PowerBI.tips

Data Governance & Adoption – Ep. 227

Data Governance & Adoption – Ep. 227

Fabric is a platform upgrade, but it doesn’t replace the hard conversations. In Ep. 227, Mike, Tommy, and Seth use Fabric + OneLake as the backdrop to talk through adoption reality: discoverable data needs ownership, preview features still need cost controls, and “data quality” work fails fast when there’s no governance framework behind it.

News & Announcements

Main Discussion

Topic: Data governance vs. data quality (and why Fabric adoption forces the conversation)

This episode is basically a reminder that unified platforms amplify whatever operating model you already have. If your data is discoverable across OneLake and your org doesn’t have clear stewardship, the first question becomes “who owns this and who’s accountable?” (not “what’s the best connector?”). And when the platform is still evolving, you also need a real way to measure cost and usage—not just vibe-checks.

They also spar a bit on practical Fabric patterns: the tension between reuse and scope (for example, whether you end up needing “a lakehouse per dataset” for clean semantic boundaries), and the reality that governance teams may need to rewrite processes because the tool surface area is suddenly much wider.

Key takeaways:

  • Data quality is about the data itself (accuracy, completeness, consistency); data governance is the policies/roles/processes that define ownership, standards, and controls.
  • Discoverability (OneLake) makes stewardship non-optional—every shared dataset/table needs an owner and a decision path for change.
  • “Lakehouse as dataset” patterns have real tradeoffs: optimize for semantic clarity and reuse, not just convenience in the UI.
  • Preview doesn’t mean “ignore cost”—capacity and storage visibility are part of responsible adoption.
  • Governance is an org capability (people + process), not a feature toggle; Fabric often increases the need for cross-team alignment/CoE.
  • Access control and compliance still apply in a unified platform; governance has to cover sensitive fields and who gets to see them.
  • You can’t sustainably improve data quality without governance—tools can score/fix data, but governance is what prevents regression.

Looking Forward

Adopt Fabric like a program: pilot one business domain with clear ownership and standards, measure capacity impact early, then scale what works.

Episode Transcript

0:32 good morning and welcome back to the explicit measures podcast with Tommy Seth and Mike hello everyone hello hello all right let’s have a lot of dead air beginning of the episode let’s do that all right guys let’s not talk well for for all of our viewers they they saw my amazing hand gestures of explosions of excitement in yet another podcast episode so much excitement so much

1:04 so much excitement so much and a happy Tuesday TJ there it is I was waiting for it I I feel as if I should wait just a little longer so that more of our listeners who pop online after a bit get to get to experience get to enjoy the Tuesday think of the audio listeners they’re thinking what did I did I lose the connection is my Bluetooth headphones not working anymore still still here still with us we’re still with you they’re thinking about how many miles do I need to run or

1:34 how many miles do I need to run or something yeah well if you are running it is finally the hottest time of the year getting right there gentlemen it is just about July July it’s summer it is sure and if if I’m making I believe our next user or our next podcast is on the fourth fourth how do you think we need to discuss about is there already the fourth the 29th no the 29th today’s Tuesday oh that’s right

2:05 that’s right I gentlemen I’ve been quite the worker bee so to speak so yeah I’m like I’m like is it Thursday it’s got to be Thursday at this point so my goodness it’s it’s nice I don’t in in our area we’re getting a nice swath of storms to come through here let’s bring bring the grass back to green right before we start launching scorched Earth massive amounts of fireworks into the air true statement is very good actually

2:35 air true statement is very good actually yeah all the grass it looks like it is very yellow yeah bring it back in bring it back in well welcome to the weather channel yeah yeah your Midwest Weather Channel because that’s what everyone was tuning in for what do we get what do we give for announcements today I know I have one I’m super excited about that I saw let’s kick it off first let’s go ahead Seth lead us in one of one of my one of my Dax and modeling guys was the I have to give thanks to for popping this

3:06 have to give thanks to for popping this one into my into my feed area but it’s the new Dax Optimizer by SQL bi has been verbally launched via LinkedIn from Marco Russo just a mere day ago promises to solve all of our Dax woes that’s what I read in there just promises to solve them all maybe maybe I’m over embalancing a little bit but either way any tool that comes out of SQL bi I’m super excited about

3:39 SQL bi I’m super excited about and if we want to share the link out Mike you can sign up to get on the waiting list of a beta version of the Dax Optimizer so they’ve if everybody should know who who SQL bis and Marco and Alberto but they have a Dax optimization course that they’ve been doing for many years and everybody who’s taken it I I hear nothing but rave reviews so if they’ve taken their

4:10 reviews so if they’ve taken their Collective knowledge and put it in a tool for us to use I’m extremely excited to see what that looks like well I think this is probably more like brought in a sense Bravo 2. 0 because Bravo had some like formatting I guess but man this is I imagine this is going to be a mix of the best practice the BPA with some things with Bravo and just some good old new new new features that we’re probably used to but just having it in the automated way so

4:41 just having it in the automated way so I’m very excited to see what this looks like like the more automation the better right yeah I wonder if they’re using AI in this thing that’s what I was thinking so Seth was and I were talking the other day day we were like with all this Ai and you we were like with all this Ai and one Lake and fabrics showing up SQL know one Lake and fabrics showing up SQL bi has been oddly quiet about all of that stuff no they haven’t really been talking about one Lake oh they haven’t really been talking about the new data engineering and the fabric roles and stuff like that so Seth and I

5:11 roles and stuff like that so Seth and I were thinking maybe they’re noodling on something well yeah now it comes out yeah they were noodling on something they were calling they were building the daxed optimizer application I and so I’m wondering I haven’t seen it yet don’t know anything about them I’ve also signed up for the preview so not sure what it will do other than every time you run it it’ll be like your model sucks build better tables do more data engineering Upstream here’s recommendation we’re going to make this measure change and this one’s

5:41 make this measure change and this one’s called by this one this one this one so here’s the other change the cascading changes you would need to make like eh the AI angle is not bad because Marco put in like he had an hour video of him using chat gbt to create a Dax measure or basically create its model I use all the Dax measures and he was going through like okay let me try to change the question so he’s been very involved there so I can also hear him regardless of the AI where he’s like yeah fabric doesn’t change your need to know Dax at

6:12 doesn’t change your need to know Dax at all Well here here’s the thing I think that’s most important on this tool it’s a phrase they used here they said not only will identify problems but it will rank them in priority that is the most I think most important thing because people one can say oh yeah well my my measures aren’t in folders okay is that really a Major Impact to the model performance probably not right so maybe it’s a best practice to do that but we’re not going to force you to do those things let’s focus on hey by the way this measure is going to run

