In July, I gave a talk at SDX 2017 in Portland, OR, on scaling support. My talk was about my experience working through dismantling my very large 18-person team prior to going on maternity leave. I touch on this in my post here. The following is my transcribed talk, along with my accompanying slides (but really, the talking part is the significant part – you don’t need the slides to understand).
Breaking up the band: Scaling support
A story in 3 parts.
Hi, I’m going to talk to you today about scaling support. Specifically, what to do once your team has grown beyond a manageable size.
When I was in kindergarten, I had two best friends – Adrienne and Sarah. We did everything together. We sat together, we played together, we went to each other’s houses and had sleep overs. When it came time to go into first grade, the school decided we needed to be in three different classrooms. We were devastated. My mom explained to me that we a clique, and we were excluding other kids. By splitting us up, they knew we’d develop more friendships, without compromising our relationships with each other. Ultimately, they were right.
Let’s fast forward 30 years, to January 2016. I’m a team lead at Automattic, and my Happiness team, Hermes, is 18 people, stretching from Minnesota, USA, to Davao City, Philippines. Besides stretching across many timezones, we featured Happiness Engineers from as far north as Lithuania and as far south as South Africa.
The team worked well together despite the inevitable problems that come with working across so much of the globe – we had weekly team chats that some people would have to call into after dinner or in the middle of the night; half of us complained about the heat while the other half complained about the cold; weekly or bi-weekly 1:1s were challenging to schedule; meetups were difficult to coordinate in a way that was fair to all the travelers; often at least part of the team was traveling so far that jetlag was a significant factor. Our team worked, but we were putting a lot of effort into making it work. We communicate in three main ways – in Slack, Zoom, and on our internal blogs, called P2s. That ends up being a lot of communication streams for each additional person you add to the team.
Ultimately, it was time to break up the band. Although it was a great team, the extra energy being funneled into just functioning as such a large band of people was no longer reasonable. I told the team that we would take the spirit of the team and spread it to other teams, just as my friends and I in first grade would grow new friendships.
I was planning on going on maternity leave at the end of June for 5 months, and that gave us just about 6 months to plan, coordinate, and execute. And that turned out to be about perfect.
Part I: How we did it
Step zero was accepting that the team really had to be divided up. Originally, when I knew I’d be taking leave, I had imagined that a temporary lead would work, but the longer I spent with the idea of breaking the team into smaller, geographic teams, the more I preferred that approach. Besides problems with coordinating across time zones, it made much more sense to divide the team up and mentor the leads, rather than try to hand off such a large team to a brand new lead.
The next step was a series of team discussions. During our weekly meetings for a few weeks I made this an agenda item. We talked about why it was going to happen (not if), and I asked the team to think about how they’d like it to happen. Getting buy-in from the team was important; no one really wanted the team to be broken up initially, so being able to control some aspects of the division helped everyone approach it positively. Just as I, as a little girl, was upset to hear I would be in a class without my friends, but with time, I grew to be excited about the new school year anyway.
We had an in-person meetup where we talked in-depth about who would be on each of the three new teams, and went through some exercises with those new squads. We had performance and long-term Happiness goals we were working on at this meetup, besides the team split, so the squads took the chance to work together and come up with solutions independently that we then discussed as a group.
The three new squads of 6 people each were divided up broadly geographically. One surprisingly useful exercise that the team did was writing their names down on post-it notes, and sticking them all up on a wall. Then taking turns grouping the post-its in different configurations. While we wanted the teams to have a strong geographic focus, we also wanted to have a mix of seniority and skills. We settled on three groupings early on, and started working on exercises.
We decided that we’d operate six weeks with the squads-within-the-team acting as independent groups. Each squad decided how to split up squad-lead duties, like prompting weekly updates and leading squad chats. Likewise, each squad tracked their own performance and reported it weekly. They also set their own squad meeting times. We converted the old team meeting into a 15 minute recap of their weekly update, but eventually we spent the 15 minutes getting some facetime and just chatting. The only thing that wasn’t settled on was the leads.
At the six-week mark, we decided to extend the squad experiment until it was time to make the permanent split. Things were working really well, and everyone on the team was comfortable with their squad. Communication was active and performance was still high. Working more closely with people in their own time zone, having team meetings during their workday, and the camaraderie of the tighter group made the members of the squads feel connected to each other.
From the time I knew we would need to do three teams, I worked with my lead to find them. One was very obvious – he had been my right hand on the team, and was perfectly placed to take on a squad. For the other two we had a lot of great candidates to choose from, and we took the squad experiment as an opportunity to encourage some folks to develop stronger leadership skills. For some, that was increased communication, or fostering and drawing out skills in teammates, or taking more initiative for things that needed doing. Once we had the leads sorted out, and they agreed, I told the team about our decision.
