One of my favorite marketing anecdotes comes from running direct mail subscriber campaigns for a newspaper. The control piece for our campaigns was what we called “the Double Ugly.” It was a black and white bi-folded postcard, with a basic “7 for 1” subscription offer – pay for the Sunday newspaper, get the rest of the week included. Its origin story predated me, but, it was inexpensive to produce, cheap to mail, and delivered strong results every single time.
Over the years, creative agencies were eager to demonstrate how their glossy, glittery, full color brochure style printed mailers could drive great response rates. They’d craft well written prose, hold photos shoots, gather testimonials, and create some truly beautiful pieces. We’d test their creative against our Double Ugly as the control, and, much to their dismay, the Double Ugly consistently outperformed whatever “pretty” piece it was being tested against.
Maybe it was reflective of the product being marketed, the target audience, or a sign of The Times (caps on purpose) as being serious business, but the Double Ugly became legendary as I shared the anecdote with other Marketers thereafter. Hard working creative isn’t always about the most elaborate production process. And, while beautiful creative is fun to make and so very tempting, it may not always be the best performer. Without testing it against a control, we might never know if the extra effort and costs to produce it are paying off.
While it’s increasingly common for marketers to “always be testing” these days, what is being tested is sometimes still elusive. Lacking a clear hypothesis, or, metrics that will indicate some statistically significant finding, campaigns start to veer towards what “looks good” or what “sounds fun” or what creative teams may want to make to showcase their own talent. This is not a bad thing, and I’m a huge fan of the overall creative practice, but, ultimately, one has to ask, “Do you want pretty, or do you want performance?”
On rare occasion, the optimized version over time ends up delivering both.
Happy testing!