6:42 by the way this measure is going to run really slow because there’s a super a crazy filter context transition that’s happening multiple times in this thing or you’re using summarize when you shouldn’t be using a summarize I think that’s going to be more impactful so that you can you could narrow your attention because if you’re if you’re given again this is probably what Marco sees all the time he’s given a model that has hundreds of tables tens of tables lots of tables with lots of measures all over the places like it’s slow it’s slow where do you start where do you get the most Improvement for your time spent on that model and

7:13 for your time spent on that model and this makes total sense they probably do this already I hope it talks to you in real life and it talks in Italian at least has an Italian accent it’s an Italian accent your Dax is no good is no good there’s no good I need to take a break and get some pasta pasta [Laughter] what else we got so that was our our one hit for today there’s another one that we had kind there’s another one that we had out here Tommy this one came from you of out here Tommy this one came from you you found this one’s

7:43 you found this one’s roaming across the Internet it’s a Twitter status message from Bernat and Bernard’s been very engaged with just this whole one Lake and fabric thing so we’re not you’re doing a great job of just really just diving in and asking lots of really relevant questions yeah Tommy just put the the question there so Bernard is asking some questions about this one Tommy give us some context here what’s what’s going on this question from Twitter so Bernat and his questions have been a man after my own art actually with the the same idea and the general gist here is but not initially asked is fabric

8:14 but not initially asked is fabric basically saying that it’s a good idea to use your whole lake house as a data set for power bi is is that what I’m hearing and as they’re going going on through the Twitter thread a very pointed statement by Bernat was wait we need a custom lake house for each data set and that’s been the general feeling if you think of direct direct import or you think about how that works it you’re creating right now the data set in the power bi service using a one lake or right and if you try to select tables

8:48 or right and if you try to select tables well automatically that becomes a SQL connection which becomes import so you lose that ability to use power bi’s Now default way of connecting to data so by this means if you think about a even mid-size organization the amount of data sets that exist do all these need now to be in a their own lake house to work with power bi or to work with the fabric because if you take everything right now

9:18 because if you take everything right now from a lake house you take everything from that data set or everything from those tables all the tables all those relationships which still mind you can only be created in the service this and lies I think the conundrum right right so this is this the Nuance of that question we talked about in the last episode where it was like there’s there was this generic understanding ahead of one like where all these artifacts were just going to be available to us to connect to but yet here we are connecting to four different types of

9:48 connecting to four different types of data sources they absolutely are still separated well at the end of the day you need if you’re going to create a data set in power bi it’s an entire lake house that’s not a subset of the lake house yeah but what if you so my understanding is again maybe this is something I’m just not understanding quite yet my understanding is if you’re doing direct leg right you go to the lake house and you click new power bi data set right and then you just pick the tables that you want you don’t need to take all the tables right

10:26 you have to recreate the measures all over again recreate what measures you’re creating a data set in power bi when you create a new data set in fabric you have to recreate the relationships you have to recreate the measures and that’s for every individual data set yeah but what in the what the idea here being you have to do that anyways when you make multiple data sets there’s no I don’t think there’s a like I understand the challenge of that but like there’s no are you saying you want the measures that come with the original lake house

10:56 that come with the original lake house yeah like here’s the lake house here’s the measures and then you just import portions of that and just gives it to you if if the general proportion of what you’re trying to do is going to serve multiple data sets there’s probably very base measures that you’re going to want especially the relationships

11:10 especially the relationships especially the relationships and the metadata the hierarchies those are all consistent if you’re using those same tables tables why would you have to recreate those again if you’re using the same fact product category table and I don’t think the lake house is intended to store relationships and measures at this point that that is a function of analysis services or SQL right we do we just we’re bouncing back and forth between my concepts of stuff that I something’s not

11:40 concepts of stuff that I something’s not sticking here the lake house as a connection is just a set of objects I can pull into a model and start to create relationships and measures on top of right sure I think so yeah Okay so Okay so if I I have access to one leg objects through the lake house connection why would I not be able to just add another object into my model well I think that maybe we’re doing the wrong I think I think it’s directly

12:11 wrong I think I think it’s directly because what you’re telling me you’re referring to it’s specifically a direct link function yes multiple Imports when we if you choose which tables you want to connect to from a from a lake house I think this is just we were I think we’re asking the wrong question I think the question really is how big is your lake house like how many objects are actually stored in one lake house okay so and I think that’s the question we need to start asking this is the

12:42 we need to start asking this is the point that I think is more relevant for Bernard’s question here is there’s there’s two things we’re talking about we’re talking about the lake house and we’re talking about one Lake one Lake one leg is this all-inclusive thing lake house or the the object inside the workspace that is a lake house might not need to be everything that might be your if you’re using like The Medallion type architecture which everyone’s seeming to talk about right now there’s a lot of videos coming out of that about The

13:12 videos coming out of that about The Medallion architecture you have this bronze silver and gold refinement of your data at that final layer of the gold layer maybe there’s a handful of lake houses that you’re exposing the data through and then Tommy and I were talking yesterday around well how does this work I could have shortcuts in my lake house that point to other things but I still direct Lake on top of those things I think so more testing is needed from my perspective but now we’re adding like shortcuts to things and so the data sits in one place but doesn’t actually get moved anywhere which makes sense I think from a effort

13:43 which makes sense I think from a effort and movement standpoint read-only tables would be great but I don’t I still don’t think I agree Tommy with the whole it is soon I have to do some more testing because I can’t I haven’t been able to get it to work I haven’t been able to play with it enough but I have to think if you’re going from the the lake a lake house into a power bi data set I don’t think you have to take every single table to make it work with directory I have been able to disprove it though well not if you’re going to from fabric to in a sense of power bi data set in

14:14 to in a sense of power bi data set in the service yes you are connecting to all the tables unless you yeah unless you click a new data set but then you’re recreating everything from the ground up which to me is redundant because if I have the majority of the tables that already have relationships and they’re already going to have those same relationships well why would I have to recreate that from the wheel there are the similar tables and I think that’s why through our questions I’ve been ruminating if that’s the correct word or at least sitting on the bigger

14:44 or at least sitting on the bigger question on how big is your lake house like how big should it be how many tables and different data sets should it store and maybe that’s the question where we haven’t really answered or haven’t discovered so much does your lake house have 25 tables does it have only does it have tables for how x amount of data sets I I I would say I don’t know if I if I’m doing like a

