Domain Roles Part 2 – Ep. 230

In Ep. 230, Mike, Tommy, and Seth continue their domain roles series and map out how responsibilities shift as you move from Power BI into Microsoft Fabric. You’ll learn practical ways to define ownership, reduce duplication, and build trust using clear governance and promotion/certification patterns.

News & Announcements

• Explicit Measures Podcast — The full backlog of episodes plus links to listen/subscribe if you want to share this series with your team.

• Tips+ Theme Gallery — A quick win for adoption: standardize report styling while you sort out the harder work of model ownership and governance.

Main Discussion

Topic: Domain roles, ownership, and governance patterns for Power BI + Fabric

This episode focuses on the “who owns what” problem: when multiple teams build in the same tenant, you need a clear operating model for domains, shared assets, and how work moves from experimentation to trusted, reusable artifacts.

• Define domain ownership explicitly (people + responsibilities), not just a diagram—otherwise governance becomes everybody’s job and nobody’s job.
• Use a centralized BI/data team for standards, templates, and shared assets, while enabling domain teams to deliver their own reporting quickly.
• Treat shared semantic models as products: version them, document them, and set expectations for change management across consuming teams.
• Make the “promotion path” visible: dev → test → prod, including who can publish, who can approve, and what “ready” means.
• Leverage certification/promotion to signal trust, but pair it with real quality checks (refresh health, ownership, and usage).
• Design workspaces around ownership and lifecycle, not org charts—clarity beats cleverness when teams scale.
• In Fabric, be intentional about where data engineering work lives vs. analytics work, so responsibilities don’t blur and create gaps.

Looking Forward

Pick one domain in your organization and pilot a simple ownership + promotion workflow end-to-end—then standardize it before you scale Fabric to more teams.

Episode Transcript

0:03 dang okay and we’ll just jump right in all we are in we’re live Tommy I just hit the button watch the screen we’ll try again I I was waiting for so this may be a secondary stream that’s gonna be spun up here now because of some technical difficulties something happened and the connection was lost all right well I rambled for like five minutes while trying to kill some dead air while we were figuring out what’s going on going on we’ll take out a moment here just to see if things are going to be light up do we see it on YouTube yeah okay yes

0:34 do we see it on YouTube yeah okay yes yes we are back up all right we’re back up all right so let’s we’ll start over again hello everyone welcome back to the explicit measures podcast with Tommy Seth and Mike July 4th okay July 4th happened all right great moving on all right let’s jump into this episode again for the second time here trying to talk through what what are domains what are roles and has the technology for fabric changed where that data engineer individual will live does that individual now need to live in the

1:05 individual now need to live in the central bi team or is it actually now going to live more in the business area I think the argument we had a couple main talking points from the end of our last episode 229 just trying to think through who are what are domains the conversation really went off a rabbit Trail around roles and access to things and I think it’s been a long time coming to get to this particular conversation with fabric for those who’ve been subscribers and followers we’ve been talking a lot about all the mechanics the

1:35 lot about all the mechanics the different parts of fabric diving into our our our sweet spot so to speak on the implications of it but I think there’s a logical Point here for us to really focus on now with fabric who really is going to own what both from Ideal World of if we were to have all the resources but what’s going to be this compromise for what are resources and budgets and departments willing to do and what’s going to work because we are in also a very large unknown this is one

2:07 in also a very large unknown this is one of the spaces none of us can say oh I’ve done this before like with this technology people doesn’t say closer than others right so fabric using the bits and clicking the buttons of fabric I would say no right that’s what fabric is doing is not revolutionary so a lake house building common light like so the technology is to in my opinion Microsoft’s just now catching up to a lot of other data platforms right like Snowflake and data bricks but like

2:37 like Snowflake and data bricks but like they’ve already been doing this for years Microsoft has now centralizing the tools part of it right because they had it in Azure right you could have built the same architecture in Azure you might not have the access now the consumer can also be in the same place or contributor and also do the data engineering so that’s that’s my point though my point is that that those capabilities are now in the same tool which is what yes synapse yes yeah so synapse in a lot of storage was supposed to be that thing but you needed an Azure subscription and then it was like no

3:07 subscription and then it was like no we’re not going to give it to you and that’s what I’m saying yes yes oh here we are there’s not like they’ve also introduced completely new technology here it’s just that’s a good point access yeah the availability of it so right right yeah yeah okay so so there’s one thing that you said it’s like oh they they didn’t build anything new here and I would disagree from the standpoint that would you would you have said the same thing about power bi absolutely not did they utilize tools and Technologies they already had to combine them into a different offering

3:38 combine them into a different offering is that what fabric is right so because of that I think it opens up if it wasn’t an interesting new business intelligence like tools set or method or ways in which like opens up access to a much wider audience in the way that they have we wonder we won’t be talking about it yeah we won’t have the logo stamped on our wallpaper so that’s true so I agree right it’s not it’s not necessary like there are enhancements to

