I’m working on three years into my journey as a dbt user, and six months into my journey as a Product Manager at Fishtown Analytics. Part of my onboarding to the product organization was a deep dive into the history of dbt, and working to understand the decisions that lay beneath today’s reality.
One of the interesting things that happen when you do this kind of deep dive is you can see patterns or facts emerge that weren't obvious at the time – and it turns out that January 14th is one of those patterns!
- In 2019, on January 14th, the first commit in Github to rename Sinter to dbt Cloud was pushed live - the next day it was announced to the world!
- Then, one year ago, on January 14th, 2020, the IDE in dbt Cloud was announced.
This wasn't planned, it just turned out to be a fun fact about our history. Kind of cool, right? But these two releases have more in common than a shared birthday, they each mark a significant deviation from dbt’s history.
As I thought about this connection, I found myself thinking about how our history shapes our present, and how in recognizing this fact, we can start to find ways to reckon with making the future different from the past.
For me, this is the story of dbt itself. What we call the Modern Data Stack is differentiated by that first word, “modern”, in that it is the current way that we approach this work. It's situated, even named, by the way in which it differs from its predecessors.
dbt transforms data, but really what it does is transform organizations, and the way that organizations handle what has, historically, been a giant bottleneck to their successful understanding and analysis of their growing oceans of information.
A lesson from history: the Toyota Production System was developed starting in the 1940s, and was expanded to their parts suppliers in the 90s - one of the key tenets of the system is that efficiency gains come only from addressing the bottlenecks in production.
dbt is remarkable for taking this idea, already applied elsewhere in the world, and making it work exceptionally well in a new arena. For data teams, the bottleneck is always data engineering.
So, what if instead of building more and better tools for data engineers, we built something that would allow analysts to to own the data engineering work that is most central to their success? (we call this analytics engineering, btw). The deviation is not just “a better solution” it is in the elimination of the data engineering bottleneck by democratizing access to the solution space.
This kind of democratization means that more folks within an organization have access to data transformation, testing, and even architecture. Transformation and testing work grows closer to the business units that have an appetite for its outputs – which are now tailored to their needs.
Building out the IDE is a fundamentally similar deviation from history – it's a move to make a developer tool easier, more accessible, and is doubling down on that same democratizing energy. The more folks who are able to gain access to and find agency using dbt, the more power it has as a tool to make all parts of the business that consume data more efficient.
The shift from renaming Sinter to dbt Cloud a more personal deviation for the Fishtown Analytics team. Tristan wrote in his blog post on the topic, “This decision is also about creating a sustainable business around dbt.” Fishtown Analytics was no longer building a random set of data tools, we were building one thing – dbt. It may have looked like “just a rebrand”, but it represented a meaningful deviation from how we previously thought of ourselves and what we were building for the dbt community.
Which brings me to our present January 14th...
About Staging: dbt Demo Days
As our team grows and the number of people building dbt grows, we recognize the potential to drift away from the dbt community. Here, perhaps, the deviation we seek to make is a course correction, to replicate the fast feedback loops Drew and Tristan enjoyed in the early days of building dbt.
We're going to begin hosting a quarterly event focused on giving the community space to provide input on the product roadmap as well as share how you’re all using new features. The “we” here is myself (Product Manager, Cloud), Drew (our CPO), Jeremy (Product Manager, Core), and other familiar faces on the Fishtown Team. This event is called Staging, and the first one is happening February 11th.
I’m not exactly sure how this event is going to go! In my mind it will be a bit like Coalesce where we all get to go deep on analytics engineering, chatter in Slack, and share some memes – but it will be specifically and only about the product roadmap. What we’re about to release. What’s recently released. And what’s coming up. So yeah...I’m pretty excited!
I hope that you’ll join me in reflecting on what the long-vision future of dbt might look like; how it might expand again, become still more democratic, what new ways it might serve to make data streams and organizations more efficient and performant – in what ways we can act today to make the future better than the present?
What do we want to be sending out into the world on January 14th, 2022?
I’m looking forward to getting more time to work with all of you in creating dbt’s next deviation from our past 📈