15:14 I don’t know if I if I’m doing like a medallion architecture right I’m only exposing so there’s only a handful of tables that I want to really expose that are like cleaned ready to go that are ready and designed for people to start consuming from data sets right there’s a there’s a there’s a cleaning effort here that’s occurring and actually this this dovetails very well into our next conversation that’s going to be around more of this anyways but I would Envision like if I had to divide these things and how I’ve been traditionally building these architectures the one lake is everything everything goes on like this yes everything one link includes everything when you start

15:45 includes everything when you start talking about the lake house or the object inside power bi. com that is the lake house that’s the item that has you lake house that’s the item that has it’s it’s groomed down for a know it’s it’s groomed down for a particular data set or topic of information right say this is all the data that we need to give to marketing or whatever that looks like and so again with these shortcut elements I could still create one table of data but I could reference that in multiple data sets or potentially multiple lake houses as shortcuts right so those are shortcuted item that means there’s only one official table of the information

16:15 one official table of the information and then everyone can reuse or re-look at that same pointer to that location what actually I think honestly would make a lot of sense but to your point Tommy Tommy you do have this wasted effort of well if I’m going to use a fact table and a dim product over and over again well there’s some relationship there that needs to be stored or captured and reused in all the data sets I don’t see that being any different than what we do today with a SQL Server having tables and we go look at them through views

16:45 and we go look at them through views through Power bi desktop every data set has to have a rebuild of that information so I would argue right if you’re talking about like Enterprise level reporting there should only be a handful of data sets that you need to service the broader part of the company and you try and make them very generic and basic you don’t build a lot of complicated effort into those data sets the only thing no and that’s a good point but you’re still now dealing with again the best practices were still yet undiscovered of how many that’s true

17:16 undiscovered of how many that’s true data sets per lake house because right now one lake house we’ve already discovered is four different artifacts plus now a data set plus now a report so what are you dealing with seven artifacts per report or per data set in power bi because there’s this four data flow artifacts that are created there’s the lake house there’s a warehouse there is a data set and there’s a report so one lake house how many reports or how many data sets should that turn out

17:47 how many data sets should that turn out for you because now think about the lineage even if they have the lineage view you’re still dealing with already something very complicated from one real true power bi artifact or two power bi artifacts the data center report this is why I think fabric has convoluted this whole Space of how we design and build engineered data right aside from Dev test and prod environments one lake or fabric is now bringing all

18:19 one lake or fabric is now bringing all the Enterprise grade tooling into Power bi so before this moment we would just easily clean draw a clean line and say well that’s data engineering that that’s out that’s before I get it to power bi it’s somewhere else and maybe you do a little bit of data engineering with data flows but yeah data flows works for if you design your data flow well it could handle 100 million rows but if it goes really large billions of Records we’re probably not using data flows right so it’s a very to me in my mind it was a cleaner break point between

18:49 was a cleaner break point between where does the data engineering occur upstream and then where does it apply in power bi this just brings all the Upstream data engineering into Power bi so does that mean we have now one workspace for all the data engineering and then that feeds a handful of other workspaces that are connecting to those sources of data or connecting to those tables I don’t know yet I think the Pat I think you’re right Tommy I think the patterns have not yet been established or best practices and what that would look like

19:21 so have you been oddly quiet up there well I I think the the unique aspect of this for me is is just trying to quickly dive into this this aspect of of like direct Lake right and one of the things I’m pulling from the documentation is like the the direct like mode does not require the lake house end point so because like we’re talking synapse right so but it does require you to set up lake house for some structure I guess so I don’t I’m not talking about like the

19:52 I don’t I’m not talking about like the complete modeling experience I guess of what synapse right like apparently this is to get you into the a a bucket where you can use the Delta parquet formats from one link so I haven’t played directly with this but if I’m gonna introduce some additional thing into a model it doesn’t at first glance seem like I should have too many problems to to still have my model in the power bi experience I’m just Express like connecting two different one Lake

20:22 connecting two different one Lake sources for additional information now we we spawned off into as I connect all these things or create these objects how many get created within one lake or my purview which probably is part of the governance conversation that we wanted to have today but like either way I think maybe diving into you either way I think maybe diving into the direct Lake and know the direct Lake and the the use cases for it would be a good thing for a different episode I think there’s Nuance in all this that like the actual testing out needs to

20:54 like the actual testing out needs to happen happen I’m just going to push back against your one comment there tell me around like one direct Lake requires you to take all tables from the lake every single time I just I’m not to me in my mind that feels like if if powerbeat forces you to do that it’s a Miss in my opinion that’d be just a miss right I would not want to be able to like the idea the idea of the lake house is to pull everything in and selectively pick a couple tables that make sense for the data set beyond that and it I would

21:24 the data set beyond that and it I would like it to still use direct query or directly when it’s using those elements but maybe that’s just something that’s going to be coming out after preview right maybe that’s something that’s going to be landed later on but I would I would argue that point like if it doesn’t do that now I need it to do that that’s a feature that’s required right I want to be able to pull stuff from the lake house and not have to have the entire thing loaded into my power bi data set for it to just be directly have you looked Tommy when you build a data set that’s doing directly a key things things do you have a different data connection

21:54 do you have a different data connection does it show you something different there that’s doing a direct link actually the few times that I’ve tested it and it could be the bugs early now where it would create a default data set which when I try to load that in using the the direct import mode I had to create another data set so yeah you can select your your tables in a lake house to create a new data set but the default data set would bring in nothing I had to create a

22:20 would bring in nothing I had to create a new data set but I would always say this is out of sync with the default data set so it was all still very seemed a bit finicky still at this point yeah yeah I’m creating a data sets like you need to use the default data set if you’re going to create this but yeah yeah oh that was blank gotcha okay yeah our parking lot’s getting pretty big so

22:50 all right with that let’s go let’s move into our main topic for today so I think today our our main topic is going to be talking about this wildly popular article written by Matthew roach and Matthew even calls out this is one of his fastest running articles or one of his his introduction on Microsoft fabric was was a record for him this is Maybe going a bit deeper around okay what does Microsoft Fabric and one lake look like for organizations how does this start driving adoption how do you think about this new framework that’s now in place

23:21 this new framework that’s now in place as you look at the lens of your organization and working inside lake house type architecture so I’ll throw the link here in the chat window as well this is from Matthew roach on ssbi polar is his blog that he uses he has very insightful stuff a lot of his content is what is out on the adoption roadmap today currently so I found immense amount of value from listening to and thinking through the things that Matthew produces so here’s the article I’ll put that in the chat window