4:10 necessary like there are enhancements to certain areas the long-term Vision I think think starts to knit all of these things together so maybe Mike I agree with you now but not like it I think Breaking Down the Walls of not having to build all those services and knit them all together yes opens up the doors for this conversation where we’re saying hey you conversation where we’re saying hey what know what business units could own all of this even though technically they do but they’re using different Tools in

4:40 but they’re using different Tools in various forms in many different applications or third-party tools or whatever the case may be yeah no true to solve the same problems that could be unified or can be unified now through fabric that’s a great point you make there because I’m a dovetail that there are other third-party tools that do data engineering and give the business control of those data engineering activities I’m thinking one of them is talent talent is a data engineering tool it helps you shape data it reacts is another one yeah how many how many third-party drag and drop ETF tools are

5:12 third-party drag and drop ETF tools are they’re out there exactly at the business uses exactly right yeah and I think the business is using them and paying for them already and if when you think about if I think about the data engineering or data the data ecosystem ecosystem the most expensive thing right now in the data ecosystem is the people the the labor to make things happen is the most expensive element of the data engineering so if you think about like power bi premium you can get premium for five grand a month if you’re

5:43 premium for five grand a month if you’re hiring a data engineer you’re probably looking at Double that just on a monthly basis for one data engineer for one business unit so the expense of that is much higher I think and if we think about our people process technology philosophy here the the big difference here and the big impact is the fact that now it’s all in the same stream if you in other departments had your own software your own tools your own login your own budget it was basically the like I See No Evil Hear No Evil if I didn’t know what that other department was doing I think there’s a big

6:13 was doing I think there’s a big difference now bringing it all into the same stream that it’s all going to be within fabric it can be viewed by multiple more people and that’s going to affect I think the process that’s going to affect the way people are doing things yeah could be viewed but not necessarily there there are absolutely opportunities for being siled in here oh yeah so when we say who owns what are we saying like who owns what in the fabric Realm I think there’s a couple places to talk about who’s owning what objects right

6:43 about who’s owning what objects right one is a challenge I think we talked about in the last episode 229 was really focusing on okay who are the people that are owning the workspace right and I think we clearly aligned on there’s there’s never really a clear ownership or or organizations are not putting in governance that says we need one or two admins we need one or two members and here’s what those roles are and here’s what they should be doing the majority of people working in a

7:13 majority of people working in a workspace are contributors people making content so I think in that in that context one that’s a weakness I think in many organizations we’re not really putting thought around what are those roles and training their people to be able to create those roles appropriately where I see things potentially being more effective here is allow Central bi is more of the governing body around

7:43 Central Central bis that governing body that decides okay you need a workspace request it from us we’ll get you the workspace oh and by the way here’s the resource groups that you’re going to have and make sure you identify who is the owner of this workspace when something goes wrong the data quality is incorrect if someone has if there’s something weird going on there we need someone to go back to and to really push on to fix the things and I think at that point in time you’re really talking about this Federated Center of Excellence now again this is a very generalized world but

8:14 this is a very generalized world but we’re looking at here just because I think if you look at different Industries certain ones are regulated more highly than others and so in certain cases you need more regulation or more control coming from that Central bi team in a regulated space so the heart of our our conversation that we ended on Tuesday was about does the business own the engineering the ETL the their data or is that going to come from the center to your point the centralized bi team and I think that

8:46 centralized bi team and I think that opens up I think that conversation really opens up now because of fabric if a company completely adopts it because the fact that you have everything accessible in one place yes and and I think to me when I as I think about this conversation I I’ve been trying to go back back and forth or see both sides of this because again this is a very new territory for all of us I would still lean on the bi side just because I think from a big part of this is

9:17 from a big part of this is you’re gonna have to invest a lot of resources if you were to have the business do this on the business understand the technology the business investing in the right process and my questions that come up is do they need to follow the same process as other departments is there a set is a a universal set of principles that need to be followed throughout the organization I’m gonna so you asked a lot of questions so I’m gonna try and break down your questions yeah I know you are so let’s let’s focus on one question at a time Tommy so let’s go in the first

9:49 a time Tommy so let’s go in the first one right one right will there need to be I think I got them all right here will there be a centralized pattern by which every Department should do their engineering I think there should be I think there should be something and that should be coming from the center of excellence that should be decided upon by those individuals that are saying okay this is how we’re going to go to market with our data and I do think in order to make the business effective there there needs to be a clear handoff between when the central bi team is developing tables and things

10:21 bi team is developing tables and things out of their Central bi space and then saying here’s what we’re gonna here’s what we’re willing to provide you we’re either willing to provide you a full report a data set or a table and I think as you work your way down towards the table level you’re giving the business more control to shape the data however they are now at that point in time whatever happens that data beyond that step that’s where I think the governance and the administration stuff really kicks in here because at when you start handing that responsibility over to a business unit

10:51 responsibility over to a business unit it’s up to them to verify and check their numbers it’s up to them to be have ownership around those things so go ahead ahead so I’m getting lost a little bit right in in General descriptions of what governance Central bi team and the relationship that exists today because when we talk about or at least when I think about how that would be structured in an organization that is that is with

