Shaking up remote retros

They put me in charge of a team retro, the fools.

I didn’t realise until recently that the remote-first way of working can be pretty unusual even in the dev industry – I’ve never done an in-person stand-up, kick off or retrospective, and so it’s not at all unusual in my mind to do these things over video. We do all of these things through online meetings where the leader of the meeting shares their screen, and we all follow along. It works really well for us, for the most part, so it’s easy to forget that most people do these kinds of team activities face to face with their entire team present in the room.

This post is about our remote retros, and how they work. We used to use google meet for our video calls, but we now use Zoom since it doesn’t make our laptops overheat when there’s many people onboard, and we can all have our cameras sharing at the same time without affecting quality. We’re not on the paid plan though, so our meetings are capped at 40minutes in length (we exit and start a new meeting, which is a pain in the butt but just goes to show how good Zoom is otherwise) but it has the awesome ability to record meetings as a video from within the application. This means we can easily record important meetings for those who can’t make it, and no one has to miss a thing!

Ok, so we use a video call – so far, so obvious. There’s more than that to a retro though. Luckily, there are fun and free apps out there for putting together a shared online retro board, like having a whiteboard or a wall for sticking post-it notes on in real life (I assume in-person retrospectives do something like this, anyway xD).


An example of one of our retros

Until recently, we were using FunRetro to do all our retrospectives, and so far as I know the other teams still do use it. It’s an online board that allows everyone with the link to join simultaneously, edit and add topics, and then vote on them together.

It’s definitely a good tool for the format of retrospective it’s designed for, but can get very samey if you don’t shake it up and change out the columns regularly.

You can name the columns whatever you like, and have as many as you like. For us it came down to four categories: happy face, sad face, ideas and actions. We would start the retro by going over actions from the previous retros, which we keep as a to-do list in confluence. (Which everyone forgets exists until we are named and shamed for not having done whatever task we definitely had the best intentions of doing last retro.)

Then we had a set mumber of minutes, usually 2-5, to list some things that went well or badly, or things that made us happy or sad since the last retro. Happy face, sad face. If we use an emoji for it then it must be FUN, right?

We each got two or three votes per column (you can change the number of votes allowed per person), and we talked about the issues one by one starting with the ones with the highest number of votes. The issue with this was that our team as it stands in retros contains a bunch of devs, a product owner, a designer and a QA. So if you had a serious problem you wanted addressing but it only affected you as the designer, and not the rest of the team, then it wouldn’t be voted on by everyone else and your things would never ever get discussed, no matter how miserable and slow it made your work. I complained incessantly about this because it affected me a lot as the only QA, until we decided to make the retro longer so that we could talk about all issues. We still voted though.

From those discussions, we would get a bunch of ideas as we went along, and then we’d make those into specific actions. These went into the confluence page, never to be seen again but we would take the general gist of them to heart.

We had an issue though. The same person felt like they were put in charge of the retro every single time by default, even though they never had the time to do anything interesting with it, and it often felt like a slog both for the retro leader and for everyone else. It was time for a change, so one of our developers stepped up and said he would shake things up a bit.

It’s not that FunRetro is bad. You can customise the column titles, number of votes allowed etc and probably do some really truly fun stuff with it. We’d just got into a rut.

So we switched to Miro (formerly RealtimeBoard) which is kinda like a collaborative digital whiteboard, on which you can draw, write and add sticky notes or images. It’s not intended as a retro tool, but it’s really flexible and you can have loads of people adding to it at the same time, which makes it perfect for those remote meetings. We had previously used it for creating our guild charter, as well as some architecture planning due to its visual nature.

Matt’s Miro

It was a retro late, but Matt did not disappoint with the new cool format. He split it into five frames, which are mini boards for mind-mapping ideas, plans or anything really.

Frame 1 of Matt’s Miro retro

The first frame was a really quick and simple intro – for weather report, all you do is add a weather-related photo that reflects how you feel about the sprint or time period you’re retro’ing. I’d just got back from TestBash so mine was an endless starry sky with Aladdin singing “a whole neww woooorld”. We love memes. Most of the images were sunny skies with a few clouds, but the designer felt like he was surrounded by fog. It was really useful for quickly finding who was having a more difficult time of it, and was isolated from the team.

