BD #14 - Use the Same Page Tool to Get Your Team Aligned

This week’s issue is brought to you by Wallaroo – the purpose-built platform for last mile ML.

Wallaroo is a great end-to-end MLOps tool with some amazing features for complex deployments (I wrote more about this here).

Download the free Community Edition here, and there are a load of excellent tutorials here.

Have you ever been part of a team that just didn't click?

It's a strange feeling.

No one person is being disruptive, people are all chipping in and trying their best but... but something just isn't working.

Today I'm going to share a tool that's helped me combat this very thing. It's great to use at the start of a new team but it can work to freshen up and revitalise existing teams. Even if nothing's wrong, per se, it can be a rewarding exercise.

I call this the Same Page Tool - but it comes from an unusual place

Dungeons and dragons

I've been a collosal nerd my entire life and very much into games. Not just satisfied with computer games I got into tabletop roleplay games (TTRPGs) when I was in my early teens.

For those of you unfamiliar with TTRPGs you essentially gather in a group, typically four to six people. One of you controls the story, the plot, the challenges to face and the monsters to defeat - acting as storyteller and referee all at once. The rest of you form a group or "party" where you control just your character and over the course of many, many sessions you collaboratively tell a story about what that party gets up to. Sometimes there's rules, sometimes there aren't - the point it you're working together to develop a story and your character.

I talk regularly about some of the benefits this hobby has for data scientists around developing perspective, teamwork, and especially empathy.

But there are a lot of challenges when playing these games. Sometimes the group doesn't knit together or the players all want very different types of game (can you see where this is going?).

See, if you're playing a fantasy setting like D&D there's a variety of styles the game can take. If one player wants a dark and gritty Game of Thrones-like setting and another wants a capers and calamity misadventure more along the lines of Shrek or Pirates of the Caribbean then the group is potentially going to struggle to get the most from the game.

So back in 2010, Christopher Chinn wrote about something called the Same Page Tool. Essentially get everyone down at the very first session and get them to rate what they're looking for across a number of axes - what kind of game people want to play, who's responsible for what, how to apply the rules, what to do with disputes, etc.

This simple exercise was genius and I've used it time and again to ensure that these types of games get off to a good start.

Applying this to data teams

After running a few teams and noticing parallels I started to think about running a similar exercise in a professional setting.

The thing is, there are a lot of unspoken rules, expectations, practices and processes that you just kind of fall into in a team environment. If it's a new team you typically end up just going with whatever the most senior people think or what management say. If it's an existing team you might even be doing things that were only ever done for the benefit of someone that no longer works here!

So running a Same Page session can be a big help to ensure the team are getting a chance to build the environment that will best support them.

I typically do this in two stages:

• Anonymously survey the team to ask a bunch of questions about the important topics

• Replay the survey results in a group for open discussion and documenting decisions

What you survey is up to you but I find, as a minimum, these things make a good start:

• What shared project/tasks management system do we want to use (Kanban, SCRUM, CRISP-DM, TDSP, shouting etc.)?

• What regular meetings do we want to have, if any?

• How will we track progress and stay accountable internally?

• How will we track progress and stay accountable externally?

• How will we prioritise work?

• How and when will we communicate progress outside the group?

• How will we settle disagreements?

• What's the escalation route?

• When should we revisit these decisions and refine them?

You might also want to talk about more data specific things like:

• How will we test ETL?

• Whose job is it to fix data quality?

• How do we approve and share collated datasets?

• What model monitoring or A/B testing approaches do we want to use?

You can add to, subtract from, or tweak this list however you see fit. Some teams might want to walk through things like definitions of ready/done, what columns to use in Kanban, coding styles etc. etc.The objective here isn't to get the definitive, final view - just to give people an opportunity to express their thoughts and open the discussion. Once that's done, get the team in a room and present the results.

It's key that someone takes great notes in this meeting, you'll almost certainly not be able to please everyone and there's an opportunity for you to make people feel like they weren't heard. Write down the final answers to each question and some of the other ideas that were "parked" for now. When you come to review and revise this might be helpful for looking at other ideas and approaches.

The output of all of this is an artefact that the team can line up behind. It also clearly signals what working in this team will be like - if that's unpalatable for any member of the group then now's the best opportunity to talk about solutions, maybe there's another team they'd fit better with or more support you can give them to bridge the gap?

Final thoughts

I've used the Same Page Tool time and again and it's a great way to kick off a new team. It also gives you plenty of material to turn into proper documentation and templates for ways of working and processes that can be shared outside the team, should things go well and need replicating.

I'd love to know people's thoughts on if they've done anything like this previously and other experiences of getting teams on the same page. Please reach out if you've any great learnings to share.

All the best,

Adam

If you're looking for support form me, here are a few options:

• Sponsor this newsletter: email [email protected] to discuss reaching 1800+ data and AI leaders that subscribe to this newsletter (booked out until Jan 2023).

• CDO-as-a-Service & Consultancy: I'll help you build and shape your data function for a fraction of the cost of a permanent Head of Data or Director hire; get in touch about this or consultancy via email to [email protected]

• Coaching: I offer 1-2-1 coaching; book a coaching session here

• Free, daily insights: on LinkedIn here or Twitter here