11:23 organization that is that is with Champions right who are business people who know how to do some reporting if all of the sudden fabric is opening up this door for like heavier ETL heavier implementation patterns like engineering work is your expectation that those Champions are going to learn all that or where we started the conversation are are we supposed to pull people from the

11:54 are we supposed to pull people from the central bi team and give them to business units to support the business unit or are you talking additional resources need to be invested in in every business unit because we don’t have the levels of expertise in there basically questions so I think when you’re talking about a power bi Champion I think you’re already looking at an individual who already understands a lot of the Core Concepts around what data engineering things are being able are required right we we

12:24 being able are required right we we already you’re already learning data engineering when you’re doing power query you’re shaping tables you’re moving things around you’re getting the basics of what how to make that star schema model for that business for that champion and hopefully again this is another thing that I would kind this is another thing that I would lean on the center of excellence is of lean on the center of excellence is expected to identify those Champions and continue to educate groom help them build better models and things inside that space so I I think that is a purely a center of excellence role that is trying to facilitate grow the data

12:55 trying to facilitate grow the data culture as that user is part of the Community Practice what I think is going to happen here is I think you’re going to find the it budgets are going to continue to shrink there’s going to be less funding for that Central team in general but what will happen is the the users or the individuals that are going to be those sharp either call them power bi Champions or maybe call them that that data analyst the person who used to be doing the data modeling and the report building I think that person that candidate will start learning

13:25 that candidate will start learning either on their own or with company guidance or from the center of excellence they’re going to start learning more data engineering practices so we’re going to introduce GitHub or git to them we’re going to introduce More Design around star schemas they’re going to learn more about Dax I’m hoping that individuals continually hungry to learn more things and those candidates become right so now I have these code of two personas the data analyst which would be modeler report Builder and I have this data engineer and I see this as like a Venn diagram

13:55 and I see this as like a Venn diagram and all I’m seeing is those diagrams are getting it’s there’s a larger overlap between the two elements now because the tool doesn’t require me to know how to set up and configure synapse Source SQL or whatever I just get notebooks and SQL and dataflows Gen 2 by default okay so now you have this this role that’s going to be evolving into the business unit and I think it’s going to be a little bit of both some of your Champions will become in engineering and I think the business unit when justified by the cost

14:25 by the cost if they need to go faster with their data they will hire they will go find talent that meets that need so you just said less funding for the central team I think it I think that Central bi team is going to get less funding so it’s going to become thinner okay so so the bi team who within the organization any organization at the moment is the team that knows how to work across all these business units has all

14:57 across all these business units has all of the technical knowledge of all these back-end systems who manages these data sets certifies them does the extended leadership reporting are the modelers are the performance tuners are the trainers and the supporters of Champions is going to get neutered and you’re going to just expect that business users who know how to use power bi are going to elevate into engineering roles okay I see it all the time I see

15:27 okay I see it all the time I see a lot of it is at the best I see a lot of people in that engineering role taking on a lot more responsibility and becoming that that is a very common I think from the organizations that I’ve worked with or seen there is a number of people who are ambitious like to learn love their data space and they start out as I’m just building a handful of pottery reports then they get effective of that and then they start managing workspaces and then they start becoming that power bi administer and so now they have more area in that space in that bi

15:57 have more area in that space in that bi realm and then eventually what happens is a career change at some point happens and they move out of what they’ve been doing building reports for a single team and move more towards this Center of Excellence space Mike you’re literally talking about a Microsoft MVP which there are 200 in our space there think about the The Grove story of just someone who gets into Power bi accidentally their best practices around even if they’re learning data engineering they’re not learning best practices they’re just usually the the

16:28 practices they’re just usually the the majority of what I’ve seen is they’re flying by the seat of their pants and they’re not yeah and that’s that’s a young that’s a that’s a newer younger user yeah I got you but Mike you let that user sit there for three to five years three to five so what you’re saying though is and this is where like and this is where you’ve already got a lot of people you’ve already got a lot of people I have been using it we’ve already been using for them for nine years it’s not the same as an engineer who knows spark who knows lake house who

16:59 knows spark who knows lake house who knows any of the things that are like like here even in the MVP realm or professional bi people within power bi if when you start talking about parquet and lake house and Spark clusters and do it they don’t know anything about that they don’t and they’re learning it yeah they’re also coming at it with years of technical experience you’re you’re suggesting that a company

17:31 you’re you’re suggesting that a company like is is going to morph into fabric

17:34 like is is going to morph into fabric over three to five years and just and so many of these people are going to come out of the woodwork that they’re going to know and understand how to utilize this ecosystem okay you’re acting like these people are just gonna discover themselves I think no I think it has to be led by the central team this is this is something this is my pushback you’re saying in one hand oh the central bi team is going to teach them and then at the same time we’re going to eliminate like start eliminating the central bi team because we don’t need them anymore yeah they’re