23:51 the chat window Tommy gave us a a stamp an overview let’s maybe start at the very beginning what’s he talking on here and where does this start us off yeah so and there’s a few quotes in here that I’m hoping Seth can can add just because oh there we go we have the quote King but the idea here is now we’re gonna finally have that conversation the good old conversation that we’ve talked about power bi and data governance and adoption and introduce this to fabric Matthew roach does have some

24:21 fabric Matthew roach does have some views in terms of how the current a governance documentation for power bi plays into fabric making it as easy as possible people to work with data is one of the tenets of the adoption roadmap and that government is not about saying no to people but it’s about enabling and empowering people so part of the Crux of this is again I think initial opinion here is don’t change anything that you currently

24:51 don’t change anything that you currently have your tenants for data governance and adoption with fabric now just because there’s new artifacts and there’s a lot of the Crux of the conversation we’re having I liked how Matthew’s initially started on there is new technology showing up but we need to sit back and really first Define what is governance I think when people think about governance they have different opinions based on how they’ve worked on companies data sets prior and

25:21 worked on companies data sets prior and for some people governance is just shut it off can’t can’t see it can’t touch it for other organizations governance has been more around working with and still enabling you to do your job but do it with guard rails and I think Matthew does a great job at the beginning very quickly calling out they wanted to read like summarize what’s in the governance let’s talk about data culture or an organization and I feel like this article is pushing us more towards hey organizations of the world look there’s

25:52 organizations of the world look there’s going to be more power given to the business users there’s going to be more ability for people to build things in a clicky clicky draggy droppy format to steal some words from Christian Wade but with this is going to be it’s going to emphasize it’s going to highlight more of the weaknesses in your data culture if you have poor data culture all this is going to do is highlight more of those those challenges you think that something where he’s kind you think that something where he’s going with the beginning part of the

26:22 of going with the beginning part of the article it is and it’s very it’s the same vein of what power bi’s data governance is again the biggest thing that and I tell clients who governance is not about saying no it’s about being able to empower people the right way and yeah there are sure committee meetings there are sure the the conversations on rules and responsibilities but the end goal of governance is not just turning it off for people it is allowing people to actually utilize data the right way now all that being said doesn’t mean it’s a

26:53 all that being said doesn’t mean it’s a free-for-all wild west and we’ve talked about that ad nauseum but here’s one thing I will have I right off the bat I will have to at least push back or at least hold reservations we’ll say in in a a nicer way with the article we are immediate pushback it’s fire away so it just won’t stuck out and this would be maybe a great conversation to have with Matthew as well but we are introducing new technology we which introduces new process it’s very hard just to say well

27:24 process it’s very hard just to say well let it go because that’s the initial tenant because we are dealing with a a new platform that which needs to come with his own questions first and going into this how does this affect the people how does it affect the current people using the technology in the process to ultimately get what the solutions we’re trying to get how does it affect new people rather than just the current people who have been dealing with it now do we introduce new people into this there are a lot of things here

27:54 into this there are a lot of things here where I still will push back a bit where it’s a OneDrive for data which is one of the tenants that Fabric’s been doing because fabric doesn’t or a OneDrive doesn’t necessarily have for select people well I guess it does but like the you could have the security around what document or what information you’re sharing in OneDrive in SharePoint there are still those roles there on who’s allowed to edit our corporate PowerPoint and if you’re to use that same way we’re

28:24 and if you’re to use that same way we’re dealing with I’m always going to be thinking and considering the the fear of providing the wrong or the inaccurate data provide and losing the difference conversation one drive for data and security how is how is the difference

28:55 how is the difference like what’s the difference between Fabric or enabling that experience to a wider audience like what’s the security concern concern so think about who created data set in the past or at least think about the the engineering side of it I see I see you smiling so go ahead I’m smiling at a comment thank you okay yes bless you Seth so but think about the engineering side before The Medallion approach or who is in charge of that right of of going through that that’s now anyone in a

29:26 through that that’s now anyone in a fabric workspace can do sure sure but they could also be doing that today I also say I would also argue Tommy like depending on where the data comes from right if people are just saying I need access to data I see probably more often than not people are just saying let me just dump it out and in many organizations yes you may be doing the whole Medallion architecture but once you get to that final layer of information that’s coming out of it so typically Medallion is owned by a particular Department it comes out of it

29:56 particular Department it comes out of it because they’re the ones that have access to the production servers that can get the data out of there and that’s that’s usually there’s a gate that you don’t want to have business users running queries against production I running queries against production that’s the whole purpose mean that’s the whole purpose here that system is not designed for ad hoc queries the goal is bring all that data down to some place and then let the business play with it and beat up this other analytical tool that is not running the day-to-day business because if we shut down the ability to make orders every everyone’s going to be angry right we

30:26 everyone’s going to be angry right we can’t sell we can’t make money so so to that point right yes we have this mendelian architecture but you’re taking a very I. T focused approach on this one and I think there’s actually a very business facing approach as well the business is already doing a lot of stuff anyways I’ve even seen organizations setting up their own SQL servers and their own data processes and everything inside a business unit that is not controlled by it so yes you have this Medallion architecture piece but I don’t think as many organizations are as organized and

30:56 many organizations are as organized and how clean that data should come out of that Central system and the business users aren’t going right back to excel because that’s what they know and they’re building a whole bunch of other copy paste moves and other data system things and yeah they’re doing the same thing that your enrichment of data is doing but they’re doing it all in the guise of the inside the business it’s inside Excel sheets well but but there’s a there’s a there’s a problem with that statement well if they’re going to go back to excel then we might as well open up the floodgates

31:26 we might as well open up the floodgates for them since that’s what they’re going to do anyways there’s that not necessarily data culture so I think I think there’s a there’s a I would not agree with that statement just because we’ve done it that way in the past doesn’t mean that’s the way we should do it in the future I’m I’m just making the argument here that is what we’re bringing into fabric now is two additional roles or personas and potentially even a third persona that is there are now other people or other

31:57 there are now other people or other other workflows that can be accomplished inside power bi with with the help of fabric at this point right so that’s I fabric at this point right so that’s my point really basically is mean my point really basically is we are now the there’s had traditionally been two teams the team that knew how to wrote a bunch of write a bunch of code and the team that took that output of data tables and were able to manipulate it into insights reports and things that that the business could look at every day and make decisions on right so to me that felt like two teams