Frame 2 of Matt’s Miro retro

The second activity was Story Oscars, with some art by yours truly. I now know that I could have drawn the oscars in photoshop and uploaded them, which would have been infinitely easier and better than drawing them with the limited (for art purposes) pen and colour choices available in the application. But there we are. xD

We used post-its to add stories from JIRA we thought were the best, the most annoying and the biggest bang for buck. When the time was up, we voted using whatever symbol we liked – so I quickly drew up my own face, obviously. It was a really good way of finding out what had been causing us pain in the last iteration. Miro is really good for teams with a lot of personality, because it allows personalisation. We were only asked to make a shape to vote with, and we got to choose symbols that reflect our personalities.

Frame 3 of Matt’s Miro retro. I’ve cut names off, I don’t know why. It’s easy to find out who works here

In Perfection Game, we each gave ourselves a rating for the last iteration, then wrote out some actions we could take as individuals or as a team to make that a higher number next time. I love the differences between how everyone presented their numbers – some with shapes and bright colours, others with plain black text.

We then took those actions, and copied them onto the next frame:

Frame 4 of Matt’s Miro retro. I have amazingly censored a colleague’s face so he is totally unrecognisable. The photographs were on the actual board, and the one on the right is a reoccuring meme in the team

Another piece of Miro art, courtesy of yours truly ❤

The purpose of this activity was to take the actions already written, and decide both how easy they would be to achieve and how much impact they would have on the team. Left is more impact, right less. Top is harder to reach, bottom is easier – hence low hanging fruit.

You can see we had a big ol’ cluster around easy to achieve things that would have a big impact. I don’t know if that’s because we all wanted our stuff there (we did move things around a lot in a democratic way, but you also don’t want to be the person to put someone’s idea in the low-impact zone) or if we really do have that many things we could easily achieve, but either way it would be hard to do all of those things before the next retro. So obviously we voted to find which issues we really wanted to get going.

Frame 5 of Matt’s Miro retro. I feel like since I started censoring then I should continue, but this is getting silly and also I missed some

Matt rounded out the retro with a nice appreciation and action board. We each wrote a thank you to someone who had helped us in the sprint, and then an action for ourselves to follow. It was kept quite open so that we could write anything we wanted in there, from “get more sleep” to “finish tickets in a timely fashion”.

We decided overall that it had been a really positive retro that had brought new issues to light and helped us see things in a way we hadn’t before. We’d kind of unintentionally stopped paying attention to the retros, and this had everyone interested and engaged.

Obviously, it’s a lot of work putting a retro together all the time, so we decided that it would be a randomly chosen person each time. We used a spinning wheel app to decide on the person, and my name was the one that came up. I wasted no time whatsoever, and shirked the last half hour of the day’s testing work to start planning.

Planning a retro

Oh my days, planning a retro is hard. There’s actually tonnes of considerations to make it a good one, and then there’s the timings too. I genuinely feel that Matt’s retro was a beautifully balanced thing, and trying to make something after that feels like going up on stage to do stand up after a Bill Bailey set.

I’m writing this in the time between finishing the plans, and actually doing the retro. This section will go through the ideas I came up with, and the last section will show the frames post-retro with probably a lot of self-deprecating comments about how terribly awful it went.

I did a lot of thinking, and some reading of really long lists of retro activities, and decided that you need four types of activities for a really good retro.

  1. Light-hearted intro. Something quick and customisable that gets people thinking about the iteration. Like the weather report, you need something to summarise feelings before going into detail.
  2. Discussion of challenges. In some format or other, you need a forum for discussing what went well and what didn’t, what can be improved.
  3. Actions. After deciding those things, you need somewhere to put all your resulting actions together, things you need to do as individuals or as a team to improve.
  4. Appreciation. I think it’s really good to finish with something positive and personal that makes people on the team feel that their work is appreciated.

Ahah, after writing that out, I’ve realised that it’s pretty much identical to the format we used on FunRetro in the first place. It has come full circle… But I guess the trick is that within those bounds, you need to shake it up and keep the format interesting.

I’ve tried to do that, anyway. I decided to start with a theme, because I’m quite good at coming up with creative ideas when I have a general idea to follow along to. The theme I chose was boats. I like this, because boats can go fast or slow, they can float or sink, and there’s lots of different kinds.