18:05 we don’t need them anymore yeah they’re no i t i t is going to be so if you think about Central bi and I. T so there’s a couple things that are being managed right now it’s not the same thing well then I don’t know what your definition is then because I think that whole that whole Center role of making data models for individuals what we’re seeing is we’re seeing the trend of what previously was happening was all the technology pieces were inside a central bi team and all that was coming out was SSRS reports or here’s a cube what is produced by business objects what is yeah but what is your what is how would

18:36 yeah but what is your what is how would you describe a central bi team today like what I guess maybe maybe this is a two-part question it appears that there’s two sorts of bi teams that you have in mind one is this old school I. T yes like the owners of the data yes you don’t get and you don’t get access to anything anything here’s your SSRS reports here’s here’s here’s reports given to you like I would love an audience poll like do you guys live in this environment

19:06 do you guys live in this environment because I haven’t seen that in a while but I would agree with you if like a central data team is holding and restricting back all of the data within an organization then yeah neuter them but that’s not that’s not the team that’s going to be talking like doing training doing adoption doing Coe doing all these like new Adoption strategies of data within an organization because that’s not like if you live in a team like that or an organization they’re not going to be the ones doing that or promoting that it’s a

19:36 ones doing that or promoting that it’s a very very it would be a very large mistake to say that the same audience that loves power bi or has adopted that is also going to be purely become engineers the reason why power bi in there so the adoption of the builders is because I can do this little ETL I can build these calculations and I can build visuals I get to do all three that has been such part of the adoption because of the personality you may like one more than the other if also you think you’re going to find that many people first

20:07 going to find that many people first just to get introduced now to The Medallion approach and data engineering which is much more intimidating the first time that you see that than power bi ever was because guess what but you’re not dealing with a simple Excel table or CSV there’s a lot more knowledge and or forget what notebooks and pipelines are going to introduce compared to five minutes to wow the different products there’s a big difference in terms of the audience and people who are going to be willing to invest their time in that too I I agree with you but I’m not saying it’s not

20:38 with you but I’m not saying it’s not going to start happening the organizations so let’s let’s talk about just general Trends here right General Trends right is are are organizations looking to make more money or are organizations looking to put more money into centralized teams I think in some organizations the answer is yes but I would say most often more organizations are trying to move money away from centralized shared services and into profit loss centers where each business unit can then create a

21:09 business unit can then create a revenue and then from that Revenue take off their expenses right so I see the central bi team as being this shared common expense across the organization organizations that I’ve been in or been working with they’re usually looking to minimize those overhead costs those large Central teams and bring down those costs so not talking about people or headcount I’m just talking purely financial pieces costs right Engineers cost money and and so the other side of this is okay if I have a a

21:41 other side of this is okay if I have a a business unit a sales team an HR team whatever those teams are those teams are having a certain level of funding so depending on the team that you’re in the that team will likely generate a revenue stream and I think what’s going to happen is with the addition of fabric we’re seeing a shift in organizations that can actually apply more cost into a business unit or an area of the business that is generating Revenue more so than a shared function team so I think

22:11 a shared function team so I think generally what’s happening is the technology so but previously you had to rely on a central team because there was hard walls or Gates between the data and the machines that you were able to touch it was on-prem it was it wasn’t in the cloud we didn’t have a whole bunch of cloud-based Services the business unit couldn’t choose which tools they were using so they were forced to use some Central Tool that was generated for them but now with the proliferation of third-party tools everything being Cloud you can buy Salesforce in the cloud you

22:41 you can buy Salesforce in the cloud you can buy netsuite in the cloud or all these cloud-based Services I think what has happened is the business unit has now made the decision with all these third-party tools that they’re going to acquire they’re going to manage their own profit and loss and what will happen is the cost will shift from that Central bi team into business units that are generating revenues now would you agree with that trend trend I I agree with the trend but what I don’t agree with is the mixing of these two things right I think I think we are in agreement from

23:12 I think I think we are in agreement from the standpoint that that fabric the workspaces the domains everything within this ecosystem actually aligns with this strategy to allow business units to manage their own spaces and B the owners of their data right correct so do I think business units should own their own work streams yeah does that mean that there are opportunities for people within those business units to elevate into these much more engineering

23:42 into these much more engineering specific roles yeah are they probably already or like are some business units already doing some of this through other third-party tools yes 100 so is there a way for them to manage their own space a hundred percent yes yes and probably easier in this ecosystem of domains and workspaces but you take that and now you start disintegrating what I like like that is not the same thing in my mind as

24:12 that is not the same thing in my mind as centralized initiatives that a a new bi team a centralized bi team should be managing managing because that like what you just threw in was all these business units are all going to have their own people that do these things and what happened like they’re just collectively going to live in a Utopia of meetings and have other priorities outside their business units that are some of the eye it’s going to be like I’m not talking about like it’s it’s gonna be like hurting cats it already is today

24:48 I’m not saying it’s the the right way all the time but I’m saying the trend here is that will happen there will be more skills in those business units this is this is where I’m falling apart with you yeah you yeah the reason I was trying to make a distinction between what you’re saying centralized bi is which is a old standard of I. T controlling all of the data that is not what a central bi team is in my in my mind right now I think we’re just talking about different definitions because I think your definition of what the eye team is going going

