In a previous post, “Actionable Analytics,” we discussed the difference between actionable analytics and vanity analytics. The core difference is that actionable analytics provide you with information you can use to do something to improve your business. Of course, the key problem is figuring out what constitutes an actionable analytic for your specific business. While understanding the nature of your business helps you to zero in on which analytics are crucial, testing provides the only viable way to determine this with certainty.
Yet, for reasons both practical and mysterious, businesses often shy away from performing tests with their readily available data. Some of the hesitancy presumably stems from a fear of resource drain. Essentially, “I can’t afford to have my employees running a/b split tests on Facebook ads because I need them doing things that contribute to the bottom line.” Without beating the dead horse too hard, dumping time, energy and money into any effort without a reasonable certainty it actually does contribute to the bottom line is suicidal for a business.
If you believe that driving traffic to your website from Facebook increases your ROI, test it. See if there is a correlation between boosting total visitors from Facebook and your total sales. If moving from 5,000 visitors from Facebook to 15,000 visitors a month doesn’t substantially boost your sales, driving traffic from Facebook isn’t profitable. Or, at the very least, you know you’re driving the wrong traffic from the social media giant. Without testing, you’re forced to guess about causation.
If your marketing people tell you that small changes in the language of online marketing content can improve sales performance of a given product, or even improve social sentiment, a blanket change to language is as dicey as doing nothing. With the existing language, you’re dealing with a known quantity and consumer response. With new language, anything could happen. Sales could spike or plummet or stay the same. The testing is where the pudding proves out or fails utterly.
While professional intuitions about customer reactions to a new product, new marketing language, and even new social media platforms shouldn’t be discounted, the technology to test it is available. Testing data is, in the long run, essential to the bottom line. Moreover, there is rapidly improving software that streamlines processing that data into something intelligible and testable. To give up those potential insights is to rob your business of the exact thing it needs to succeed: knowledge.