Data, Not Preference, Is What Drives Digital Marketing Improvement

It’s said that stats can be used to prove anything.  That is a true statement when we allow our preferences to bias how we conduct digital marketing campaigns.  Digital marketing should be data-driven and changes should be honestly tested to see what is most effective. Dictating changes based on preferences will suit your tastes and make you feel like your gut feel is spot on, but data will drive real performance improvement.

The trouble with preference bias is it’s something people often aren’t self-aware of.  Trainers, consultants and professional coaches unknowingly craft experiments that make their preferences shine through as the best way of doing things.

We had an obvious case of this happening recently with a client. The client attributed his email marketing campaigns success to putting questions in the subject line. The problem was that the open rate had been in a noticeable decline over the last twelve months. Our client was resistant to testing other types of subjects because he was certain that wasn’t the problem.  He had used questions in the subject consistently and had several best-practice articles that sited questions as the best converting subjects. In fact, he said, “I tested subject lines that weren’t questions seven months ago and the open rate was worse.”

After experimenting with some other potential causes, including changing email marketing platforms to make sure that delivery was not a problem, we reviewed the test he had run.  It turned out that he had used the subject right after adding a new list from a trade show.  Many of those first time subscribers were lured into signing up for his email list but weren’t motivated to read his campaign, at least not right away. The bounce rate data confirmed that the new contacts were the catalyst for the open rate drop, not the subject line. Upon this realization he agreed to trying subjects without questions.

While the other tests produced small or moderate changes, the updated subject lines produced the most notable improvements.  Of course this is not to say that questions make for bad subject lines.  But it definitely does mean that exclusively using them in this client’s case was negatively impacting his email marketing performance.

There are many potential biases in digital marketing and none of them should be universally adopted without testing.  If you are not implementing changes that are counter to your preference on a semi-regular basis then there’s a fair chance that your preferences are driving your decisions rather than the data.

Photo credit: Wikimedia Commons, Atlasowa