Challenge(s)
I guess I faced two challenges this week. The first is I completely forgot to post last Sunday. 'Twas a bank holiday here in England and it slipped my mind -- oops!
More seriously, one of the things I grappled with this week was kicking off the growth experiment process.
With growth experiments, it's easy to fall into fairly obvious traps.
The trap I want to focus on today is the lack of benchmarking. Hypothesising and benchmarking are at the core of every experiment. Failing to specifically define those will turn otherwise decent tests useless.
So, what's the challenge?
Well, as with anything else in life, growth experimentation has to start somewhere. The challenge I faced this week was getting back to the roots and figuring out (once again) how to avoid the dreadful trap.
Decisions
Whether you are launching a new product, a new channel, a new strategy, or a new tactic, you're often starting from zero.
This may be scary to the average growth marketer. If you start from zero, what do you grow? What's your hypothesis? What's your benchmark?
It's zero. And this realisation may freeze you into place. So here are the decisions I made this week:
1. Start with benchmarking
You gotta start somewhere.
Let's say I'm hoping to work on my activation lever. A good place to start would be setting up a series of onboarding emails that will teach new signups to use our platform.
What should my open rates be? What should my response rate be? What should my activation metric (whatever that is for us) look like after one, two, ten emails?
No one knows.
So I'll start there. Set up my emails and define this as a benchmarking experiment.
2. Hypothesise, still
Hey, I'm not out of the woods yet buddy. IMO, I still need to hypothesise on the results of my benchmarking experiments.
"But you said..."
I know what I said.
I'll take your best guess. Define what success would look like for us and go with it.
Turns out I grossly overestimated the results you'd get? Good, I've got a goal to work towards.
Went too low and smashed my expectations? Time to challenge the team.
This process should free me from the mental freeze I sometimes get. Where do I start? Anywhere. Just start.
Experiments
I'm backlogging a bunch of experiments this week. Nothing worth sharing yet, though. I'll give some insights on this later on, particularly the prioritisation method I like to use.