Have you heard something like this before: “Marketing’s no longer an art, it’s a science.” I hate that stance, because it implies that science and art are somehow polar opposites. Nothing could be further from the truth. If the phrase above makes you think anything remotely similar to these following points, you need to toss it out now:
- Start with the facts
- Ignore hunches
- Follow convention
This is probably what most people think of when they hear that “ marketing is a science.” But that’s not how science works. It works like this:
- Start with a hypothesis, which is probably little more than a hunch
- Perform an experiment to test your hypothesis
- Constantly test conventional wisdom to see when and why it breaks down
I fully embrace the idea that marketing is a science. But science is a creative discipline. It’s about discovery. And if you aren’t performing experiments, testing your assumptions, or indulging gut instinct, you’re doing bad science, and hence, terrible marketing.
If the internet has given marketers only one benefit, it’s the ability to perform quick, low cost, accurate experiments at a moment’s notice. It’s so much easier to measure our target audience and the influence of our actions on KPIs than it has ever been before.
Good marketing, good science, mixes the creative and the logical. It starts with admitting that you have opinions.
Nobody has ever failed to prove they were right by starting with “just the facts.” You can always cherry pick your facts and present a biased argument. When you start with your opinions, you can put them to the test. And that allows you to define exactly what to expect from your actions.
The purest example of this is the split test. You can start with an opinion, like:
- This headline will perform better than this one
- Customers will be more likely to buy if we’re upfront with them
- Scarcity is going to compel visitors to buy sooner
- The core concept of this landing page will outperform the core concept of the other one
Then you randomly assign visitors to one page or the other. And you don’t make a call until you reach genuine statistical significance.
It’s not always possible to conduct pure, controlled experiments like this, but that’s no excuse not to experiment or measure. As long as you are continually testing, measuring, and keeping records, you will eventually start to recognize what is working, and what isn’t.
More importantly, you will make discoveries. Without discovery, how will you ever outcompete the others in your market?