Which X-Axis is the Right One?
December 31st, 2008 by Doug BrightI’ve somehow ended up with a copy of Men’s Health in my house as a result of my holiday travels. It’s been a few years since I last picked up a copy and now having checked out the most recent issue I can confidently say that Men’s Health stinks at statistics.
Actually, that’s unfair. Editors are generally pretty smart people but are no doubt under lots of pressure to fill page space. I can see how one might embrace a few logical fallacies if they make for compelling copy. Here’s my favorite (probably) misleading tidbit from the latest issue:
| 2000 | 2010 | 2020 | 2030 | 2048 | |
| Prevalence of Obesity | 57% | 71% | 79% | 86% | 100% |
| Health Care Costs | $82 B | $194 B | $438 B | $957 B | ??? Trillion |
Let’s ignore the hard-to-swallow 2048 projection of 100% obesity (they cite a Johns Hopkins study but 100% anything rarely passes a sniff test). Also, let’s forget we ever saw the I’m-phoning-it-in-on-this-one projection of “??? Trillion”.
We see here an example of the age old fallacy of attributing correlation to causation. Men’s Health chose their x-axis to be prevalence of obesity and, lo and behold, found that as obesity goes up, health care costs go up. Of course, that x-axis could just as easily have been average prescription drug price, % of people without insurance, or average miles driven per person per year. In fact, considering that nobody expects health care costs to go down, the x-axis might as well have been my golf handicap.
As quantitative marketers, it’s easy for us to point and laugh at people who confuse correlation and causation. But when we plot revenue against page views and see that the line goes up and to the right, aren’t we engaging in the same fallacy? Isn’t it more likely that an increase in revenue almost requires page views to increase? After all, there are plenty of things that could be driving revenue, and page views will always tag along for the ride.
There are a million x-axes out there that will make a revenue plot go up and to the right, and almost none of them are actually driving the revenue growth. It’s our job to find the one, two, or three x-axes that actually cause growth. If we aren’t answering that question with our analytics initiatives, can we be sure we’re spending our time on the right things?
Tags: correlation, statistics, x-axis
Add Personalized Product Recommendations to Your eCommerce Site
Istobe is a powerful recommendation engine that makes it easy to add recommendations to your eCommerce site. How powerful? Shoppers who click on Istobe recommendations spend 20-50% more than the average visitor.
