top of page
  • Writer's pictureMonica Kay Royal

Event Review: Datanova | Taking Data Mesh from Theory to Practice

Brought to you by: Starburst


This event was designed to provide an accelerated learning opportunity to help data leaders take their best next step on their data mesh journey. And it delivered!



This was a new topic for me that I just learned about this month. I attended an event earlier this week that talked about data mesh at a high level and had some questions and was skeptical about how it actually worked. For more on this, see my review of the Semantic Layer Summit 2022.


At first, it seemed that everyone was celebrating data silos in organizations, like they were a good thing. Talking about how decentralized models are the only way to be successful in creating data products. It wasn't until like halfway through this event when I had a full on

AH HA moment 😲


Thank you to Prukalpa Sankar at Atlan, the way that she explained data mesh cleared everything up. For some reason, I could not connect the dots and realize that data mesh was the SOLUTION FOR these data silos and decentralized models. This is a wonderful example of how important presentation and public speaking skills are in getting your point across and having your audience understand what you are saying. No jab at the other presenters, these presentations were meant to be deep dives into the nitty gritty of implementing the data mesh. I just really appreciated the simplicity shared and hearing Prukalpa's perspective.


With that said, rather than going into each of the presentations themselves and risk the possibility of confusing everyone (including myself), I will give a high level summary. If you do desire to learn about the details, I provided a link at the end so you can watch the sessions on-demand 🤓


If you want to skip this review because of that, no worries, thanks for stopping by. BUT... please scroll to the bottom and check out my summary of the fun and humor provided by the amazing Scott Taylor.


What the heck is a Data Mesh?


How I first learned of the data mesh was that is was an organization principle that is used to deliver analytics to the organization but allows the business users / data stewards to control and create the data products. This was the definition that made me question all existence.


With more clarity, Starburst defines data mesh as a strategic approach to modern data management and a way to strengthen an organization's digital transformation journey. The main objective is to evolve beyond the traditional centralized data management methods. Additionally, data mesh is described as a 'socio-technical' approach that requires changes to the organization across all three dimensions of people, process, and technology.


There are 4 pillars of Data Mesh

  • Domain-oriented ownership & architecture

  • Data as a product

  • Self-service data infrastructure

  • Federated computational governance

The main takeaway with this is to allow the people that have the most knowledge of the data be the ones that are creating the data products because they know the data best, they are the subject matter experts.


Prukalpa talks about the concept of metadata and how it is the foundation that your data mesh needs. Metadata is typically used by data catalogs to organize the inventory of data assets within an organization. The problem with these traditional data catalogs though is that they are not personalized to the individual needs of particular organizations. They are super generic and treat all users the same, when actually every user is quite different. She then gave an awesome example of how the recommender systems at Netflix are able to give personalized recommendations based on the data used for that product.


To close things out with the definition of data mesh, a few other topics I have seen shift lately

  • Data is no longer an asset, it is a product 🤔

  • Data lakes are now termed lakehouses 🧐

and as mentioned previously

  • Decentralized models are actually not so bad (*if you use a data mesh) 😃


People, Processes, and Technology!


This was discussed multiple times throughout each of the presentations, it must be important. Let's look at the highlights:


People:

  • We need to get people on board to use the data products, training may be involved

  • Every organization has different people in different roles, each will think differently

  • Have the people that know that data the most create the data products

  • Leaders are KEY, they are the direct line to the people

Processes:

  • Data mesh streamlines the data product creation process

  • Process depends on data literacy of the organization

  • Transformation requires change, change is a process

Technology:

  • Technology may have improved, but the data team structures and tools have not

  • Data mesh can benefit BI and downstream tools


Typical Data Mesh Journey.


Kris Karaikudi and Sharjeel Khalid at Slalom gave one of my favorite presentations because they brought in behavioral science. 🧠


The typical data mesh journey includes the following: ideate, build, govern adopt.

Emphasis on the adopt!


Ideate

Moving to a data mesh is both a technology and a business change and it starts with data product ownership, from the business.


Build

You must set up the underlying platform and tools for your data mesh and the business needs to co-own the capabilities.


Govern

Establish processes that support end-to-end governance across your data lifecycle.

new data products > ownership and governance > data consumption


Adopt

This is the hard part and what makes this so difficult is the people. As previously mentioned, everyone thinks differently and will have different opinions, reactions, emotions, and narratives.


You can have someone who shows enthusiasm, is excited for change, and thinks that this is a great opportunity.

You can also have someone who shows nervousness, change gives them anxiety, and thinks that the old way makes more sense.


While transformation requires change and change is a process, it is also key to note that change is a mindset that could require changes in behavior. Therefore, it is important know these different behaviors so that you can help everyone get on the same page, by understanding the how and why, and start solving problems together.


Side Note

Behavioral science is my jam because of a previous role I had in providing data products that were built with behavioral data. Fun fact, I was trained in behavior design by BJ Fogg. His program focused on helping people live happier, healthier lives through understanding the ways you can influence and change behaviors.


Book recommendation: Tiny Habits by BJ Fogg


Fun and Humor from the amazing Scott Taylor

a.k.a: The Data Whisperer


Favorite Topic: Federated Computational Governance


Great at data jokes!



meta data = data about data
meta meta data = data about data about data
Seinfeld data = data about nothing 😂

What chemical regulates a rabbit's mood 🐇
... carrot-tonin  🥕

Parting Advice: 4 things data will do for your business
It will help you...
grow
improve
protect
sustain

In Conclusion

Kudos to all those that participated! I am glad to now have a better understanding of the data mesh world and hope to work more with these partners in the future!


Featured Partners:


If you would like to watch the sessions on-demand, visit here


And if you want to learn more from Starburst, sign up for next year's Datanova



Happy Learning!!

Recent Posts

See All

Comentarios


bottom of page