K#4. The best productivity system to safeguard your focus

The goalie system: crazy impact + actionable tips

Apr 29, 2024

Heyyy 👋

Welcome back to K’s DataLadder ✨! It’s been only one month but we’re already 707 strong now and all thanks to you! ❤️

This week, we’ll dive into a data role you might not have heard of before but which has revolutionised how we handle our data science workload in my team at Spotify. This system could do the same for you or your team and it’s called the “goalie” system.

Estimated Reading Time: 4 minutes

But first, how are you doing this week? Me pretty well! A lot has happened. I played a lot with my new camera around the city and I ADORE it. Paris is so cool right now in Spring especially since we’re preparing for the Olympics Games soon. Other than that, I met with my friend Nabil [The Bald Prompter Newsletter] and we motivated each other for a writing session, it’s always so nice to work alongside like-minded people!

Agenda

This week’s story
Why it’s a game-changer
Daily duties of a goalie
Tips for future goalies

This week’s story

This week, I’m the designated “goalie” of my data science team.

What does that mean?

I’m the go-to person for all ad-hoc requests, maintenance, and last-minute tasks. It’s a hectic role, but someone’s gotta do it, right?

Being the go-to person for a week means I’m the one tackling all sorts of questions and tasks that used to disrupt our workflow, and just flow in general. During that week, all of your ongoing projects are held on pause.

It completely transformed how I feel about my job.

I have ADHD so getting myself started on a project and keeping the focus going is a challenge everyday, so imagine having this focus abruptly broken many times a day by stakeholders asking for specific data ASAP.

I used to shed tears of blood each time I got a slack request, but that red sea is all dry now and it’s thanks to our goalie system!

Why it’s a game-changer

Here’s why the goalie system is a game-changer for any data science team:

Streamlined communication: One point of contact for all data queries.
Faster response times: No more waiting around for someone to be free.
Enhanced focus for others: The rest of the team can dive deep into projects without interruptions.
Skill-building: As goalie, I get to touch on various aspects of our operations, which boosts my adaptability and knowledge.
Dopamine boost: It’s nice to ship many stuff a day, you feel productive, and we’re all pumped on that dopamine juice, so you get your daily shot.

Since we implemented this role, it has seriously been a lifesaver for our team dynamics and productivity!

So whether you already started working or will do in the near future, try suggesting a similar system to your team. It could definitely help:

Showcase your proactivity to improve ways of working
Boost your visibility within the team through impactful contributions

Daily duties of a goalie

Lead team meetings: Kick off weekly syncs and manage demo sessions.
Pipeline management: Decide whether to deprecate outdated datasets after checking dependencies and usage.
Ad-hoc data requests: Compute metrics, eg. user growth over specific periods or how specific user groups interact with features.

Example task from this week: I handled a request to analyse user growth since early 2022. I pulled the data, calculated the growth percentage, and visualised it. Then I sent it to the PM who requested the data and closed the ticket for that task on our tracking system.

The other tasks were more complex, so I chose to share with you this simple one.

SQL snippet I used:

WITH p0 AS (
    SELECT COUNT(DISTINCT user_id) as p0_metric
    FROM your_table
    WHERE date_column = '2022-04-25' -- the starting point
),
p1 AS (
    SELECT COUNT(DISTINCT user_id) as p1_metric
    FROM your_table
    WHERE date_column BETWEEN '2022-04-25' AND '2024-04-25'
    AND EXTRACT(DAYOFWEEK FROM PARSE_DATE("%Y%m%d", _TABLE_SUFFIX)) = 2
) -- focusing on only one day per week to reduce running time

SELECT
    p0.metric AS p0_metric,
    p1.metric AS p1_metric,
    ((p1.metric - p0.metric) / p0.metric) * 100 AS percentage_change
FROM 
    p0, p1;

Tips for future goalies

Handling random data requests can be daunting, but here’s how we keep it together:

Ask stakeholders to provide the following info when sharing their requests:

Context & Objective: Provide a description of the issue being addressed and include any relevant links to projects.
Impact: Explain the potential impacts of this request.
Due Date: Specify when the data is required.

Scope the questions: Try to narrow down broad requests to deliver faster.
Manage expectations: Clearly communicate what’s doable within the week.
Scope small requests: If it can be done in half a day, take it on directly.
Collaborate: For specialised or extended requests, loop in the right experts from the team when needed.
Keep track of tasks: Always have a system to manage these ad-hocs—whether it’s a shared tracker or regular check-ins, make sure everyone’s on the same page.
Handover: Make sure to document in a short text whatever needs to be handed over to the next goalie in terms of asks or issues that you could not address.

Being a goalie means being proactive, resourceful, and ready to tackle anything that comes your way.

It has taught me a lot about our workflow and helped me learn more about what my teammates are doing because I get to deal with requests spanning all of our department.

So next time you’re swamped with requests, think about how a “goalie” system might work for you and your team! That way you get to kill two birds with one stone: requests, and your daily dopamine intake.

If you haven’t done that already and you want to show me support:

Subscribe to my YouTube channel. It would mean the world to me!
Follow me on Instagram, LinkedIn, X, whatever works for you really!
Join +4.8k readers on my Medium blog
Check out my latest data science piece on how I landed an internship at Spotify

See you next week 🫶