My first frame is similar to Weather Report – except instead of pictures of the sky, we have gifs of boats. Miro is great for this, as you can drag and drop, copy-paste or upload gifs really easily. I don’t know how well it will manage with having many gifs at once though.

People will be placing their boaty gifs here, I hope for a variety. Mine is doing backflips.

Next up is the discussion frame. I could have spent ages trying to come up with something original, but there’s already a great boat-themed retro exercise called the Sailboat Retrospective. Apparently invented by someone called Luke Hahmann many years ago (unverified information), this exercise asks people to think about goals, aids, risks and anchors.

I could have saved myself some time getting a picture from the internet, but I like doodling

The above image is separated into quadrants. The wind in the sails represents things that helped us move quickly. The sun is our goals (even though the boat is facing the wrong way…). The squid monster is whatever we think is lurking in the depths, waiting to bite us in the arse later. The anchor represents things that are slowing us down. The team can add post-it notes to this board with appropriate comments for each part.

Next up in the Brucetro is the actions section. I’m not sure why I went with turtles. except that I had been looking at turtle pictures earlier in the day when I drew this.

I should really label these. The top left is “Actions I can take to carry my team”; top right is “Actions my team can take to carry me”; bottom left is “Actions the company can take to carry our team”; and bottom right is “Actions our team can take to carry the company”.

I’m not sure how this one will go down, as the differences are quite ambiguous and there will probably be a million requests for clarification on what each of those means. I do think it’s really important every now and again to check in not only with how we sit as individuals in our team, but how our team sits in the company though. I guess in this instance, ‘the company’ means both other teams and the upper echelons that make all the decisions, but are a few steps removed from the people actually implementing what they want.

Lastly we have number four, wrapping up with some appreciation:

Arr, me harties…

I thought it might be fun to (once again) split this into four categories. It’s the one consistent part of this retro, considering the boards are all different shapes and sizes.

In this frame there is a ship, with four characters in purple. The idea is that we will all nominate people for each category, vote and then I will draw or add a photograph of that person’s face to the picture. The first character is the navigator up in the crow’s nest, this is the person we think has been keeping us on track to reach our goals. Then we have the captain, who is a general MVP. The mermaid on front will be whoever has been great to work with, keeping up morale and such. Then the duck legs are for the person who has been frantically working away to keep the boat gracefully gliding forwards.

Again, it might be hard to distinguish groups, but we’ll see how it goes. They who dares, makes a great retro. Oof, singular they makes that sentence confusing, doesn’t it. xD

I just had a 1-1 with my direct manager and we went through the retro as I’ve designed it. He pointed out that I should definitely have some planned timings for each activity so that they don’t stretch on too long and cause people to lose interest. I think he’s totally right, and winging it is not a good plan when it literally takes a few minutes to decide on this.

We have a 2 hour time slot for our retros, but I’d ideally like for it not to take that long considering we have these every two weeks.

My current plan is:
Boat gif intro: 10mins (5 mins for finding gifs and people sorting out whatever issues they have connecting, and 5 mins to quickly go through them and laugh/banter about the gifs)
Sailboat Retrospective: 30-40mins (10mins explaining & writing notes; 2mins voting on issues; ~20-25mins talking about the issues. Depending on how many there are, we will only guarantee discussion on the top 5 voted issues. +/-5mins opportunity to bring up less-voted issues that individuals are strongly affected by)
Team actions: 20mins (5mins action writing; 15mins discussion)
Appreciation: 10mins (3mins naming; 2mins voting; 5mins bantz and winding down, adding people’s faces in)

So how did it go?

The feedback was positive, but not everyone has given feedback so it could be that it sucked haha. In terms of timing, there were a few points that came up that took up more time than I’d have liked, but I think we managed to get actions from them. We took up the full two hour time slot, including a few minutes at the end for messing around with the board, but didn’t go over. So I guess it was perfectly timed?

There were only 8 of us, so the gifs worked out okay. I think if there had been more, it would have burnt out my Mac to try watching all of these on the Miro board while screensharing on Zoom…

The ice-breaker was pretty fun, though I think it was more a case of people trying to find really funny or cool gifs instead of ones that actually represented how they feel about the last iteration of work. There was a mix of people who felt they’d had a great two weeks (dinosaur, backflips etc) and people who had struggled a bit but could see the end of their suffering (cresting a wave). It did take a bit longer to watch and explain each gif than it had with Matt’s weather report, but it was a nice light way to start the retro.