32:27 that felt like two teams because you need to have that user have that business domain knowledge what we’re doing now is we’re now allowing the same ecosystem for all those data engineers and now even data scientists and potentially even production environments to start playing inside power bi and fabric and a good amount of that I agree with we’re there’s finally that enablement of just from the business side of it I could think about oh well with data flow is now we can do a lot more with power apps and now we can enable just a marketing team my goodness what you

32:58 marketing team my goodness what you could actually set up for them and then push to other the other systems one big thing that one Lake’s been talking about is the ability for the apis and be able to connect to a lake house via API so you can really push this information out which is a huge win with that being said I think this I I honestly think there’s still a simplistic view that we’re looking at this on we’re thinking of a few data sets we’re thinking of a few scenarios not an entire department or a much larger Department with a lot

33:30 or a much larger Department with a lot of data already created where how does that actually look in in one on one lake or how does that look in a fabric workspace to where’s the source of truth if a organization what is what is what is the source of Truth for a business unit right now without one link it probably comes from the data engineering team and it probably does still come a bit what comes from the data engineering team this source of Truth the tables the the library table every table every table that ever created for business units right now that the reports that

34:01 units right now that the reports that they’re generating with power bi all come from government and I’m not saying 100 and this is our yeah I’m definitely not saying what you just said just said I’ll I’ll check the tape let me let me re-ask my question sure what is the source for business units for the data that they’re using in their reporting today today and in no not all of it’s from governed places so where else is it from it does come from apis and it does come from non-governed places at the same

34:32 from non-governed places at the same time and I want to make this clear that doesn’t mean that that’s always the right way which is usually a lot of problems that companies have is looking at a report going is this correct is should this be checked with Finance that’s part of power bi’s I don’t want to say flaw but that’s part of the curse and the blessing of power bi is anyone who created a report okay let me let me rephrase my question yes power bi is a very powerful tool for the business to connect to many different sources of

35:03 to connect to many different sources of information including Excel files including many different things that they may store behind security in their own teams areas whether that’s OneDrive SharePoint like whatever whatever how do you govern that in today’s world what is data governance look like even in a like a very well like a an organization well down the path of adoption what does governance look like in in that world for that business unit today today you’re asking me these completely

35:35 you’re asking me these completely straightforward questions that you and I both know have a lot of entanglements to them so I’ll do the I’ll do the best I can I’m asking a simple question right you’re you’re making a lot of points about why Fabric and all these things are bad for data Governor or challenging to challenging the I guess positive vibe that that Matthew has on the fabric solving some of these governance problems but like if we’re bringing the negative what does today look like from a governance standpoint for those

36:05 a governance standpoint for those objects in an Ideal World or are you just saying I’m just saying in real world today in Real World what does governance look like like if you’re to have a Governor’s program that actually was standing up so to speak there are levels of call it certification call stamps of approvals at least farther Upstream I think at the end of the day regardless of the enablement there’s at least this approval okay governance is guidance guidance and guardrails hey business units you we would like you to do this

36:36 units you we would like you to do this with your data this is where we well how you pull from how you access like like and or and or store or whatever the case may be but it’s not like they’re all pulling from standardized data sets sure they can connect to many different things but they are creating artifacts that they’re using within reporting which you do have like technically rules around how they

37:06 like technically rules around how they should manage but as a governance body how do you ensure that they manage that in today’s world part of governance is understanding are people actually doing what we want them to because there’s a reason why we’re spending all this time to figure out the patterns that we would like them to follow because it’s going to a serve the company better because it’s not going to be random data there’s going to be some quality around it right people have access to the things that they need

37:37 access to the things that they need is why we we developed the governance and but part of that is also making sure people are doing that it’s the simplification of like guiding people who like may or may not know about the particulars of data and reporting to the levels that we do into the easy path here’s how you navigate all this stuff here’s the data access that you have here’s the things we want you to do in your business unit and many different business units do things slightly differently but how do that they’re doing

38:07 but how do that they’re doing it in today’s world like we’ve talked a lot about governance and adoption and the challenges with specifically just power bi are what there’s a lot of the data movement ETL where data lives Etc that’s well before the purview of power bi right like these are just different options this is part of the difference now in the world or power bi is the ability to create a data set that now lives in a place where the con organization can consume and and sense

38:39 organization can consume and and sense believe that’s either true or not where it’s easier than ever to do that Great era certified data set way Downstream right if that’s an object and how many times yeah yep and I think that’s to me though we’ve talked about we’ve talked about ad nauseum the need for the certified and promoted data sets coming from the bi team that that’s not just a random user doing I understand what

39:09 random user doing I understand what you’re saying in terms of how is that different from fabric it a bit it is because it’s going farther Upstream now where at least the data set lived on this particular Plateau or particular plane of where the data was getting created from where the model manipulation was coming from but we still talked about that important need for hey if it’s going to be certified it has to be reviewed by the bi team we’ve still talked about that sure that’s not it and the same conversations we had doesn’t mean you can’t create data but we how many times have especially you

39:39 we how many times have especially you and Mike brought up well has to be certified we’ve talked about labels or put a badge on it sure what it what does that process look like without in today’s world where a governance body would be like hey we would expect that as as this group is accelerating in their pretty heavy user power bi but we haven’t gotten any any feedback around like their certified data sets like unfortunately are they doing things right I think it’s rare I think the the having this Ideal World

40:10 think the the having this Ideal World we’ve talked about with the certified data sets why is it here because it’s hard to manage it’s hard you the resources put in investing the resources investing the resources in the right people and the process for data governance this is something where people have migration say I haven’t like I have I have my central bi team and part of those guys have their responsibility is governance is the volume of data that’s getting created let’s say yes this business unit

40:41 created let’s say yes this business unit this business unit heavy user power bi but with that like we don’t have any certified data sets that have come out of them of them okay okay what what do you do as a governance group like are we just assuming that they’re operating on bad data do I have to engage with them manual like to figure out like what data they’re using and and how would I use how would I do that in today’s world Seth you’re bringing up situations where we’re talking about the lack of whatever the situation of the program brought up in your I know you already have a answer

41:12 in your I know you already have a answer here so no I I guess my point is if governance is about making sure people like follow a process that benefits the organization there isn’t a whole lot outside of discussions or building a process within an organization that people people have to follow to follow but how do you administer that how do you ensure that people are following your processes ensuring that as as popularity of a

41:42 ensuring that as as popularity of a report goes up all of those data sets start to become curated or certified or valued like to the point where we’re not going to have ambiguity anymore right like there’s a visibility problem is where all this is driving at right in a current world my building my processes and guard rails for individuals to do things or not do things is yes security related yes it’s making sure they have access to things but the point here is governance should