25:18 going that is very technical in nature that can solve all of the problems that we talk about for 230 episodes right are not business are not business people problems that they can solve because they don’t it’s not their area they’re not technical experts right it requires a team to work across business units to do at minimum joining of data sets to pull together in a centralized like

25:48 to pull together in a centralized like be it ad hoc be it a warehouse be it a centralized model it is extended leadership reporting and dashboarding across an organization not specific to a business unit because you’re not replace even in today’s world you’re not replacing reporting from a business unit and nor is it the bi team job to go do reporting for everybody in an organization that’s what those Champions are for that’s what the business units are for are for fabric should Elevate the the levels of

26:18 fabric should Elevate the the levels of trustworthy data that that team can

26:21 trustworthy data that that team can interact with but most importantly that team’s job is to be the experts in the organization to drive those really high level objectives and help train members across the organization on how to do things I. E hey Mike you’re thou you were the power bi expert now you’re the fabric expert even though you all these other components were part of your job before now it’s just within this work stream but you’re the admin you’re the owner of all the data ecosystem you work

26:51 owner of all the data ecosystem you work with these teams collaboratively like to suggest that that team doesn’t need to be fully funded and or is just going to magically appear or have these teams cross-collaborate with each other to hit these high level objectives I think is naive I think you’re missing my point though because that’s not what I’m saying you you just you just you just put words in my mouth because I’m not what I’m not saying is what I’m not saying is that there’s this Rogue now every business what you’re thinking is well all the bi Engineers are going to

27:22 well all the bi Engineers are going to lead the central team and there’s now no direction and then the central team is sorry that’s the that whole picture I was you were repeating for me yeah but if you read the Microsoft adoption roadmap documentation they don’t recommend that at all they actually recommended a Federated approach there’s a smaller number of lead people at the very top of the center of excellence that are being funded by a an executive sponsor around bi initiatives so I would totally agree with you I would never in a way say to throw away these larger bi initiatives

27:52 throw away these larger bi initiatives because there is big company organizational goals the directions that they want to go they have to report on these larger initiatives however that Central team I think is going to focus more on governance and guidance again the guard rails of the road they were still going to have to convey those main pillars of Direction but in no air we’re trying to say we’re going to fracture the business and and make a lot more smaller business Centric units that are just going to stand alone in their own function in

28:22 stand alone in their own function in order for this to work well you need to have a multitude of of individuals right we have the everyone who touches power bi is your Community Practice there’s going to be power bi Champions that are going to be people that are funded by the business that are experts in their area in their data spaces and then there’s going to be the center of excellence that is governing bodying on the top part of this and so they’re setting standards on and again I think this is where maybe I do align with you is the role of that business unit or that Central Business unit

28:52 unit or that Central Business unit person that Central bi team where previously it was we make all the reports we make all the data we make all the cubes you just consume from us what’s happening now is the the central bi team is now focusing more on governing guiding there’s a Federated approach where if your business unit wants data from anything Central you have to have a representative or their there’s time being shared from your team to that Central team see see but this is where this is where I don’t agree

29:22 where this is where I don’t agree because now you’re talking heavy cross collaboration and shared ownership of resources that have completely different priorities and that’s regardless that’s going to happen well no it’s not it’s not there’s going to be there’s going to be either that’s why you have to have that Center of Excellence the executive sponsor because they’re the ones that are trying to have those cross they need it the central the executive bi sponsor has to go talk to the sales you

29:53 sponsor has to go talk to the sales you sponsor has to go talk to the sales executive and say hey we need to know executive and say hey we need to align we need to understand what is common what are models what data do we need to support both teams here what does the central bi team need to deliver either it’s just trade straight tables or it’s models or it’s certified things and so this this is those executive level individuals are responsible to make sure that the next level down are communicating and working and you are communicating and working and Mark it sounds like you’ve had a know Mark it sounds like you’ve had a similar experience with this one right

30:23 similar experience with this one right so it it I think it’s already trying to occur so and I I would agree that I think it’s trying to occur but I think to me the this complete straight story there’s the one big barrier here that has always occurred when we do a technology and we saw with power bi we’re going to see it a lot now it’s the technical lack of knowledge everything you’ve said to this point has required a lot of resources budget and people that are not just someone out of college

30:53 are not just someone out of college that require a pretty particular skill set and resources devoted to the organization the or and most organizations are always hesitant to spend this much on something that they don’t know is going to work or they don’t see the the final outcome and especially we’re dealing with fabric and if we’re dealing with people at this three to five year Trend that you said that I don’t think any organization is going to be willing to wait that long and I don’t think if they hire someone who’s going to start making mistakes which in power bi I think

31:24 mistakes which in power bi I think there’s a lot more wiggle room for hiring someone out of college or a little wicker room with a report if you’re dealing with data that’s now going Downstream to other systems and as soon as there’s mistakes Mike what have you always said has led people to stop working with power bi or adopting power bi stop pulling from Excel it’s still no it’s the lack of skill you’ve said that how many times when someone left an organization and so what you’re doing but yeah I do disagree with you I disagree because you’re talking about skills on people I’m ignoring that whole

