Make everything an experiment

If your team can't agree on a course of action, conduction a short-term experiment instead.

The best course of action isn’t always obvious. Should you use an RDBMS or NoSQL database? Should you use Scrum or Kanban?

Sometimes the best course of action may be obvious to you, but you can’t convince others on your team.

In either case, there’s a simple tactic that can help break the deadlock:

Experiment!

Conduct a short-term experiment, using one of your proposed solutions. After a short trial period (perhaps a week, perhaps a month), re-evaluate, and determine next steps. If your chosen database isn’t turning out as great as planned, it’s easy at this stage to change course, with minimal wasted effort.

To conduct a successful team experiment, follow these simple steps, taken from my video course, Ship better code faster:

  1. State the goal. What is the reason for the experiment? Might be a technical improvement, streamlining of meetings, or any other team goal.
  2. Propose an action. What action are you agreeing to do as an experiment? Remember: It’s not permanent. Nobody’s committing to staying this course forever.
  3. Agree to a timeline. How long will this experiment run? 1 or 2 weeks is usually sufficient, although some changes may require more time to meaningfully evaluate. In other cases, an hour may be enough.
  4. Conduct the experiment. Do the proposed action for the agreed duration. Try to stick to the plan, and not abort the experiment early, unless there’s a really good reason!
  5. Review the results. Get the team together and discuss the experiment. What went well? What went poorly?
  6. Iterate. What are the next steps? Was the experimental action successful? Should it continue? Should it cease? Will you try some other experiment?
Share this