Causality is the relationship between two variables whereby one variable is a direct consequence of the other. For a scientist in a lab, identifying causality is fairly easy because the causal variable can be controlled and the consequences observed. For marketers, such control is a dream, not a reality. Identifying causality, then, can be a real challenge.
Why is causality so important? Assume you’ve observed a drop in sales that you think is caused by a competitor’s lower price. If you reduce your price to combat the competitor’s when, in reality, the poor sales are due simply to seasonal factors, lower prices might give consumers the impression that your product is cheap or low quality. This could send your sales even further downward. Drawing the wrong conclusions about causality can lead to disastrous results.
Control is an important related concept. Control, in this context, means not the degree to which you can manipulate an outcome but rather the degree to which you can separate the effects of a variable on a consequence. For example, you have complete control over what the customer pays for the offering. You are able to manipulate that outcome. However, you have no control over seasonal effects. Nonetheless, you can identify what those effects are and account for their influence.
The first type of control is managerial control, whereby you have control over how variables in a marketing plan are implemented. You decide, for example, how many stores will carry your product. You can vary that number and have an effect on sales. The second type of control is statistical control, whereby you can remove the influence of the variable on the outcome mathematically. For example, you have no control over seasonality. If you are selling a product for babies and more babies are born in August than any other month, then your sales will go up in September. Statistical control allows you to smooth out the seasonal variance on sales so you can then determine how much of the change in sales is due to other factors, especially those you have control over. Statistical control is something you learned in a regression class. However, the numbers in a statistical analysis can be as easily approximated. You don’t necessarily need to utilize complicated equations. Consider the following scenario:
- Over the past five years, you have observed an average decline of 20 percent in sales for the months of June, July, and August, which also happen to be months in which many salespeople and buyers vacation.
- This year, the decline was 28 percent.
- You can therefore safely assume that about 20 percent of the decline this year was due to people taking vacations, as they have in years past; you can further assume that the amount of the decline due to factors other than vacations was about 8 percent.
Doing a simple analysis such as this at least gives you some idea that something new is going on that is lowering your sales. You can then explore the problem more completely.
So how do you figure out exactly what is the cause of such a decline? In some instances, marketing executives speculate about the potential causes of problems and then research them. For example, if the product’s price is perceived to be the problem, conversing with a number of former customers who switched to competing products could either verify this hunch or dispel it. In a B2B environment, salespeople who are aware of a competitor’s new lower prices might be the first to identify the problem, rather than marketing executives. Nonetheless, the firm’s marketing executives can then try to verify that lower prices led to the sales decline. In consumer-goods markets, there are often many segments of consumers. Rather than asking a few of them what they think, formal market research tools such as surveys and focus groups are used.