for any business.
Business cycles are driven by the operating adjustments constantly being made by every business to get to target margins, market shares, inventory/sales ratios or reserve capacity. Because the same economic changes (credit costs, foreign exchange rates, consumer confidence, tax legislation, etc.) course through the economy, affecting every market, there is usually a predominance of companies making the same type of adjustment. For example, an increase in credit costs that makes inventories more expensive will prompt production cutbacks to reduce inventories and cause an economy-wide slowdown in spending.
Some cycles are long; some are short. Some are very pronounced; others are very dampened. A cycle’s individual characteristics are imparted to each market as the cycle courses through the economy, typically taking about four years.
Chart 2 shows where selected markets are in the business cycle at a single point in time. While some markets are expanding, others are still experiencing the recession the leading markets passed through two years earlier. The trailing markets are linked to the leading market. What will happen next in the trailing markets is happening now in the leading markets. Each market reflects the impact of cyclical changes with its own unique timing and cyclical sensitivity, depending on how it is related to the rest of the economy.
Similarly, Chart 3 shows the business cycle position at a single point in time for various measures for a single market or brand. Orders, sales, inventory, selling price, capital investment, margins, etc., are linked. The impact from the general economy being felt today on orders will be felt successively on sales, selling prices, inventory and capital spending. For example, rapidly rising consumer income increases, in succession, customer orders from car dealers, dealer orders from the factory, the dealer’s selling price, the dealer’s inventory and finally, the dealer’s investment in facilities.
The income increase improves buyers’ attitudes about the purchase of cars and buyers’ evaluations of the average dealer and manufacturer. A few months earlier, when their income was either lower or less certain, buyers rationalized that they should delay a car purchase because the available cars were not good quality. Now, with more income, they rationalize their decision to purchase because cars (and dealers and manufacturers) are good quality.
The items on Chart 3 could be relabeled as "sales," "customer satisfaction with brand X," "probability of buying within the next X months," "probability of buying a top of the line model," "probability of rebuying the current brand," "importance of image relative to function in buying decision," etc. Plotting any of these attitude measures would show a consistent relationship over the course of repeated business cycles with any measure of hard data, such as sales.
Identify business cycle in brand data
This is a graphic illustration of time series trend/cycle decomposition, a topic in many statistics texts. It can also be done with variants of the Census X-11 seasonal adjustment program which is widely available or with the Early Warning Forecast program from Cahners Economics (Cahners Publishing Co., Newton, Mass.). And it can be easily programmed in spreadsheet or database packages. But the manual illustration shown below should be sufficient for occasional use. It will let you separate external from internal impacts on a brand so you can target the problems that you have the power to change.
(1) Plot the product sales data in the same format as Chart 1. It will have the characteristic business cycle pattern as in Chart 1. Use market data not brand data because this eliminates variation due to market share changes and minimizes the impact of any irregular events (a competitor’s facility was destroyed in a fire so you got much of its business for a few months). Use monthly or quarterly data for a long enough period to plot at least three turning points. If the turning points are not obvious, you can identify them with reference to an overlay plot of a related market or the next highest level of aggregation, such as all household appliances, which includes all washing machines.
(2) Calculate the length of the last complete cycle and the relative amplitudes of the last two periods between turning points. These calculations define the business cycle for this product.
(3) Plot the attitude data for the product in the same format as Chart 1. Assume that the attitude data is a customer satisfaction measure. Compare the sales and attitude charts to see how much, if any, the turning points in the attitude data lead the turning points in the sales data. Satisfaction changes before buyers act on the changed attitudes. My experience is that attitudes lag sales by about the length of the buying process. So attitudes are coincident with sales for low value, frequently purchased products and lag up to four to five months for the most complex purchases. You will have to estimate the lag from the advice here if you do not have a long time series of consistent attitude data.
You will also have to estimate the relative amplitude of the attitude data compared to the sales data if the amplitude turning points are not well defined. My experience is that attitudes change more quickly than actions. A decline in the evaluation of one of your many product features may not cause a shift in suppliers. And even a decline in a summary attitude, such as overall customer satisfaction or supplier preference, does not always cause a switch to a new supplier because of contractual constraints.
(4) Plot the "pure business cycle" for the attitude data by using the cycle length and relative amplitude data extracted from the actual sales data plot for the product, adjusted, if necessary, for the cyclical lag of attitude behind sales and more or less amplitude in attitude versus sales taken from the plot of the actual amplitude data. See Chart 4.
(5) Now add the actual attitude data to this plot. The deviation of actual from business cycle is the impact of brand management on customer satisfaction. If the actual data is above the "pure business cycle" estimate then brand management actions are improving satisfaction with the product. Satisfaction levels declined in the 1991-92 recession but the decline was partially offset by brand management actions that boosted customer satisfaction relative to the average for all competitors in the market. Satisfaction levels increased in the 1993-94 economic recovery but the increase was less than that recorded by the average competitor in the market.
Using business cycle data
Even if you have only four or five data points over a two to three year period - not enough to make one of the charts shown above -- you can judge where to put a few points on Chart 4 and determine if your brand management actions are sufficient to improve attitudes about your product.
Working with hundreds of companies, I have never failed to find the business cycle element in sales data. Actually, I prefer to use order data because it is not affected by shipping, accounting or inventory practices. The business cycle impact is larger for products bought less frequently and may be trivial, even though identifiable, for some consumer staples, such as canned corn.
Which data should be examined to see if the business cycle impact is large enough to be regularly tracked? Any data relating to preferences between high and low price point products should be first because buyers switch to economy from image, to short-term benefits from long-term benefits and to fewer functions when their incomes decline or become less certain. Summary brand or company preference measures should be examined because more "no preferences" are recorded when purchases become less imminent. Consequently, brand loyalty is cyclical. Satisfaction declines in a recession when buyers are pessimistic and less forgiving of minor faults. The cyclical impact on buyer attitudes will vary by market segment because the business cycle impacts different regions and industries with different timing.