31:54 skills on people I’m ignoring that whole concept I’m ignoring the fact that there’s skills on people here because the idea is everyone’s going to have to learn more of this it’s the fact the fact that tool now exists and that people can turn it on and it’s going to be in sweat organizations you’re gonna have to learn it and I think what as we look at like yeah we’re not hiring high school students so we’re not hiring fresh out of college students to come in and do this data engineering thing from the social bi team you’re gonna have to learn what you’re gonna well we’re gonna take people who are skilled in certain areas right now dbas SQL those efforts right now right I

32:25 SQL those efforts right now right I don’t think it’s a major leap to go from I used to write a bunch of stored procedures to now I’m writing notebooks completely yeah so I’m saying we’re going to take people that are already skilled and and and and educated in very data Centric things all we’re doing is Shifting some of the mindset can can I teach a SQL DBA how to use lake house and mendelian architecture of course I can is that going to be a multi-year process no it’s not no so I I think what what I’m saying is this is this is the role of like what power so

32:57 this is the role of like what power so the technology is now pushing people to be able to educate themselves more in different areas we’re now going to need less management of the infrastructure the piece like literally before fabric it was I gotta figure out how to roll things out in Azure I may need Dev test prod how’s that going to work I have this blob storage account thing but I have one of them what happens if I have data in a different region how do we see like there’s a lot more technical challenges that were there and Microsoft said showed up and said hey here’s

33:27 said showed up and said hey here’s fabric I’m going to give you this thing called one Lake and now we can say on a user by user we’re giving you a platform as a service as a as opposed to software as a service with the the other elements here and I think what’s what I’m seeing the shift in the fact that the technology has changed we have to fundamentally rethink like what we’re what we’re where are our skills and talents today what do we think the skills and talents need to be for tomorrow and then does that in any way realign what does

33:59 that in any way realign what does central bi look like versus what bi looked like in the business units now I’m not disagreeing there’s still going to be friction those are going to be directional challenges across where the organization is going there’s going to be departmental goals and objectives but what I again what general Trends I think you’re going to see that Central bi team moving more away from creation and moving more into governing and education and that raises the ship up and this is where I would say this tide raises All Ships across the organization but does it right so my

34:31 organization but does it right so my point a couple things so like one is I I do agree with you I think like with the Advent of fabric right there’s a new ecosystem that people can play in and business can leverage it in different ways so over time right and if I if I don’t think right now like I’m overtime is fabric is out here as business units start to understand it and pick it up does it mean that there’s opportunities for people we would term business people business users that use power bi to elevate right Elevate into being much more engineering specific and

35:03 being much more engineering specific and hey with copilot maybe it’s even easier than before like as that gets implemented that’s true too yeah it

35:09 implemented that’s true too yeah it does does that leave the organization open for some huge errors in cost absolutely it does right like it does because these are people who have no idea what they’re playing with over time fine that’s great right the other interesting thing as you were talking though is I I don’t think this is a tie like the right that rises All Ships like I think somebody made a comment in in chat earlier about it right does does this actually dismantle further any

35:39 this actually dismantle further any strongholds of data ownership of old school I. T business units and maybe Mike this is the point you’re making where like those teams start to shrink because like this opens the door where if I’m a business unit owner and I know or have heard that fabric gives me all these things and I can’t see sales people are pushing this oh totally what it means is I need more Technical Resources in my area that I didn’t have to pay for before exactly right so your distribute

36:09 before exactly right so your distribute you’re you’re moving you’re moving like the builders the creators out of out of a technical space and putting them in business units correct and like I I like I said I think that’s fine with the exception that there should still be an overall strategy where you’re still going to need the same level of maybe not volume of people if you’re Distributing that workload but then

36:43 there’s just so many ramifications of that decision like separating it out because that’s saying like hey Finance you now own all of your stuff and by the way not only are you running your business unit I need we need you to start building these data sets for the organization and when you’re done you need to hand it off to this team not even head off that is a high priority or they need to be able to plug into it so that so that’s that’s more yeah so that’s where I’m more thinking this is going to be right so Finance team should own all the financials of the company that’s totally true but what happens when the sales team says hey we need

37:13 when the sales team says hey we need some better financial data what does that look like we need better sales data what’s occurring here for from that perspective now this is where I think the challenges again I think there’s going to be more challenges in this approach now because some of the barriers to things I would do in power bi is I wouldn’t want to use dataflows for very large data sets I wouldn’t want to use apis calls getting data into a place was much more challenging with power bi versus using a pipeline right so there there’s there’s an unlimited amount of information that

37:44 an unlimited amount of information that you can learn on the internet and my argument here is now that the learning cost is now purely a Time function as opposed to more of a I need to pay for training there’s a there’s a an education education level that you need to get here like I don’t think you need to go out and go get a computer science degree to go become a data engineer no no but at the same time though and this is what I was saying there’s a little more leeway with making mistakes in a report unfortunately because it it was at the end of the stream with a power bi report

