• by David Baker , Op-Ed Contributor, October 10, 2016

With all the hype about machine learning, I’ve been thinking a lot about testing lately. Experimentation should help answer questions for you, not cloud your view, right? 

I asked a blind question of a variety of retail marketers, technology marketers and “insiders” recently,  just to gauge if people are thinking about this opportunity differently now.

The question was simple: “What are the three biggest challenges you face in doing any type of testing for email marketing?”   

I got an interesting dichotomy of responses.   For instance, I asked this of a brilliant marketing technologist, and we ended up threading for 10 minutes on the merits of my question, since “testing” meant something completely different in his world. 

I asked a media and search guru, who was very “by the book, there are ways to do it, and we all know those sets of playbooks.”

When I asked retailers and email marketers, I got the responses you’d expect: 

1. Not enough content.
2. Limits on resource.
3. Not enough time to test.

 When I asked strategists, both tenured and those I would call novices to email marketing, the answers were so long-winded I was literally exhausted trying to glean good thinking on the subject. 

So, I decided to draw what I thought the biggest challenges were on a whiteboard, without words.   I often find that forces me to the root of the problem.   Am I thinking about this the right way?

Well, my simple board turned into an algorithmic equation that was likely more confusing that when I started.   Smart and practical don’t often end up on the same board.   But it was a whiteboard session, so you’re not supposed to judge, right?   I had an A , B, C, Z, multiplied by T, D, with circles around it, then I had a line under it to represent division of T/S/R multiplied by an S, C, E…  

Confused yet? Made perfect sense to me, but didn’t help me answer the questions.  However, it did help me break down something complex. I landed on three simple questions:

1. What type of marketer are you?  A scientific marketer, all about the stats and statistical significance of the experiments?   A conservative marketer that won’t pull the trigger unless you have a confidence level that will CYA? Or are you an erratic marketer that will shoot at everything, hoping something will stick?

2. Do you think testing is about finding winners? To your boss, the answer may be yes, but deep down, you likely just want to be able to adapt to the market and try new stuff. Isn’t that why we all got into marketing — for the creativity and cool stuff related to consumer and market behaviors?

3.   Do you love marketing because you like selling stuff?    At the end of the day, is that how you measure your development and what you love doing?   There isn’t a wrong answer   But the root of your response is important, as it will help you see where you gravitate and where you’ll be best at experimenting.

The killer moment in all this was when I looked at that equation on the board and the comments from others and ended with one principle as a motivating factor: The world has started moving so fast that the old ways of thinking about testing have to change, and you will end up gravitating to habits driven by your internal career aspirations.   This is important for how you design and sustain a testing/optimization culture in a company.   

As they say, the best thought leadership helps you become better without telling you the answers.  So I’ll end with a few sound bites that you should take to the whiteboard therapy session with you.

— Machines help compress time and space, but won’t ever replace intuition.
— Optimization should not have a steep adoption curve.
— Experimentation is like yoga, testing is like aspirin. 
— Slow down to speed up.  Think about what you really need to optimize.