Design your own customised experiment flow to avoid mistakes

Want to test your project without wasting too much time testing? Follow the principles of experiment flow.

David J. Bland, the author of “Testing Business Ideas”, tells you how. To start with, testing or trying out new business ideas to minimise uncertainty is no easy task. Teams often apply the tactics of experimentation without comprehending the principles behind it. They follow certain practices because they seem easy to replicate, but… this can be dangerous, warns Bland.

Think about the automobile industry. Manufacturers tried to mimic Toyota’s Andon system: a cord (later replaced by a button) which workers would pull whenever they noticed a serious problem, the aim being to stop the assembly line and alert a supervisor. That way, they could prevent the fault from moving on and becoming more difficult to repair further down the line when more pieces had been added to the car. Removing those components to get to the bottom of the issue would have meant hours of work and production gone to waste.

Other brands thought Toyota’s idea was great and decided to implement it in their factories. Disaster! Employees would detect a fault and pull the cord, but their supervisors would tell them off for interrupting production - the factory had a target to meet and stopping the assembly line meant that the numbers would not add up. Ultimately, workers would avoid pulling the cord so as not to get fired.

Why didn’t that happen at Toyota? Because they had shaped the practice according to the jidoka principle, which loosely translates as “automation with a human touch”. This concept is essentially written into the Japanese firm’s DNA. At Toyota, every employee knew what to they had to do:

1. Detect the mistake.
2. Stop the assembly line.
3. Solve the cause of the problem with their supervisor.
4. Implement lessons learned in the workflow.

Similar scenarios have occurred with experimentation tactics, where teams have copied what other companies did without understanding what lies behind those actions. They could benefit from grasping the principles of experiment flow.

• Visualise your experiments

Running a test can be a tricky business unless you visualise it first. To do so, simply write down the tests you plan to carry out (in a note book, on a blackboard, on the wall, in an app…). Draw a table to help you track progress and see when things are getting stuck. It could look like this:

Pending

Planned

Underway

Outcome

Now, all you have to do is add your experiments to the first column in the order you intend to do them. Put the tests you plan to run first at the top.

Pending

Planned

Underway

Outcome

Online campaigns

Customer interviews

Landing page

Survey

Limit the number of experiments you have underway

Do not try to tackle too many tests at once, especially if your team is small - you will only end up frustrated with a string of unfinished experiments. Set limits in your table and do not let another test start until the first few are moving forward.

Pending

Planned

Underway

Outcome

Online campaigns

Landing page

Customer interviews

————————-

Survey

————————-

————————-

Continuous experimentation

When experiments start flowing with the team, it is time to focus on action and reconsider the way you work. You might want to dig a bit deeper when a problem arises or change the limits on tasks that are already underway. Once you understand the principles of experiment flow, do not be afraid to alter your methods: you will get better tools to make the most out of every test. For example, your table could change to look like this:

Pending

Planned

Underway

Outcome

Survey

Online campaigns

Landing page

In progress

Customer interviews

On hold

Remember, the goal of the experiment flow is to help you reduce uncertainty, so… good luck!

Source: https://www.strategyzer.com/blog/the-principles-of-experiment-flow

Photo by ThisisEngineering RAEng on Unsplash.