42:15 but the point here is governance should not only guide but make sure that things are operating the way you would want them to them to and and for all intents and purposes because a lot of raw data a lot of transformation a lot of back-end things are invisible unless done in a certain way within power bi data flows or like even data sets whatever it’s still very missing and like I find it interesting you’re

42:45 and like I find it interesting you’re pushing back on really key points of of potential improvements to this that fabric is bringing forward mainly around insight and oversight on an administrative level to understand whether or not your processes that you’re creating as a governance team are being followed well allow me to allow me allow me to I don’t want to say push back to what you said you made some very good points that go into what I’m trying to say here if

43:17 go into what I’m trying to say here if you’re going to follow let’s say mac roach’s maximum the farther Upstream there’s a reason for that it’s not just optimization it’s about the accuracy and where you’re pulling from data quality data quality we’re not talking about data quality we’re talking about data governance you can’t say they’re not related at all they are two different things things I think they stem from the same root no do they do they directly from the documentation think of

43:48 directly from the documentation think of governance as a set of established guidelines and formalize policies all governance guidelines and policies should align with your organizational data culture and adoption objectives governance is enacted on a day-to-day basis by your system oversight Administration activities from a higher level level like non-power bi related as defined by the data governance Institute data governance is a system of decision rights and accountabilities for information related processes executed according to agreed upon models which describe who can take what actions with

44:19 describe who can take what actions with what information and when under what circumstances using what methods it’s very much what is going on in my ecosystem and is it operating the way I would want to from a security and data access level perspective I data quality certainly really close

44:40 I data quality certainly really close but not the same thing I never said the same thing I said there’s a correlation one affects the okay that’s all I’m saying and I’m not saying yeah inequality sure sure there’s also different there’s also different roles so I’ve been taking a lot of notes on the side as you guys have been have been debating back and forth I’ll say so as I as I I feel a couple points here that are very relevant so I think there’s this conversation of

45:11 I think there’s this conversation of governing data and I think Seth I think I think you hit the nail in the head on a lot of points there right it’s it’s around letting people play with information but then being able to say when people are building things how do the decision makers of the output of that how can they rely and I think there was a comment in the chat said it’s about trust right do you have a process that’s documented do you have are you using power bi certified data sets what is your process to get something that is certified I think these are all actions that go

45:41 think these are all actions that go along with the governance piece that help you trust what’s going on with the data right there’s there’s rigor behind certain parts of this I’d also argue in some cases here not all data is equal an equal value right so there’s definitely going to be this portions of data or tables or information you’re going to make is going to be the report of the personal reporting level fine it’s less governed than it’s less combed over right as opposed to like that financial data would have to be highly governed and people get

46:13 to be highly governed and people get fired when financials are wrong so I think there’s also like layers of information here that you have to be able to make decisions on and so at the when if I’m if I’m an executive showing up to reports that my teams have made the data is going to be messy my organization may or may not have a good data culture around that information I have to be able to as a leader come in and say can I trust the information that’s being presented in front of me and I think that’s a combination of you and I think that’s a combination of having good data stewardship know having good data stewardship talking about what the quality of the

46:44 talking about what the quality of the data is what is our process to get data from where it started from to where it gets to my eyes and then being able to have confidence that I can trust the team under me that’s developing that data and has done the rigor they need to to get that that data or information into a decision-making stance so I so I I think to to me like trust is data quality data governance is the thing that rises like raises the bar where you get better quality objects to

47:16 where you get better quality objects to interface with that if you’re saying governance is yeah you feel governance is the is the side of the world where we’re saying so does governance is the guard rails essentially play Within this bound so for his right so for as much as I leaned into Tommy in the OneDrive in data security because I think there are I think fabric does bring a very interesting ecosystem to play that actually helps governance exponentially

47:46 actually helps governance exponentially with Discovery and how data is used if fabric is the platform that everybody is working on because as an administrator like even some of the simple things around the increase in administrative oversight of workspaces within an organization right no visibility only then only accessible to API now it’s part of the UI where you can dig in and see who the users and the owners and the administration administrators of all the workspaces are and now like all this

48:16 workspaces are and now like all this means is I have more knowledge to understand how the organization is using things and whether or not they’re following the policies that I would I would outline one of the key points that you Breeze by Tommy that I think is a valid pushback granted it’s like since this conversation is so heavy and large when you’re launching a new new like feature or something it could take a long time to describe

48:46 it could take a long time to describe but what’s what’s interesting to me or I said or I think is a Miss from Matthew is if if if if so much of governance is about developing those guidance and guardrails and the processes in an ecosystem that we we’ve just had this fictitious description of like how chaotic and crazy it is if gov if companies have gone through governance strategies and platforms to that work heavily with power bi and it’s all

49:18 heavily with power bi and it’s all around how to navigate platforms and processes processes you just introduced a complete rewrite rewrite of how your how you would want an organization to navigate fabric like this the this from their own documentation strategy challenges like having an understanding of how people navigate through things people challenges process challenges skills and data literacy challenge like you’re starting over you bring fabric into the

49:48 starting over you bring fabric into the ecosystem rewrite all of your governance a little guidance might have been helpful to like figure out how workspaces and the explosion of objects and all these things that all of a sudden show up in an ecosystem for us are supposed to be explained let alone the workspace paths and how teams are supposed to be working within here like dude there is a a wealth of information that I would have loved to have had in this conversation as we just re-platformed everything data related in

50:18 re-platformed everything data related in power bi as opposed to just saying hey what what this should help out [Music] no one said you need Jurassic Park level clearance in order to use fabric in order to get onto the island no one said that but if we both agreed and I’m sure Mike would agree with this too the majority of organizations have a hard time governing power bi that fabric is now something that we should have a word of caution with to quote the great Steve

50:49 of caution with to quote the great Steve Martin Excuse me no I don’t I don’t agree with that because like from a governance perspective you’re saying like yeah it was hard enough with power bi the hard enough part is power bi started a piece of this which is the way Far Side Far Side like it drove for these conversations and we were already having them with one piece of a puzzle what my argument is is okay okay awesome we just unblocked a whole lot of visibility we unblocked or increase

51:21 visibility we unblocked or increase the potential problem of business units doing things but not like from a governance standpoint it’s fantastic because like all that data would come into an ecosystem that I could now have visibility over that that opens up huge possibilities for us to create better processes in in guardrails for people but I think that’s the problem we don’t have them yet we we have yet to create those processes and to me before you get to the wild west it was so easy to do that with a data set and a report before data flows were even

