No Estimates at Scale in the US Federal Government
Adam Parker

When conversations about improving our backlog refinement ceremony started, we had no intention of getting rid of points. We did not set out prove or disprove the #NoEstimates debates waged on social media. “We" were the three teams of E-Verify Modernization (EVE-MOD) product group. E-Verify is web-based system offered by US Citizenship and Immigration Services - an agency under the umbrella of the Department of Homeland Security. An experiment had barely gotten underway for refinement before our whole landscape shifted. Our planned work became overwhelmed by mid-sprint interruptions and decisions made in sprint planning became moot within hours. We had to act fast, so we stopped all work for a one entire day. We held a workshop offsite to decide how to re-design our approach to work. Beginning the very next day, our teams switched to Kanban. And with the unrelenting stream of incoming work, we also decided to stop the practice of estimating work and thus began EVE-MOD’s NoEstimates journey.

In this session, you will also learn how EVE-MOD adapted refinement conversations, how NoEstimates continued after a transition back to Scrum, why we eventually returned to estimating, and what this agilist learned during this unique 12-month experience in the US Federal Government.

Learning Objectives
I seek to build healthy, trusting teams and in doing so develop a culture of learning. Experimentation allows for structured learning. This experience impacted me in the following ways:
• NoEstimates can work, but it’s not for every situation
• A reminder that pointing is not mentioned in the Agile Manifesto
• What “inspect and adapt” and "being agile" looks like
• The benefits of stable, feature teams
• If you can experiment in the federal government, you can (and should) experiment everywhere.
Session Type
Experience Report