38:15 end of the stream with a power bi report if I’m doing things in fabric it’s probably going to and I’m assuming to your world or the world that we’re speaking at right now the fabric in engineering that we’re doing and this was actually going to one of my questions the fabric engineering that’s going to be conducted is going to live outside of just a report the end goal is going to be talking to other systems because you can live in a lake house or database sure Etc and that’s the goal so the the implications of a mistake of something breaking or being inaccurate is much

38:47 breaking or being inaccurate is much more severe than just a power bi report and the willingness to accept a mistake that I can do on the fly in power bi I think it’s going to be a lot smaller so yet part of learning is making this I agree and that’s been the great part of power bi but I think in this new world though we at least have the ability now to let people experiment and try more things and I think this is where I’m saying with with this added capability that we’re bringing to the business we need better controls or guard rails

39:18 need better controls or guard rails around what data is really certified what data has a process to follow right we are we are no longer like the fact that we’ve been in this space so far all this is done is just added more opportunities for us in this space in my opinion so who designs the certified certified data then is it the business or is it the centralized bi so this is this is a great question who’s who is the one that does the certification right if you’re talking about a certified if you’re talking about a Federated Coe meaning there are

39:48 about a Federated Coe meaning there are a couple people that are from the central bi team and there are individuals or representatives from the different business units I think the cue should be the team that is certifying data sets now the center of exon’s Team should Define what is the process and this is an agreed upon pattern that the organization leans on right for example a certified data set can’t be getting data from Excel files it’s 2 finicky it’s going to break too often that does not make sense right a certified team should say we should have some level of data Ops along these certified data sets

40:20 data Ops along these certified data sets so how do we check how do we verify what comes out of the source system actually matches what’s in the report those are the questions I think the center of excellence team needs to address and when the the financial data becomes an important topic for the business because the data that that team is working on is strategically valuable to the entire organization so I really do think like that team needs to is they’re building value they’re adding value to their information and likely the knowledge that they’re

40:51 likely the knowledge that they’re building into that data needs to be shared across the organization the center of excellence or that Central team is the team that’s working with that Finance team and saying okay let’s let’s review how do we certify things what are the guardrails we’re going to put in place here how do we is the data model correctly does the performance of the model work well and as we grow up that model right the central team is is working with that team to then figure out how do we certify that data set and distribute it out and then there’s again a lot of this is again governance piece

41:21 a lot of this is again governance piece you could say I’m going to give you just reports I’m going to give you a cube a data model or data set or I’m going to give you access to Flat files or tables that exist there so to me these are these are layers of depth closer to the raw portion of the data that are now available to us that we never had before if you have the right resources if you have a CDO I think organizations are going to require or they’re going to be flying by the seat of their fabric pen

41:53 and that Fabric’s on fire but I think this is going to require CDO to understand what the investment is because everything you said would work that’s exactly if you had the investment CDO executive sponsor same thing in my mind it’s fine I I think it’s going to require the the devotion of a CDO not just a sponsor and but I think everything you said would work if you had an organization that’s willing to invest that much resource and time because again I don’t think organizations will be willing to say oh no we this is the requirement the

42:24 no we this is the requirement the requirement to most organizations going to be the investment in the cost they see on a sheet of paper that whatever I pay for fabric every month or whatever I’m paying for the engineering more than that it’s more I guess I know I know but I’m saying their understanding of what the investment’s going to be that we’re going to need which what you made sound almost is a hub and spoke approach just in a different way organizations can choose how they want to roll us out all all day right you you can do power bi multiple different ways you can roll out power bi with a central team or a Federate

42:55 with a central team or a Federate approach there’s a central team that will have Hub and spoke approaches or you can roll it out as a fully centralized team no one touches anything and everything goes through Central bi that’s fine too that’s a choice in the organization there’s also the approach of we don’t have any Central bi and the entire it’s all business-led all business just runs their own stuff so you can take multiple approaches I think there are cost and there are implications of going down each path not saying any one of them is wrong I would say in my opinion if I’m if I’m going to pay for an organization I’m going to

43:26 pay for an organization I’m going to focus on focus on trying to efficiently right so again I’m cheap so so right I’m not saying just spend money to spend money I’m very cheap and what I would what I would prefer my organization would be is I want all the business units or all the functions over my team with data to run as efficiently as possible I want to enable people as quickly as I can to build invest build their own elements that make their business unit running because the last thing we can’t

43:53 running because the last thing we can’t we cannot do anymore it doesn’t afford us just to restrict access to data that that and that’s probably the most cost probably has the most cost but probably the best quality I’m not I’m not sure I agree with their statement but I would say the amount of time we spend on people picking up data out of one Excel file and moving it to another file and doing some Transformations is probably way underestimated in your organization oh no I would agree with that but I think when we’re dealing farther Upstream now

