Skip to content

Split Testing in Real Life: Data-driven career decisions

Kai Lovel
Kai Lovel
4 min read
Split Testing in Real Life: Data-driven career decisions
Photo by Suzi Kim on Unsplash

There is a question I’ve heard many times before. I’ve pondered it with high-school mates, work colleagues, neighbours, and regularly with myself.

It’s a question with no correct answer, an unfathomable number of possibilities, and an often-changing response.

It’s not something I can answer for you, nor can you answer it for me.

“What should I do with my life?”

Many people act as if they expect the answer to this question to arrive magically. I think this is rarely the case. But I figured out through an unlikely means that there’s a way to get closer to an answer.

It was when I was doing work with Vedi (then VetDB), and I was introduced to split testing. Let me explain.

Split testing: a brief explanation

Also known as A/B testing, split testing is like a friendly competition between two versions of a website or app.

Let's say you have a website that sells awesome, handmade dog toys.

You're not sure if you should change the colour of the "Add to Cart" button from blue to green. You may have a hypothesis of what colour (or headline, or image etc.) may be best, but you can’t know for sure until you test it.

With split testing, you can split your users into two groups, and then show one group the blue button and the other group the green button at the same time.

The performance of each group is then compared to see which version does better. This can be done using a variety of metrics, such as click-through rate, conversion rate, or time on page (and tools like Google Optimise and Optimizely help you do this really easily).

Split testing is a useful tool for optimising websites and apps because it allows businesses to make data-driven decisions about changes, rather than relying on assumptions or gut feelings.

It acknowledges that we can guess, we can hypothesise, but we can’t know until we test.

But here’s the thing.

The question, “What’s the best version of this button?“ is kind of the same as asking “What should I do with my life?” in the sense that there is no way of knowing without trying.

So when you’re scratching your head wondering what you want to do with your career, it’s a sign to get testing.

Of course, you can’t spin up an experiment in Optimizely for a real life test. But you can take the principles of split testing, and apply them in your world.

I’ve put a few steps together to help us do that.

Split testing in real life

1️. Use curiosity to determine your options.
To work out what career options to test, think about what’s recently sparked your curiosity. What rabbit holes have you gone down? What opportunities sparked your interest when they crossed your desk? It could be a part-time job, internship, side hustle idea, online courses / community or something else.

Rather than having to pick the thing you’ll ‘do for the rest of your life’, instead adopt the mindset that you’re just picking something to experiment with in the short-term. This reduces the barrier to jump in, mix it up, and maybe give that ‘crazy career idea’ a crack.

So once you’ve picked a couple of potential pathways to test, how would this look practically?

2️. Set up the test
There’s a few ways to do this.

You could do two internships, one following another, and compare the options that way.

Or, ideally you would split up your days or weeks to fit multiple test options in (ie. 2 days/wk at Job X and 3 days/wk at Job Y). This allows you to compare them virtually at the same time. Until we can literally work two jobs in parallel universes, this is the best option!

The key is to try and keep the timing of testing each option as close together as possible (max a few months). You want to eliminate the potential for other inputs or variables from influencing the test, such as environmental or personal life factors.

3️. Choose the key metrics you’ll measure.
This is how you compare the options against each other at the end of the test. Some ideas:

  • how much you look forward to ‘Mondays’, on a scale of 1 (dread) to 5 (excitement)
  • your energy levels at the end of the day

You might also look at how each option matches up to your personal values (check out the Personal Moral Inventory from Atlassian’s Dom Price for more on this).

Whatever the metrics, they should be based on what matters to you in your career and life.

4️. Go forth, test, and reflect.
Stick at it for a while. Get past the honeymoon phase, the teething problems, that one tricky project or hard conversation, and sink into the day-to-day. Measure often, reflect, and summarise the results of the test.

In Summary

A career split test begins with curiosity, it’s validated by testing, and your path forward can be informed by the results.

You might run this sort of split test every 2-5 years, to calibrate your new interests and abilities.

However you use it, it’s a powerful way to find more career clarity without the guesswork — or waiting for that “dream job” to magically come to you.

Get testing!