May 1, 2013
If you ever had the chance to walk by a meeting at Monetate HQ, there’s a good chance you’d hear me saying the Latin phrase "Post hoc, ergo propter hoc" to everyone in the room.
The phrase translates loosely to “After this, therefore because of this.” And there’s a reason I’m saying it so much lately, particularly since we’re in this era of big data and more companies are making it a priority to truly know their customers.
They say this phrase often because in the quest to optimize the online customer experience, it’s really easy to mix up our old friends correlation and causation. This Latin phrase begs a question: Does one thing really cause another? Often, we assume that because we’ve tested something on a website, it’s led to a result. But is your action really correlated to a customer’s action?
Here’s a great example: One company that shall remain nameless decided to test a thicker divider line on an email campaign to see if a stronger dividing line between sections would impact results—and it did. The company, in fact, said the thicker dividing line improved customer engagement since it improved a few KPIs. And I have no doubt that there is correlation, but I have a huge doubt that there’s causation. It’s pretty hard to imagine the thickness of a line actually impacts the effectiveness of a brand’s engagement with its customer base.
To me, this is like the story of the man who wore his lucky pants to a job interview and got the job. Based on his logic, he should wear those lucky pants every single day for the rest of his life because that’s the only way anything good will happen in his life.
And that’s why it’s important to understand that some things just happen. The world is full of randomness. Maybe the thicker line in the email campaign had an impact, but maybe the company’s audience opens more emails on a Tuesday and that email was sent on a Tuesday. Again, correlation isn’t causation.
Navigating the Seas of the Customer Relationship
So what’s the solution? How do we stop confusing correlation with causation? Part of the solution requires a shift in mindset that moves from focusing on small tests to optimizing the relationship with the customer.
Sure, you might run an experiment and find that using a thicker line increases conversions. But is that really all you want? To move the needle for one transaction? The reality is that your company is in this for the long haul, which means riding little waves throughout the long customer relationship. A transaction is just a positive wave in the ocean of customer engagement, not a way to navigate the sea.
It’s not about testing what I call parlor tricks, like the thickness of a line or the color of a button. Testing a thicker line might drive some incremental revenue, but it doesn’t teach you anything about your customers. There’s no enduring value to knowing customers might prefer a thicker line. And that’s not an effective way to develop an effective, long-term relationship with your brand.
Instead, brands should run experiments to uncover how they can better listen to and engage their customers. Experiment with messaging and branding, try a new verb, test out a new voice and see how it impacts your customers. Experimenting with those factors will uncover learnings that are more valuable than just a few more sales in the short term.
The right experiments uncover learnings about how to communicate effectively in the same way online. It’s about finding new ways to listen to your customer and determine whether you understand their problems, their context, and meet their needs accordingly. You still have to be cautious, but when you experiment with customer engagement, you’re closer to seeking causation than you are correlation. And that will have a long-term impact on the bottom line, instead of just helping you hit this month’s sales goal.
And as we continue to navigate the challenges and rewards of the age of the customer, it’s incredibly important to take every step you can to optimize the customer relationship—not just one transaction.
Customer Relationship image courtesy of Shutterstock.