Without a doubt my favorite statistical effect is the moderator effect. The technical definition of a moderator effect is a "variable that affects the direction and/or strength of the relation between an independent or predictor variable and a dependent or criterion variable." Everybody got that? Maybe an example will help.
Reality is that everybody thinks in terms of moderators all the time. Here's a very common business example. Let's say you work in the marketing department of an international company, and you just completed an analysis that found customers' purchase behaviors were strongly influenced by perceptions of how innovative your company is. However, you believe this effect may vary across regions and have your research team conduct regional analyses. Sure enough, you find that innovation perceptions are incredibly important in North America but much less important in Emerging Markets. You just found a moderator effect - Region moderates the effect of innovation perceptions on purchase behavior.
So, why am I giving you a brief statistics lesson? Well, we've spent the last three months on the customer feedback analysis section of this blog discussing customer centricity. We've effectively established the importance of a customer focused company culture and strategies (at least I hope we have), but there is one more important thing to discuss - not all companies will get the same benefit out of their customer orientation. In effect, the positive impact of customer focus on company performance is not the same in all contexts.
This means the business contexts and environments in which you operate can impact the level of customer orientation that your company should strive for. If your company operates in an environment with high returns at all levels of customer orientation, then it makes sense to put a lot of effort toward achieving the highest possible levels of customer focus as opposed to contexts where there are diminishing returns to higher levels of customer orientation.
Now is a good time to make a very important point, the best research suggests that customer orientation (as defined in my earlier blog on the topic) has an impact regardless of context, but the effect size does vary. And one more important point, before you go off and say that your company operates in a context where customer orientation isn't very important, do some research. That may be your hypothesis, but do yourself and your company a favor be actually testing these assumptions before acting on them.
There is secondary research out there to help you understand how your business environment may impact the effect of customer orientation. As with most areas of research, however, moderator effects are only investigated once we've clearly defined the concept and the overall effectiveness of it. Here are a few current hypothesized moderator effects that have received some empirical support so far:
- Greater market turbulence (the speed at which customers and customer preferences are changing) increases the effect of customer orientation on business performance.
- Greater technological turbulence (the speed of technological change) decreases the effect of customer orientation on business performance.
- More competitive markets see a stronger effect of customer orientation on business performance.
- Stronger macroeconomic performance decreases the effect of customer orientation on business performance. When the economy is weak (i.e., a recession) is the best time to increase your customer orientation.
- Regional and cultural differences have an impact on the effect of customer orientation with "Western" cultures having a stronger effect of customer orientation and "Eastern" having a weaker effect.
- The size of the economy in your home market has impact on the effect of customer orientation with larger economies having a stronger effect of customer orientation on business performance.
- Finally, mature markets demonstrate a stronger effect of customer orientation on business performance.
Troy Powell, Ph.D.
VP, Statistical Solutions
BTW, The existence of moderator effects is nearly universal, existing for almost all cause-and-effect relationships, which is a big reason why I believe blindly benchmarking yourself with specific best practices or norms comprised of unknown companies is not a best practice.
List of all posts in this series: