Happy July, dear reader. We've been combing through our log files to prepare a view on how our systems are being used by the Kiwi public. Here are our charts and commentary for the second quarter ... more
Written by Trade Me's Agile Coach, Sandy Mamoli
... let teams self-select!
At Trade Me we’re in the process of getting everyone into small, cross-functional teams that will persist over time and across projects. For this we have been deeply inspired by Spotify’s ideas and terminology of squads, chapters and tribes and their culture of autonomy, mastery and purpose.
In this post I’m going to write about how we got to a place where people work in autonomous squads and will focus on the process of self-organisation that we used to move people into squads.
Up until last week we had six established squads within two tribes and the rollout had been purposefully slow and controlled. Now we felt there was enough knowledge within the organisation to speed up the process.
The first thing to do was to define our next target condition and to agree on how to make it happen. For this we used Jimmy Janlen’s Toyota Improvement Theme Kata (see picture below):
We decided our next step would be to hold a “squadification day”. Squadification day is a word we made up and it means “day where we form squads and select who is going to be in which squad”. (That’s on the sticky notes on the picture but a bit hard to read.)
How to select a squad?
The next question was how to select squads. One option was to have management choose who was going to be in which squad. That’s not necessarily a bad idea - it would be mostly right and it’s a quick way to get it done.
However, I couldn’t help thinking about our recent Fedex day. Fedex day had worked so well, see my post “Every day should be Fedex day”. During that day people worked on a project of their own choosing with the people they had chosen to work with and they were perfectly capable of forming small cross functional teams to deliver 24-hour projects.
Everyone was so happy and motivated that I couldn’t help wondering whether we could find a way to allow people to choose what product area to work in and who to work with.
There are of course higher stakes in “real life” than during a Fedex day as squads persist for more than 24 hours and delivery is crucial. But still, the results, buy-in and ownership we got were something that made it seem worthwhile to at least give self-organisation a try.
How to make it happen?
David Mole, our head of projects and I identified three things we needed in order to make this happen:
- Permission: Convince people that this is a good idea and get them to buy into the process.
- Structure: Confirm the number of squads needed, their missions and Product Owners.
- Facilitation: Find a way to facilitate self-organisation without this turning into the Hunger Games.
Every single person’s first reaction was something like “Whooooaaaa! How’s that supposed to work? I’m really not sure we can do this.” But after some time and a discussion about what could possibly go wrong everyone came to really like the idea and agreed to give it a go.
Overall, getting buy-in was way easier than I had expected and we agreed to run a trial in our Auckland office with the prospect of spreading it to other locations if it worked.
On Fedex day people can come up with their own ideas for things to build and while I find it somehow appealing to run a company this way it doesn’t necessarily make business sense.
For now all we needed was a number of “empty” place-holder squads, each with a Product Owner and their mission defined by the business so people had something to sign up for.
Our main concern wasn’t so much that people wouldn’t “get it right”; after all we had had to make changes on two “management-selected” squads earlier, but more the social consequences of selecting people into a squad.
Our greatest fear was that the process could turn into an “ooh, ooh, pick-me, pick-me” contest with some people being left over; much like the new kid nobody knows yet or the kid with no ball skills on the soccer field.
To form functional squads we also had to tell people what was needed and decided to put in place a set of rules to guide the group’s self-organisation. The plan was to let people sort it out and to accept their choices as long as they stuck to the constraints.
According to our rules a squad needed to be:
- 3 - 7 people
- capable of delivering end-to-end
The overarching rule for everyone was “Do what’s best for Trade Me!”
We also came up with a process for how we wanted to facilitate the squadification meeting:
- Everyone turns up to to a squadification meeting
- The product owner presents the squad missions
- We facilitate conversations about how to solve the puzzle of creating three squads
- We keep going until everyone agrees on what the squads should look like
David has made a beautiful drawing of the full process:
Here’s the detailed script:
- Put empty flipcharts representing each squad with the Product Owner's name and squad mission on the wall.
- Put up a flipchart with the rules and constraints.
- Introduce the process and emphasise the guiding rule of "Do what's best for Trade Me".
- The Product Owners present their missions to the group.
- Everyone writes their name on a sticky note.
- Round 1 (time box 10 mins): As a group put stickies with names on squad flipcharts. Try to make it work for Trade Me.
- Together assess all squads against the constraints (Can they deliver end to end? Would it work?)
- Round 2 (time box 10 mins): Repeat the process from round 1 while keeping the constraints in mind
- Round 3 - x (time box 10 mins): Repeat until it's solved and there is consensus this will work for Trade Me
We knew that there was absolutely no chance that any company going from ad-hoc project teams to squads just coincidentally happened to have the right number of people and skills available. We were aware that Trade Me probably had to hire to fill gaps, and in order to surface and acknowledge these gaps we provided a stack of blank index cards that people could write “imaginary people” on to make the squad puzzle work out.
With the above plan we went into squadification day. I felt well prepared but still a bit nervous. In fact, we were totally overprepared with a long list of what-if scenarios and really nervous ;-)
What really happened?
We started the day off with Lyssa Adkin’s journey lines exercise to make sure everyone in the group at least knew each other a little bit. That was especially important as not everyone had been working together before and we wanted to make sure that people had a chance to form a squad with people they didn’t necessarily know that well.
We all learnt a lot about each person and how they felt about Trade Me and the work they did. Everyone took it very seriously and had some pretty honest stories to share.
Next, before sending everyone off for lunch, we introduced the idea of self-organisation, the reasons behind it, the process and the constraints. We did this because we had learned earlier that people usually needed a little bit of time to get over their initial shock and to love the idea.
The afternoon was focused on the squadification meeting. We needed four rounds before coming up with three squads:
- In round 1 people often put themselves on more than one squad. Many seemed to find it hard to commit to one product area (motors, property or dealers).
That’s when people came up with practical questions such as:
- Will this change reporting lines? (No, this has nothing to do with reporting lines)
- What will happen to current work? (We won’t start squadding tomorrow. We’ll come up with a plan together with you)
- Can I ever switch squad? (Squads are supposed to be stable but we’ll run a retrospective 3 months from now and if people want to switch squad or squads want to swap product area we’ll find a way to make this happen)
- We couldn’t make squads that could deliver end-to-end because two out of three didn’t have the required skills on the squad.
As it was hard to see which skills were lacking on a squad we decided to have everyone write their name on a colour-coded sticky note to show their main skill/role.
- This time people stuck more to being in only one squad than in the first round
- Everyone seemed to want to do several roles. We are looking for t-shaped people but one person as a developer, BA and squad master was pushing it a bit.
- Lots of people wanted to do the squad master role - in fact, more than squads available. (At Trade Me a squad master is a bit like a Scrum Master but as not all squads run Scrum the name seems a bit silly. The role is probably closest to a mix of Scrum Master and permanent Agile team coach. It is also not exactly the same on all squads.)
- Going through the rules we realised we still couldn’t form full squads so as an input to the next round we wrote each squad’s issues on an index card and placed it on the squad flipchart.
- This was when people really started to collaborate and to discuss the pros and cons of each squad composition.
- People started consulting us and each other.
- I felt this was the round where we really got buy-in and everybody started to work things out
- For each squad we checked if the issues from the index cards were solved and if not we recorded them on a whiteboard.
- We had come as far as we could moving people (i.e. sticky notes) and decided to come up with options and scenarios for the issues we had recorded after round 3.
- We finished round 4 with a list of things we could do to make things work.
- As we were incredibly exhausted we stopped here.
On day two we found a solution by moving one developer to a different squad which made it possible to have two full squads and one not fully skilled squad we will need to hire for.
This is the final result:
Each person seemed to be very happy with the squad they ended up with and in what role they would be doing.
After agreeing with everyone we followed up with a Lean Coffee for each of the newly formed squads. This went incredibly well and in 30 mins we got all of the big questions discussed and any remaining fears addressed.
- I was very impressed with people’s alignment with Trade Me’s purpose and the constructive culture. Everyone truly wanted to do what was best for Trade Me and people didn’t even need to be reminded of the rule.
- I thought two days would be plenty but we almost ran out of time.
- It was awesome but also incredibly exhausting. The process needed more active facilitation than I had thought.
- It was really important to have the Product Owner be there for the entire day. (Thanks to Megan Linton for her engagement and help)
- I found it interesting that the pattern of initial shock and subsequent processing and full buy-in for the idea was fundamentally the same for everyone. Thinking about it, my reaction was the same: After writing my blog post about Fedex day I kept thinking about how sad it was that we couldn’t do this in real life. Only after a while did it occur to me that there was nothing stopping me from doing this except for habitual thinking and the completely untested assumption that other people would reject the idea.
What would I do different next time?
- I’d improve the way we visualised squads: Next time I’d use photos of people instead of sticky notes and I’d visualise the check whether a squad has all skills needed in a better way.
- I’d improve our pre-event communication: We forgot to inform a couple of people properly and assumed some things that we should have explained (What exactly is a Product Owner or Squad Master?)
Did it work?
We got two fully-skilled and engaged squads out of the process and people bought into the process and the results.
As a group they came up with solutions we wouldn’t have come up with ourselves -solutions we would either not have thought of or thought they weren’t what a particular person wanted.
There were some hard decisions that had to be made but as people were part of the decision making process and ultimately responsible for the outcome they understood the reasons and knew that this was the best possible thing for Trade Me.
Overall, everyone seems to be happy with the squad they’re on and I have since heard from the Auckland crew that everyone is really excited to keep things moving and really enjoyed the two days. Even the not fully fledged not-yet-squad are more optimistic and engaged.
So, yes it did work!
Would I do this again?
Yes, yes, yes - it rocks!
(And because we’re using so many ideas from Spotify I’d like to thank Henrik Kniberg, Anders Ivarsson, Jimmy Janlén, Joakim Sundn, my former colleague Victor Cessan and everyone else at Spotify for sharing! You guys are an inspiration. Thank you! )