It’s been a busy month for me since my first speakers’ appearance at SQLBits in March, enjoying the Easter break before getting back to a busy period with our clients at Dufrain. Reflecting on the talks I attended, I am finding more and more that events like SQLBits are a must for data professionals. To be around other experts is invaluable for keeping up to date with emerging trends in the data space – especially in the new and exciting world of Microsoft Fabric.
A speaker and an attendee
I presented my thoughts on Creating a Power BI Culture: Governance as a gateway to Adoption, not a barrier. I shared the highlights of Dufrain’s Power BI Adoption framework, including our pyramid of self-service, our 4 Power BI Governance principles, and the shopping list of deliverables our client receive when embarking on a reporting journey with us. I received a feedback score of 8.88/9 which was a great result for my first SQLBits appearance!
I attended some fantastic talks – from Hylke Peek’s Govern your data in Fabric, and Chris Webb’s Fabric Direct Lake, I left with a book full of notes and ideas to ponder. If I had to sum up my one key takeaway from the week, it would be that Microsoft Fabric really is the tool for Data Professionals. By this, I mean that it blurs the lines between what we expect from Data Engineers versus BI Developers. In my blog from last summer, on the announcement of Microsoft Fabric, I urged organisations to consider the setup of their data functions. If your Engineers and your BI team are separate teams with minimal collaboration, upstream versus downstream, then you will not feel the benefit of Fabric’s true potential without redefining what it is to be a “Data Professional”.
The AI trend and empowering BI teams
With tools like Chat-GPT being so in vogue, I have noticed a trend in professionals leaning on AI to bridge the skills gap to make Fabric features such as Notebooks more accessible. From a learning & development perspective, I think this is brilliant, and empowers upskilling on a whole other level to what was possible pre-Fabric. Traditionally, a BI Developer looking to have more control over the modelling or ETL of their data would either find themselves in a battle of access with IT departments or result to Power Query and Dataflows to get the job done. Implemented the right way, with good governance in place, the nature of Fabric as a SaaS product breaks down these silos and gives us the chance to create empowered Data Professionals in organisations.
Would I endorse Python scripts from CoPilot being implemented by BI professionals on a production environment at enterprise scale? Well, of course not. This is where we come back to my advice for organisations on collaborative data teams. The best and most productive projects I have been part of are the ones where architecture, engineering and business intelligence teams shared stand-ups, DevOps boards, office space and team socials. “Solution Pods” of 1 Architect, 1 Engineer and 1 BI, maybe with a BA thrown in for good measure, are the key to delivering real value for your organisation.
Lastly, as I consider the hours of innovative demos from Data Activator, Fabric Notebooks, Real-Time Reporting and more, I found myself thinking about how often I am still supporting organisations with unpicking troublesome Power BI models that have been plagued with badly written Power Query, and even worse DAX. If Power BI teams have not yet mastered clean data models and DAX – how on earth can we unleash this functionality on them?! I often think that Power BI is like skiing – easy enough to strap a pair of skis on and snow plough your way down the mountain, but to master parallel turns and the precision and control required for exciting jumps one has to invest a considerable amount of time and effort. Accessible at first, but an uphill climb to master. One of my favourite things about Power BI is how accessible it is – I was once an admin with no data background who found herself empowered with a Power BI License (and the rest is history!) – and so the need to invest in quality assurance, training and development, targeted governance oversight, and time to reduce technical debt is crucial. Without these investments, organisations face a decision between shutting down innovation, or allowing a free for all on your data landscape.
Good governance should ensure the right controls are in place to keep data secure, maintain one version of the truth, and reduce maintenance overhead while encouraging innovation. Controls which stifle access to change lead to disengagement from users, and poor adoption rates.
Speak to the data pioneers today
At Dufrain, we’re on a mission to help our clients strike that balance. If you’re reading and relate to the challenge, get in touch and let’s talk about our Power BI Adoption offering.