51:52 and a report before data flows were even introduced this concept of the Wild West was already apparent in the same UI well it’s a fundamental thing challenge yes business intelligence a funny yes yes but now we’re introducing all these other Concepts and services into the same Mainframe so to speak no one’s saying don’t play with it no one’s saying don’t have a ball play around with it try it out however there is yet to be a concise and solid grow-up story with what we’re going to do and yeah I

52:23 with what we’re going to do and yeah I agree with you okay I agree with you and my pushback is I would have loved to you’ve you’ve some people have seen this too a lot longer than we have right like it’s been great to like this is absolutely part of it because if like preview is a very short period of time like the this is a challenge that all of a sudden it’s like okay how do we navigate this new space and we’re having conversations about this so yeah like and and I think that’s my don’t agree with it I will agree with this don’t claim victory with the tool for all of these things if

52:55 these things if without without knowing right or at least giving some breadcrumbs this is this is the Crux of the my argument today no one’s saying what only fabric for it I’m saying yeah let the marketing team do it however let there be a process for the grow up story for all these artifacts and components being created way before we get to the point where there’s a hundred pieces of information in a workspace and that technology aspect of it will influence if people are going to use it when it also becomes becomes very confusing how

53:25 also becomes becomes very confusing how can we prevent that how can we prevent the complication and the all of a sudden this hype and whatever that truth or disillusionment is because well I don’t know what to look at anymore there’s 50 pieces of information who’s doing that and people will go back to excel I do agree from the standpoint that if if they’re already like you look at the adoption and roadmap documentation and governance specifically around power bi the same challenges of like strategy people process skills and data literacy

53:58 people process skills and data literacy it the those are magnified I don’t want to say it’s like 10x or if I were like whatever but if you think about how off the rails something could go from a governance standpoint is it easier or harder in fabric to get your head around I would say it’s easier it might be easier easier West adoption strategy all of a sudden is like not so wild west because I’m an admin and I’m like okay I see what you’re doing we want to guard we want to we want to

54:29 we want to guard we want to we want to put here’s here’s how I want to guide you here’s the objects I think you guys could use and maybe it helps facilitate that conversation a little bit more I think if nothing else you’re going to get better conversations around if you don’t have solid processes today around what governance and looks like inside your organization I think this was going to accelerate those conversations and it’s going to force people to really start thinking about okay now that we have this capability and again part of this is there the one Lake element here there’s

55:00 there the one Lake element here there’s a concept here that you can one Lake it right to your computer right so I could literally have a file on my machine that I then drop into this one Lake object that immediately pushes it out to the service so that it can be available to me in any of my reports yeah what I’ve what I feel like is happening is there’s been a lot of historical information where there’s been such a hard division line between what it does and what the business does it was very easy and the lines were very clearly drawn data governance data quality these are things that come out of the it organization because that’s

55:31 of the it organization because that’s that’s what they were able to own but now the technology is changing and I feel like since things have gotten a lot less code based it’s low code or no code Solutions because a lot of this lake house stuff transforming things we can now do a lot of things that we weren’t able to before in just a UI so I

55:49 weren’t able to before in just a UI so I I know all the code that was written before this it appeared there’s a lot of headaches of getting things to talk to each other across different Azure Based Services so now that this is all together we’re now able to have a whole a holistically different conversation around okay what does this look like for organization what should we have for policies and I think this is one of my comments here I made earlier was this is just going to accelerate the need to have that data center of excellence or that collaboration across multiple teams

56:19 collaboration across multiple teams because again it an organization has many different goals each department has their own objectives they want to see the data shown their way that makes their team look the best right there’s there’s so many competing opportunities someone’s going to sit down and say what’s the opportunity cost on those things where do we spend our money where’s the best place to put our effort so this stuff adds value and if you have I I would have I would fully believe or I’m going to be so naive is we can’t always assume every

56:50 naive is we can’t always assume every organization we’re going to walk into is going to have that strong leader at the top understanding exactly how data should be handled there may be just leaders at the top that just say just figure it out just get it done and and you have a looser governance policy at the top which will then that again that that everything I read Around data culture start to the top whatever the leadership adheres to or or messages to the organization is how the rest of the organization will figure things out data wise

57:21 figure things out data wise this is just a more Consolidated effort and so I feel like this fabric has been very good I feel like it has a lot of opportunity I definitely feel there’s a lot of rough edges but I think we have to focus on the technology has now leveled up we now have to level up our people and we have to level up our process to accommodate the additional features that come with the technology like overall my final thoughts are I think ultimately this certain like fabric would certainly simplify aspects

57:51 fabric would certainly simplify aspects of like the business experience in obfuscating some of the things that are just innately happening but also opening opportunities for them to leverage more Enterprise platforms think fabric does provide a super powerful platform that that simplifies that oversight and produces a better which ultimately produces better results for the organization like if we build a bunch of governance processes are people following them do we have that oversight within the organization like yeah man like it definitely opens a lot of

58:23 like it definitely opens a lot of possibilities and even from some of the administrative stuff I’ve seen so far I’m super impressed what I think it does initially though is it complicates the governance team’s development of new processes and procedures because it’s so wide open like I agree with that like it’s bringing a whole new way in which organizations could or need to engage in this platform and yeah you’re rewriting stuff like it’s com like how do we do that what are what

58:53 com like how do we do that what are what are the best practices guidelines how does it fit into the adoption of fabric right and maybe those documents are being generated and we’re just a little ahead of the curve here but at the same time like if we’re trying to roll onto this it’s a key part of the conversation and actually one of these points that irritates me in the like the marketing fabric like could lean into this one man right fabric we’ve we’ve introduced all these things there’s now oversight Administration this helps you accelerate

59:26 Administration this helps you accelerate your adoption roadmap strategy within an organization like Mark there you go marketing yeah and that’s and that’s the conversation it should be having right now right it should be it should be this is look at this new tool look how how fast we can help you move forward here’s some best practices and guidances around what data quality and governance should look like here’s a whole bunch of things at examples because what it what is the differentiator right you’re not first to Market with some with the oh no way in here like you’re you’re coming coming to

59:57 here like you’re you’re coming coming to play ball play ball but at the same time what is fabricab that no other platform has well it’s unified all this yeah so the more you lean into that and the more you can sell around the fact that like organizations can accelerate certain aspects of this might might be helpful yeah like let’s go maybe I miss my calling maybe maybe you did you’re an author no and I think the last thing from a great conversation today by the way love sparring with you Seth so always it

