Experiment more

This year instead of resolutions, commit to experiments.

Experiments are where you have an idea (hypothesis), a rough plan, and commit to trying a change.

Unlike resolutions, they aren't a succeed/fail situation.

With a diet resolution, you succeed or fail every day. Same for reading, personality changes (e.g. be nicer), and just about every change.

An experiment through, that's you testing something different to see if it works. If you don't do it on one day, no big deal. That day's data won't count in the results.

Let's take this into the marketing realm with something I hear all the time from Shopify store.

• Resolution: I'll write a weekly email for my email list.
• Experiment: By writing a weekly email, my store traffic will increase.

Miss a week with the resolution and you beat yourself up or make-up excuses why it was "okay" to skip a week.

Miss a week with the experiment and that means that week's traffic data won't be counted.

Experiments do have an additional step at "the end" that resolutions don't. At the end you need to do a bit of simple analysis. What were the results? It could be a confirmation of the idea (e.g. weekly emails created more traffic), disproving the idea (e.g. weekly emails didn't help traffic), or not enough data (e.g. missed too many weeks).

Notice how negative the resolution's end result can be (failed or broke the resolution) vs the experiment (not enough data). With the experiment, you are more likely to adapt the plan and give it another attempt.

Give experiments a try this year. You can even buy a white lab coat on the company card if it helps.

Eric Davis

P.S. These experiments are different than A/B split tests. More five-year-old, less statistician. Though if you want to run split tests too, go ahead as they can be great if you have the infrastructure for them.