My favorite Albert Einstein quote:
Not everything that can be counted counts and
not everything that counts can be counted.
So what do you absolutely have to measure? If you don’t have time, energy, money, staff, management support, a degree in data diving, access to brilliant experts, or enough curiosity to kill a cat, what’s the bare minimum of metrics?
Avinash Kaushik was asked just that after his inaugural presentation at the eMetrics Marketing Optimization Summit in 2005. His answer was thoughtful. After delineating the metric trinity of behavior, satisfaction, and outcomes, he chose satisfaction. He concluded that if you can move the needle on satisfaction, then you’re doing all the navigation, persuasion, and segmentation right and the outcomes will follow.
Measuring customer satisfaction in a meaningful way is complex. A simple number of “likes” on Facebook or a Net Promoter Score by themselves are indicative and feel useful, but you really need to get to the level of an iPerceptions to understand what they are satisfied about or the sophistication of ForeSee Results to understand the drivers behind that satisfaction and the impact of improvements to those numbers.
Given the limited resources presaged above, I would recommend starting at the end and working backwards.
Your end goal might be the almighty purchase. Given a longer sales cycle, a webinar registration might be your most elemental event. Count those as your first gauge of success and then as an analytical yardstick for analysis. What to analyze? Measure them against your traffic sources to determine where to invest in more traffic. Measure them against alternative landing pages to see which is most effective. Then get ever more granular with segmentation as your guide until you reach the event horizon.
Segmentation is every marketer’s best friend, as everybody doesn’t do the same thing. Even if you only segment by time of day or by the ad message that got them to click, once you start chopping up your website visitors into multiple buckets, you’ll start to see how different types of people respond to different stimuli. (See Change the Message for the Medium for more.)
So what is the event horizon? That’s the moment between orbiting a black hole and getting sucked down into it. It’s when you know you have stepped into the measurement quagmire just a little further than is wise from an ROI perspective. It’s when you have been taken over by the puzzle and forgotten about your goals.
On this side of the event horizon is the A/B split testing and multivariate testing that let you sharpen your advertising and messaging and persuasion techniques. On the other side is the black hole of honing that blade for the sake of hearing the metal against the whet stone.
Let’s say you make a dynamically driven offer to everybody who comes to your site. The less likely they are to buy, the more enticement you offer. But even those who are in the 98 percentile of purchasing likelihood are proffered free shipping and a 10 percent discount if they’ll just click the “Buy now” button…now. You have tripped over the edge of the event horizon and are falling headfirst into the black hole of testing for testing’s sake. These people were going to buy. All you had to do was leave them alone!
While I don’t recommend staying so far from the edge that you miss easy sales, I do recommend that you keep an eye open for where too much measurement costs more than it’s worth. Between customer satisfaction and visitor segmentation, you can get a lot of value from marketing optimization. Just stay back from the abyss. Even Einstein wasn’t sure what’s on the other side of that obsession.
Emily Ma, product director of Tencent’s advertising platform products department, was a keynote speaker at ClickZ Live Shanghai where she discussed the ... read more
Nurcin Erdogan Loeffler, head of strategy and innovation, Vizeum China, outlines the seven ways businesses can future proof their digital strategies.
The terms that customers type into your site search function can help you to gain an understanding of user behaviour and can be used to optimise ... read more