Facilitating the Spread of Knowledge and Innovation in Professional Software Development

Write for InfoQ


Choose your language

InfoQ Homepage News How to Grow Teams That Can Fail without Fear: QCon London Q&A

How to Grow Teams That Can Fail without Fear: QCon London Q&A

This item in japanese

Blameless failure starts with building a culture where failure is acknowledged, shared, investigated, remedied, and prevented, said Emma Button, a DevOps and cloud consultant, at QCon London 2019. Visualising the health and state of your system with CI/CD practices can increase trust and ownership and invite people to help out when things fail.

To help us avoid the common pitfall of looking for blame, we need to embrace failure and acknowledge that even the most awesome technical engineers working on the most elegantly architected systems will have to manage failure of software and systems, said Button.

Button used the term "brutal transparency" to describe the level of honesty and visibility that we should provide in the work that we are doing. Using inclusive and open communication methods, even when we are chatting and collaborating online, helps to provide visibility into our activities, she said.

When you’re working through an incident in a production system, it can sometimes feel awkward admitting publically to your peers or managers that there is something wrong, or that you need help - but honesty is always the best approach, said Button. Don’t try to hide the scale of a problem away in a private conversation.

CI/CD practices can help to build a culture of trust and ownership within teams. Button mentioned that visualising the health or state of your system or your processes helps to spread ownership. Having a screen to display your CI/CD pipeline is the first step to inviting everyone in to help out when things fail. Button said, "Don’t get me wrong, it takes more than a green/red screen to encourage shared ownership of a build failure, but it is a good first step."

Button mentioned that building short sharp test suites and short sharp build and deploy processes means that if things go wrong, you can remedy them quickly and easily; teams using continuous delivery tend to be happier taking risks because they know that they can correct mistakes quickly.

InfoQ is covering talks from QCon London 2019 and spoke with Emma Button after her talk.

InfoQ: What can organizations do to prevent that people look for blame?

Emma Button: My talk at QCon covered a few key areas that can help organisations to stop pointing a finger of blame and instead work on preparing people to be ready to respond. Blameless failure starts with building a culture in which failure is acknowledged and shared, and encouraging everyone to contribute to investigating failure, remedying failure and preventing failure. Monitoring tools help us to predict and spot failure. CI/CD tools help us to visualise failure and share ownership. Chat tools help us investigate and collaborate, and some of the scrum, agile and lean processes help us learn from our shared mistakes.

InfoQ: What can you do as a leader to help foster a culture of experimentation?

Button: Finding time in everyone’s busy schedules to learn and experiment can be hard, but the rewards are amazing. With some of the teams I’ve worked on, I’ve been able to offer everyone a few hours every week to spend working on their own technical projects to learn, innovate and experiment with new technologies - simply giving an enthusiastic developer a free Amazon Echo and some AWS credits can result in magical innovative new creations!

In other teams, I’ve set a team challenge - something that isn’t part of their everyday job but which gives them an opportunity to try new technologies in a reasonably "safe" environment; with my last employer, I encouraged a team of people to build a serverless single page application for our recruitment programme so that they could learn cloud technology skills and leadership skills - it wasn’t part of their day-to-day project, but definitely contributed something of great value to the company and they all learned new development and deployment tools.

InfoQ: How do you build psychological safety in your teams?

: Regarding psychological safety - this level of team safety can be built in team retrospectives. Even if you’re not running Scrum, reflect regularly; there are things you can do in your retrospectives to make the environment safer so that everyone feels comfortable providing honest and constructive feedback.

For example, teams who are finding it hard to talk honestly in front of each other could use an online tool which makes things feel a little more anonymous until they get used to sharing honest or tough feedback. Remote teams often use an online tool to conduct their retrospectives. Even if you are not remote, use an online tool to capture the talking points and talk from there. This gets over the first hurdle of people being brave enough to raise their concerns.

InfoQ: What have you done to reward people when they show positive behavior?

Button: I used to be very timid and would never give praise to my peers out loud; working for a large US company helped me to overcome that reservation and I learned to celebrate the great work of those around me. Reward needn’t be financial or a physical gift; sometimes, giving someone a very public "thank you" in front of their teammates really helps to lift someone’s spirits. Something I’ve experimented with in the past is a "thank-you wall" which is a public place where people can share positive feedback for their peers in a public place for others to learn from and celebrate.

Rate this Article