Why Most Growth Hackers Should Stop Experimenting
Experimenting is a waste of time and resources.
… in many occasions.
“Experiments” are a part of many definitions of growth hacking, while in reality, the true definition of growth hacking is “having a focus on growth”. The way you achieve growth should only be supporting the end goal.
The big problem with experiments
The big problem with experiments is that they take time - if done correctly:
- Carefully designing an experiment with success-/fail criteria, a hypothesis, and an action plan.
- Executing an experiment so that you properly split the audience between the different variations, not making too many changes between the variations, and making sure that the viewers of the alternative version don’t notice the glitch of their alternate universe.
- And analyzing the outcome, including following up with next steps.
Most of the time… you cannot justify that time spent on the experiment for the lost potential in growth.
Because for most of you, you will still have enough low-hanging fruit to grasp first on your journey to faster growth.
Imagine this scenario:
You are an e-commerce webshop, and your product page does not have a buy button. What do you do?
Obviously, you just add it.
You wouldn’t set up an experiment for this, would you? Because you could be 99% sure that it would improve your conversion rates.
But at what percentage do you stop acting and start experimenting? Probably it’s way too soon.
Growth ≠ Experimenting. It’s about taking steps. Growth Steps.
In the time spent designing, setting up, and analyzing your experiment for one growth step forward, you could also have made two or three steps forward.
Look at your backlog. Now back to me 🐴🦸🏿♂️🚿
Probably, your top ‘experiments’ should just be action points.
Only when your next growth step could potentially hurt you more, should you run an experiment.
For example, when you improve the hero title on top of your homepage to something different, one might be much better than the other.
… but consider, when you hear a lot of complaints about your header being super unclear, that it might be better to just do it ✔️
It’s about your gut-feeling
Most probably, you’ve included something of a ‘Probability’ criteria in prioritizing your backlog, if you’re using a P.I.E. or I.C.E. variation.
I do.
I always include a question of ‘How sure are we that this will work? (yes=5, no=1)’.
If you give this experiment a ‘5’ because you’re super sure that this will work… you probably shouldn’t test it, but consider just implementing it without testing. Sometimes progress is more important than limiting your results.
Growth is your responsibility, and sometimes you get there faster if you don’t double-question everything.
Especially for startups
If you’re a startup, you probably should just be executing.
Not just because an experiment might take waaaaay too long, because you have less traffic.
But most probably because putting your first campaign live is more important than AB testing that first campaign. And the same goes for almost every first thing that you build: something is better than nothing.
As a growth practitioner, you should keep the company moving.
Coming up next: How to make experimenting your superpower 😏
This article came to life thanks to a call with Mohammad Alselimi & today’s LinkedIn post by Ramli John… about the opposite 😅
