Listen to the article
What is continuous improvement?
A team in any organization can only become more efficient by improving, and not by waiting until they realize things are beyond bad. For a team to output truly a good product, it’s important to take frequent inventory of what is working well, what isn’t, and how to improve the aspects of the team that could use improvement.
Why does it matter?
In our experience, teams have control over many more aspects of their day-to-day than they generally realize.
- They have the power to work together.
- They have the power to say no.
- They have the power to produce a good product.
The only way to become better at anything is through practice, and by practicing good habits and proper communication as a whole, teams can fulfill their goals of producing more work with fewer defects and less stress.
What are they exactly?
Agile metrics are data points that express how well your team is performing at producing work. They will vary slightly from team to team, but there are many generally accepted metrics for various types of teams.
How do I measure them?
First, your team needs to define the metrics that they feel need to be monitored. The chosen metrics should be multi-faceted, to hold different levels of the organization to provide insight into how one part affects another. Four commonly reported-on parts of a team are business, product, engineering, and support.
- Business creates themes and initiatives for the team
- Product creates and maintains the roadmap, setting priority that will result in a product that satisfies the themes and initiatives of the business
- Engineering builds the product
- Support monitors inbound defects and reports them to product
What should I measure?
It’s better to track any metric than not track at all, and may vary depending on your industry, business, and agile methodology. That said, here are a few to consider:
- Business: Percentage balance of work on themes and initiatives
- Product: Lead time and cycle time for features
- Engineering: Lead time and cycle time for stories
- Support: Average age of bugs
Each of these reports on how well the team is performing as a whole, or the process will break down. For example, if the business does not clearly define goals, then product will not be able to properly organize a roadmap. This will lead to context switching with long lead and cycle times, and the number of defects will increase over time because nobody is focused.
The important thing to keep in mind is that it’s not a race to game the system to improve your metrics, rather the metrics are a real-world representation of how well your team is performing, by attributing a number to your performance in different areas.
How can I use metrics to continuously improve?
Start by tracking a few metrics in a spreadsheet over time to see how the numbers change. Then, use those numbers to make some improvements. Little by little, the team will have more well-focused time, which will free up resources to build a self-maintaining dashboard that’s integrated with the organization’s tools.
Who should I show them to?
This is a short story that represents a long journey. The organization needs to be ready for improvement. They need to want it – and they need to understand that they need it. The road to continuous improvement could be months, or even a year away.
Start out by putting a meeting together that includes one key person from each of the parts of the team – someone who represents each of business, product, engineering, and support. Bring in someone that all the teams report to for a monthly meeting to have an open discussion about what’s going well and what isn’t.
How can I show those people my improvement over time?
The best way to show your improvements to the executives and your teammates is by creating a continuous integration dashboard. This dashboard will integrate with your organization’s project management tool and update at some frequency. It should show your metrics week over week, and should roll up through different parts of the organization so that the metrics can be seen at different levels of granularity.