Engineering at Ro isn’t constrained by traditional agile as we are keen on experimentation. Most recently, we’ve tried several different variants of stand-up, and we’re excited to share our results with you. The stand-up experiment was particularly interesting to us because we are a distributed team and were curious if an alternative to the classic model of stand-up could yield better results. Let’s take a look at a few different methods we experimented with and the results of each one.
Traditionally, standups take place in a conference room or huddle area with all the team members physically standing up in a circle. One person starts and then the rest of the team gives a status update in order around the ring.
Our team is mostly remote, so many times we’re not standing and we’re in no particular order. Using the Moderator and Popcorn methods below, we’ve tried to emulate the round-robin style in a remote context.
This version required one person to act as the moderator and call out the next person in line for a status update. This created a problem because often people on the line weren’t ready when they were called or would give their update and then tune out. It also added an extra burden for someone to “run” the standup.
In the popcorn style, the person who just finished their update is responsible for calling on who goes next. This was useful for a while because it kept the team more engaged because they had to track who had and hadn’t had a turn yet. But with a medium-size (15 members) team it became difficult to track who was left, causing minor confusion to net out the value it originally added.
What worked best for our team was to toss out the idea of going person-by-person and instead focus on the work. We now review tickets, not people, one at a time via a screen share. We look at tickets in Jira from right to left, omitting anything that was completed or hasn’t been started.
The major benefit of this is that tickets in limbo don’t get stuck. Sometimes an engineer might assume that a tester has taken it, or a tester may assume that an engineer is working on a fix and then nothing gets done. By looking at the work, we know immediately who is responsible for the next steps. We are able to easily identify blockers by checking where a task is in the flow.
Instead of the typical questions “what did you do yesterday?/what will you do today?” we’re asking questions like “what progress has been made on this ticket since yesterday?” and “How can we move it forward?”
This method creates a lot of engagement and participation from everyone on the team.
For in-person standups without a task board to look at, round-robin still seems to be king. For remote standups and screen shares, ticket focus is our new favorite.