Scrum Foundation for data teams: what works and what I adapt
My first certification outside the technical side was Scrum Foundation, from Certiprof. It is still useful because data projects have a specific problem: the work is hard to see.
by Gabriel Henrique · · 2 min read · Ler em português
I'm Gabriel Henrique, a data engineer, and I hold the Scrum Foundation Certification from Certiprof. It covers the Scrum basics as defined in the Scrum Guide.
In this article I go through what the certification teaches and how I use it on data teams, where Scrum needs a few adjustments.
Scrum basics
Three accountabilities
- Product Owner: decides what has the most value and orders the backlog.
- Scrum Master: makes sure the team uses Scrum well and removes impediments.
- Developers: the people who build the increment. In data, that means engineers and analysts.
Five events
The Sprint, which contains the other four: Sprint Planning, Daily Scrum, Sprint Review and Sprint Retrospective.
Three artifacts
Product Backlog, Sprint Backlog and Increment, each with its commitment: Product Goal, Sprint Goal and Definition of Done.
The challenge for data teams
In software, the increment is a screen or a feature someone uses. In data, it is often a table nobody sees until the dashboard is ready. That makes progress invisible and the Sprint Review empty.
It's also common to discover mid-Sprint that the source data isn't what was expected. Estimates miss for reasons the team doesn't control.
How I adapt it
Definition of Done with data quality
For me, a table is only done when it's documented in the catalog, has quality checks running and a consumer has validated the result. Writing that into the Definition of Done prevents deliveries that only work on the author's machine.
Reviews with real data
In the Sprint Review, I show numbers and the business question they answer, instead of code. That's how the business side understands what moved forward.
Spikes for unknown data
When the source is new, I set aside a short task just to explore the data before estimating the build. That reduces surprises mid-Sprint.
Where it helped me
Today I work with several support and business teams to gather requirements and align delivery expectations. Scrum's shared vocabulary makes those conversations faster. I went deeper on the people side later in the Aspire Leaders Program.
Read about my path as a data engineer on the portfolio of Gabriel Henrique, Data Engineer. I'm Gabriel Henrique, and I write about data, methods and career.