Local data value in large organisations
Happy Friday!
Today I write about how we maintain the ability to provide local value through data in large organisations.
There’s also links to articles on buying a data catalog and whether that really meets your goals, how to communicate about data like a leader, and a game to see who is the best CDO!
Local data value in large organisations
Have you ever worked as a data person for a small organisation, <100 people?
It’s a great place to be!
You know everyone, you know everything about the companies priorities, and you know all the data that’s available and could be used to support those priorities.
I saw a good example of this when I saw Ties Boukema of Dawn Capital (a VC) speak at a meetup in London a couple of months back, and which you can read about here. Ties knows exactly what’s important to the organisation (connecting with founders) and how data can support that (building a “roladex” of connections).
There’s ~38 people working at Dawn Capital, and it’s probably not going to grow significantly more than that. So there’s no need for data contracts, data mesh, shifting left, or any other initiative.
They can rely on human communication and collaboration to deliver the right value to the right people.
The problems start when a company grows beyond that.
As a data team you can no longer engage with everyone, you’re further away from the priorities of the company, and the amount of data and its complexity increases far beyond what you can keep up with.
This is why we move to more decentralised models, such as data mesh.
We’re trying to empower that same local data value across large organisations.
And we do that by promoting their autonomy to own and manage their own data and take actions with it, without needing central approval and review.
But it would be wasteful if every local data owner had to decide how to create and manage their own data and build the tooling to do so.
So, we provide a self-serve data platform that takes away that complexity from them, allowing them to focus on the data itself.
And then there is the challenge of governance and ensuring data is managed consistently and in line with policies across the organisation, no matter who owns it.
So, we also need to implement governance through this data platform, ensuring compliance without relying on manual actions by dozens or even hundreds of data owners across the organisation.
It’s these investments that allow that local value to be created from data throughout a large organisation.
Sponsor
Keboola MCP Server connects AI copilots like Claude or Cursor directly to your Keboola environment via the Model Context Protocol (MCP). Now, your chats can spin up real extractors and transformations, write accurate SQL, deploy production-grade pipelines in seconds, debug failing jobs or auto-generate clean, up-to-date docs.
Data engineering just got conversational.
Interesting links
We Were Young and Impressionable. So We Bought a Data Catalog by Kasriel Kay (LinkedIn)
We’ve come a long way since those early days — when we were young, impressionable, and convinced that structure alone would unlock value.We built catalogs, launched councils, and mapped the mess. But the organizations finding real traction today are thinking differently.
They’re not chasing more visibility — they’re building systems that make data move with intention.They’re embedding trust not in process, but in flow, and they’re not asking who owns the data. They’re designing environments where the answer doesn’t have to be manual.
That’s the difference between organizing your data — and operationalizing it.
Great post.
4 Strategies To Elevate Your Communication Style From Engineer To Leader by Ben Rogojan (Seattle Data Guy)
Good tips on speaking and acting more like a leader. Worth reading even if you’re not (currently) a leader.
Who’s the best CDO? by Charlotte Ledoux
A fun little game on being a CDO and the impact your decisions might have.
Being punny 😅
I worked at a toy shop as a teenager and some weeks I was bringing home six figures. Then I got sacked for stealing.
Thanks! If you’d like to support my work…
Thanks for reading this weeks newsletter — always appreciated!
If you’d like to support my work consider buying my book, Driving Data Quality with Data Contracts, or if you have it already please leave a review on Amazon.
Enjoy your weekend.
Andrew