All three were new leads, so the entire final month before my leave, I set aside for the hand-off and mentoring. The first two weeks I asked the new leads to sit in on ticket reviews I’d conducted, and also on routine assessments. About halfway through these assessments, I asked the new lead if he or she would like to lead while I assisted, and they all lead the last one solo while I followed along. This gave both the new lead and the Happiness Engineer time to adjust to each other’s style and understand each other’s strengths and goals. The ticket reviews covered the HE’s previous 3 months of work, and drawing from these show the HE’s growth over time and how they relate with our customers. So by bringing in the new lead, they gained the benefit of this work and could start their relationship with the HE at the outset with insight.
During the next two weeks, I talked frequently with the three new leads, and gave them support in any way I could. This was my time to take a step back, be hands-off with the front-line HEs, while still acting as something of a safety net for the new leads. Sometimes support for the new lead was specific to their new team, and sometimes it was more general about being a new lead at Automattic. We had a private channel on Slack, where they could ask questions about the role, for my opinion, or anything else. By the end of the two weeks, they were ready to be on their own.
Part II: Why it works
Dividing up the team into squads worked for several reasons. I’ll come back to these points again at the end. It not only was infinitely more manageable for new leads, it created more lead opportunities – as our Happiness division continues to grow, we see more people who are a great fit for leading.
At first, the team and I discussed creating two new teams from Hermes, but eventually we decided on three as it made even more sense – each of the new teams would be able to grow for quite awhile. The time we spent in discussion was vital as the process continued, because everyone was sure of the end goal, even as it evolved, and was committed to it.
Feeling like we were still a unified force was important for moving through the process together. We did a trial period with squads because we wanted to make it easier to shuffle folks around if they felt that their current squad wasn’t the right fit for them, but the squads we agreed on at the meetup ended up not changing at all. In essence, we were already three separate teams, but with one name.
One unexpected result we discussed was that having a smaller group made everyone feel more accountable for support goals, and more aware of how vacation time affected the rest of the team. It was a small revelation, but important for confirming that the move away from such a large team into smaller teams was the right move.
The biggest factors that made the division successful however are as follows:
The split went well. We didn’t have any significant or lasting resistance. We didn’t see a performance dip that you might expect after such a big change to team structure. It worked because the team had buy-in. They had some measure of control over their future working environment.
We also had time to make adjustments before committing to a particular team structure or lead. We could show that each squad was operating effectively before ever officially dividing the team up. The squads had flexibility to make decisions as a unit, and they maintained a close-knit sensibility.
Finally, the new leads were given a lot of support. They got to co-lead with me for two weeks, then lead with my support for another two weeks. They were able to interact with the leads group from the outset, and benefit from those interactions as well. By co-leading with me for the first two weeks, they could see how I handle specific things, like 1:1s, support reviews, and the assessment process, and start developing their own style right off the bat. It gave them a chance to interact with their team members and get to know them before everything became finalized as well.
Part III: How it can work for you
- Give your team adequate notice to absorb the information. Make time for questions, but also let them get comfortable with the information at their own pace.
- Give your team the ability to make decisions. Anything that you can delegate, let the team own. Take a step back and see who naturally steps up.
- Choose leads you believe in, and give them plenty of time to acclimate and develop skills. Adjusting to becoming a lead can be really hard, so be sure to provide support after the hand off.
- Try to establish a reasonable timeline – long enough that everyone can adjust, but not so long that it feels like the thread is lost. For us, 6 months was just right.
- Communicate every step of the way, even if it feels like you’re repeating yourself a lot. I have many posts written about what we were doing, and what was going to happen. Make sure to ask the team members during 1:1s about it so you can talk to them about any apprehension they may have.
- When you hand off a team member to a new lead, really hand them off. Don’t undermine the new lead or confuse your former team member by continuing to act in that team lead role. Wrap up your 1:1 sessions with the team member and explain what will happen to the information they have shared with you. In general, career discussions and any items you have worked to improve together, as well as strengths and victories will go to the new lead. Personal matters will be left to the team member to divulge to the new lead when – or if – they feel ready.
- Make sure that everyone knows what will happen on the first day of their next chapter. You don’t have to decide what that is, but be sure that the leads know, and that they are communicating that to their teams.
If everything is coordinated carefully, with ample time and support, then everyone can look forward to their next adventure with excitement. I know that by the time I started first grade, in a class away from my best friends, I was ready to find out what came next.
One response to “Breaking up the Band: Scaling Support”
[…] most of Eshu, the WP.com team, are from my old team, Hermes. Before my maternity leave in 2016, I broke my enormous 18-person team into 3 teams of 6. Eshu was one of those newly-formed teams. Eshu is now 12 people, so alongside the familiar faces, […]