The power of experimenting

Chris Atkinson is a Delivery Coach here at Atom. In the first of our monthly deep dives into the world of our Technology team, he shares why experimenting can help you find the perfect solution.

In the fast-paced world of technology, the ability to experiment is not just an advantage, it’s a necessity. For tech teams, embracing a culture of experimentation means continuously testing new ideas, iterating on solutions, and learning from both successes and failures. This agile approach allows teams to adapt quickly to changing landscapes, optimise processes and ultimately deliver more effective and impactful technological solutions.

We live in a world of uncertainty, and organisations are becoming more and more complex, adaptive systems. As per the Cynefin framework, work in this domain requires emergent thinking — there is no ‘best practice’. Atom bank’s challenges, especially in product development, are unique and need a “probe–sense–respond” mindset.

Navigating this uncertainty requires an experimental approach, where the only ‘best practice’ is that of embracing trial and error, running experiments and continuously learning.

How often are you running experiments within your team, value stream or as an individual?

When was the last time you experimented with your ways of working, processes, social contracts, or communication methods?

Here, I’ll cover some simple tips to help guide you to define, run and measure experiments:

Step 1 - Identifying a problem or opportunity

When you spot a problem, or an opportunity, it can be tempting to just go right ahead and take action to fix or implement it. But, unless you’re in an emergency, it’s worth taking a small step back and asking yourself a few questions:

What is the problem you are trying to solve?
What tells you it is a problem?
What data do you have or can collect that validates the problem exists?

Answering these questions will help you frame the problem you want to solve, and establish a baseline that you can measure later to see if you have solved the problem.

If you’re going to need help from someone else to solve the problem, data is the difference between fact and opinion. The more you can quantify a problem, the more rational an argument you can create in order to drive change and gain support from those who you need to achieve it.

Step 2 - Understand the root cause

Now you have framed and quantified your problem, try and get an understanding of the root cause. Fixing symptoms of problems might get you some benefit, but never as great as addressing the root cause of a problem. You can use techniques like 5 whys or Ishikawa (aka Fishbone) to understand the root cause(s) of your problem, and Pareto principle to prioritise which of them to go after.

Step 3 - Ideate solutions

Now you know the root cause/s - how can you solve them? Create a list of solution options and prioritise them in terms of feasibility, viability and desirability. It’s probably a good idea to go after elements that are low cost, easy to implement and easily undone if it does not work.

Step 4 - Create a hypothesis statement

So you have framed your problem, have an idea of the root cause and one or more solutions to put to the test — now it’s time to hypothesise. What do you think will happen when you implement your solution, how will you know it is working? What can you measure to validate this? (hint: that should be the same data that is telling you about the problem).

You could use a template like this:

Due to <root cause+problem+data>
We believe that will
We will prove this by

Step 5 - Go!

Run your experiment, remove things that don’t work and make permanent the changes that do. Don’t be afraid to restart at step 1 and repeat the cycle until you find a successful solution.

If you’d like to learn more about the exciting work our team is doing, visit our dedicated technology pages.

Or, if you’re interested in joining our thriving technology team and helping us find the next perfect solution, check out our careers page for open vacancies.