After going through all that effort to hide names, our faces are everywhere.

I didn’t think I explained the four quadrants of the sailboat retrospective very well during the retro, but the team cracked on and wrote a lot which was good. Ahead of time, I prepared colour-coded post-its and pictures of people’s faces in an adjacent area for people to copy and paste onto the board. This took out the time spent constantly asking who had written which tickets, since everyone had an assigned colour – I also added a white ‘anon’ post-it for people to write things they didn’t personally want to talk about, but felt should be discussed. We then voted on the issues we wanted to talk more on, with two face votes each per quadrant. The faces are taken from a set of team stupid-face emojis we have in slack, which probably tells you a lot about our dynamic as a group. xD

For discussion, I set a 5 minute timer on my phone for each quadrant and we went through the topics from highest to lowest votes. I wasn’t very strict on keeping this timer if there was a conversation going on when it buzzed, but I reckon we didn’t go too far over on each. At the end, I also asked if there was anything we hadn’t discussed or had got fewer votes that anyone thought needed discussion anyway.

Additionally, we took actions from these discussions, which we placed as post-its above the board.

The third activity went quite well, and I love how people avoided putting post-its on the turtles’ faces. xD We got some great actions from this, such as:

Setting agendas before meetings, with some questions or topics people should think about ahead of time. We have a problem at the moment where we have lots of meetings, which have titles describing what they’re about, ie which service or tool or process is being discussed – but only the person running the meeting knows what’s really happening so no one can prepare or research ideas ahead of time.

Document decisions. Since we’re a sort of remote-first company, lots of decisions get made in video meetings and calls between small groups of people, which aren’t documented anywhere. No one ever has the full picture.

Highlight good things on blog. We want to talk more about the cool things we do (we do a lot of cool things) in small blog posts and the like. Showing off what we do would be really good for creating a sense of pride around what we create.

Clear vision. Our team struggles a bit with the ambiguity of our role in the company, in terms of knowing which duties and services are within our domain or not – since we’re the “ads” team, but as a company what we do is make and serve ads… So pretty much anything could come under that. We’re pretty much a catch-all for anything that doesn’t come under console or data.

My manager put his cursor over this frame as I was taking screenshots. I waited 3 entire minutes for it to move, holding the mouse pressed down, but it did not.

The last frame was a nice ending to the retro. There was potential overlap between the different roles, and I’d kept the MVP definition intentionally ambiguous. I reckon this worked nicely, because separating the general sense of appreciation from the whole which-person-worked-hardest thinking by having multiple categories, opened up the field for us to vote for people we wouldn’t usually go for given only one option. I don’t know if that sounds bad, but it just means that we could appreciate the people who we had directly seen to be working really hard, while also showing that we appreciated the less obvious work done by others. The designer won general MVP with a landslide victory, and I won Mascot, even though we weren’t nominated once in the frantic-swan-legs or the keeping-us-on-track categories. Because we do other things that are less tangible, like going above and beyond to help people with queries, communicating well, helping to keep the board moving smoothly, boosting morale etc.

So there we have it, our journey so far in improving our remote retrospectives. We’ve had two really awesome, fun and useful retros so far so I hope this continues.

Also, the day after the retro I got asked by a member of another team if I had “used that website Matt found” to get my ideas. I had no idea what he was talking about, but after it was shown to me I wish I had. It’s called RetroMat, and it has a tonne of activities which you can mix and match. It splits the activities similarly to how I did – but better – into five categories: set the stage; gather data; generate insights; decide what to do; close the retrospective. You can just slide through different options until you find combinations that are suited to the feel you’re going for.

I would definitely have used it if I’d known about it, as the ideas on there are great and it would have saved me a lot of time haha. Then again, it was cool to create something from scratch that was enjoyable and added value.

Generally, the point is that you can do really fun, engaging and personal retrospectives with your team even if they’re not in the room with you. You can have meetings with personality, even when you’re not there in person.

And that’s another post from me! I hope you’ve got some ideas to take back to your retros too. :3

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.