60:28 sparring with you Seth so always it wasn’t sparring Tommy was getting an argument Tommy I would love to tell you maybe that’s the problem no well just a great a good conversation that’s needed because there’s a lot here that again it’s all undiscovered there is no fabric governance probably because it hasn’t happened yet there’s there’s not a business that’s actually rolled this out at all we’re still in a trial period this is all very new and how does

60:58 period this is all very new and how does the people affect the technology how does the technology affect the people we don’t know this yet we also don’t know what that truth or just addition we also don’t know how much is going to cost us we also don’t know the cost of this this is something that’s not bring up what I really wanted to today because that’s like my biggest question mark right I’m gonna I would like I love the idea of this thing I like that so many things work together yeah there’s a couple and you let everyone play but that’s a whole I didn’t even want to bring that up today but yeah that’s a big point but I think there’s there’s a lot here on

61:28 think there’s there’s a lot here on discover that we’re trying to figure out without actually seeing it in the reality in the real world on where where would people fall apart with this where would people really what what type of people is this going to bring to the Forefront we don’t know this yet so so I’m gonna I’m gonna end on the chat GPT and or Bing chat conversation here so we’re going to land on this thing so the question for today I think was very appropriately from the chat window was talking about how do we Define the difference between data quality and data governance it sounds like we’re doing a lot of banter back

61:59 like we’re doing a lot of banter back and forth between those two topics and it sounds a lot so here’s what chat gpg says data quality is refers to the accuracy reliability completeness and consistency of relevant data and relevancy of data it’s a measure of how well we meet requirements and expectations that is of its intended use data quality focuses on characteristics and properties of the data ensuring that is free from errors does not have duplications and any inconsistencies which I I would like agree with that definition I think it’s a very clear like we have rules we apply rules

62:31 like we have rules we apply rules against data then we can score it and say there’s a score out of the data quality on that thing that to me makes a lot of sense however I like the definition of data governance data governance on the other hand is a broader framework that encompasses the policies the process and the responsibilities of managing and controlling an organization organizational data assets it’s a guide of practices that ensure proper management protection and utilization of data throughout its life cycle data governance aims to establish a framework for data data related decision making

63:02 for data data related decision making accountability and compliance it involves defining data ownership and that’s one of my things I had on my list was Data stewardship right this is this is now something if we’re going to let data become discoverable from one Lake who owns it like you just can’t just throw data into the into the organization and say here you go and and not have anyone actually take ownership of it yeah but shortcuts are great representation of who owns it that data comes straight out of Finance right yeah exactly boom that’s the point here it’s

63:33 exactly boom that’s the point here it’s like like the technology is like adapting here to help us okay maybe we can actually have a policy around what that looks like right the last thing it says here it says it involves deep itself it involves Divine defining data ownership establishing data standards and policies and enforcing data access controls ensuring Regulatory and legal compliance and that was my point earlier right people get fired if you don’t do something right we’re not going to let everyone see a bunch of email addresses in our data sets so that’s that’s where those things need to come into play into play so I felt like I thought Bing chat

64:04 so I felt like I thought Bing chat did a really good job doing that one and writing a summary around that I also have one here from Google or Bing chat as well so that was that was chat gbt this was Google Chat it basically says about the same stuff oh is that Bard or beard no no it’s Bing chat that’s Google’s version ghoul’s version is Bing chat no no Google’s version is bared that’s what I just Bing chat is whatever Bing chat is yes GPT something something

64:35 yes GPT something something and and basically Bing was saying I’m essentially the same stuff basically at the very end of the the Bing chat said some interesting things around data equality and data governance are both indispensable for organizations that want to become data driven both have separate practices but they are fundamentally related often organizations purchase a data Quality Tool hoping it will solve their data accuracy and Trust however they first need to address their data governance to create a foundation for Enterprise scale

65:06 create a foundation for Enterprise scale data quality to put it simply you can’t have data quality without good data governance I thought that last sentence was like ding ding ding ding spot on like that felt really relevant to me and it was actually quoting a source there the data source was libra. com or something like that where’s finding that information so I thought that was a really good answer on that information anyways I think I think we’ve very much elaborated on this topic at this point

65:37 elaborated on this topic at this point so I will say if you really enjoy us arguing and getting hated about these different topics around data if you also enjoy us discussing and thinking these things through and trying to really push into these new tools like fabric we would love your listenership and we’d love you to share this with somebody else so if you don’t mind share this podcast with somebody else who may enjoy this conversation who may be also thinking through what’s the difference between data quality and data governance and how does this now interact with your fabric environment where is this going to fit I think there’s a lot of unknowns

66:08 to fit I think there’s a lot of unknowns at this point we’re gonna have to figure this out we’re gonna have to think through these situations and figure out okay what is our data governance what will this look like and how will it impact our company excellent Tommy where else can you find the podcast man you can find the podcast anywhere it’s available Apple Spotify Google podcast make sure to subscribe leave a rating it helps us out a ton join the conversation live every Tuesday and Thursday 7 30 a. m Central on all power bi. tips channels if you like these fabric conversations and power bi in general make sure to submit a mailbag

66:40 general make sure to submit a mailbag on what do you want us to are you talk about have a discussion on at the powerpi. tips slash podcast webpage excellent that works oh that’s fine sure fine sure it’s good enough sounds like I’m 55 but it’s a web page on the interwebs on the interwebs yes thank you all very much and we’ll catch you next time

Thank You

Want to catch us live? Join every Tuesday and Thursday at 7:30 AM Central on YouTube and LinkedIn.

Got a question? Head to powerbi.tips/empodcast and submit your topic ideas.

Listen on Spotify, Apple Podcasts, or wherever you get your podcasts.

Previous

Fabric Capacities – Ep. 226

More Posts

Mar 4, 2026

AI-Assisted TMDL Workflow & Hot Reload – Ep. 507

Mike and Tommy explore AI-assisted TMDL workflows and the hot reload experience for faster Power BI development. They also cover the new programmatic Power Query API and the GA release of the input slicer.

Feb 27, 2026

Filter Overload – Ep. 506

Mike and Tommy dive into the February 2026 feature updates for Power BI and Fabric, with a deep focus on the new input slicer going GA and what it means for report filtering. The conversation gets into filter overload — when too many slicers and options hurt more than they help.

Feb 25, 2026

Excel vs. Field Parameters – Ep. 505

Mike and Tommy debate the implications of AI on app development and data platforms, then tackle a mailbag question on whether field parameters hinder Excel compatibility in semantic models. They explore building AI-ready models and the future of report design beyond Power BI-specific features.