44:25 when we’re dealing farther Upstream now the the severity of a mistake is much more it’s much more severe to make a mistake now Upstream compared to Something in Excel correct and we can’t be willing to make these oopsies oh no your quota is actually completely wrong this month sorry guy we’re learning a new technology and I’m just learning this Medallion approach sorry that your tart your quote is off this month well I don’t like I don’t think that would be I don’t think well I would hope that organizations wouldn’t go that shoot from the hip and just light up fabric and just say just go build stuff right yeah ready to be

44:57 go build stuff right yeah ready to be disappointed I think if you’re if you’re so things that are important to organizations are your Revenue right and then those those bonuses or benefits that go back to organizations and people and every organization that I’ve been in that has any benefit paid on performance scenario they’re not just whipping those out the door and then sending them away there’s there’s lots of people reviewing them there’s lots of people at the executive levelership leadership approving the process of what that is and again I think your example is a bit naive in the in the fact that your

45:29 naive in the in the fact that your example is yes you’re right people won’t just change your benefits halfway through but however because there is a certain level of data in your organization is more important than others that is a very important data set I would say and so there’s going to be much more rigor naturally put upon the organization to make sure that that data is right yes I took an extreme example to make an impactful point but I understand well I agree with yeah I I agree but I also would say there’s a lot of organizations too who still don’t have the resources that they need in power bi

46:02 and then they’re they’re shooting from the hip too unfortunately again I’m not going to disagree with you yes I think you’re right but again I’m looking at this going so as I work with organizations I look at them and going we don’t have the skills we need on this team today currently to do what we want to do with fabric we just added a whole new realm of things we need to know and learn that we already have people who can’t even open up like powerba. com and correctly navigate to reports in our organizations so what’s the expectation that’s going to happen from the organization on we have a

46:35 the organization on we have a whole lot more education that needs to be provided into the organization around all this stuff this is a whole new world that people are trying to figure out like and this is where I just push back like or where I’m pessimistic about how many people are from a business unit are just gonna like adopt this because if they’re not doing reporting right now if they already have struggles with challenges of people engaging of doing all these things right now like to find a power bi Champion is a miracle and a great thing in an organization every

47:06 great thing in an organization every business unit doesn’t have one and then to elevate them into engineering level things like dude okay yes it’s possible I guess my question to what they want to close though that I’ve been thinking about is maybe maybe there’s a different opportunity here where if there is a larger centralized bi team that we like fabric gives an opportunity for us to clearly demarcate a business unit and ownership of work of data work streams to support their own reporting as well as be a be a source for all of

47:39 as well as be a be a source for all of the Upstream like data sources that we would need maybe the opportunity is to not dismantle a central bi team but put one of the engineers in each of these workspaces so that collectively like their job is work with and support to deliver reporting like you did for a business unit but the it’s not a Coe thing it’s not where we’re all we’re all coming together and figuring out like what are the things to do Coe would

48:09 like what are the things to do Coe would be like hey now that you’re embedded in this business unit you’re 30 of your job is still to make sure that we’re pulling centralized cleaned governed data into our models for these other purposes and training right and maybe maybe you just embed those users into the business units to do what you’re talking about Mike because you you can’t do a holistic like hey blow out we’re going to Train everybody how to use fabric when no I agree like the engagement just isn’t going to be there as opposed to

48:39 as opposed to integrating or sticking your fingers into these business business units with the overall objective of like hey down the road here’s how we’re gonna get engaged or pull more people into this as opposed to push down informational messaging and I think Mike I gotta I’ll end on a question I love that for both of you is everyone ready for fabric or is there a certain part of the maturity skill or like levels where that you can tell like

49:09 like levels where that you can tell like where so everyone and I think that’s maybe the question so I’ll answer your question quickly because we got to kind question quickly because we got to wrap here so I would say first and of wrap here so I would say first and foremost everyone is not ready for fabric and I will say also a lot of traditional I. T roles are going to be challenged with fabric so a lot of people that have been a traditional business intelligence team this fabric thing is really going to push them to need to learn more and different things and so this I I clearly see the skills of data people trying to shift and for

49:40 of data people trying to shift and for and for one for for every one data engineer or data modeler we’re going to have three to five report builders for every three to five report Builders we’re going to have 10 100 I don’t know a larger number of people just consuming reports so to say everyone in the entire organization needs to understand fabric the answer is no I don’t think that’s the that’s the scenario here but I think what we’re doing is we’re trying to figure out what those centralized roles around data engineering looks like and figuring out

50:10 engineering looks like and figuring out how to best leverage our current resources upskill them so they’re able to use this new thing called Fabric and hopefully get all these teams to work together and again I think the the important thing I think the thing that I feel like is missing most here is the pattern of governance there there is that is the the missing part of this that I see the technology is changing but we the governance is not as clearly well defined from Microsoft and there’s more articles coming out from them that are helping with this but I think this

50:41 are helping with this but I think this is more of a nuanced area that it will vary between different organizations on what they feel is comfortable or not comfortable with what they do with their data data so and well and I’ll ask both of you Seth too could you identify theoretically on the maturity levels on the adoption roadmap where someone could potentially be a prime candidate better suited for family I would I would totally read through that the employment and I think the deployment roadmap will have to change in lieu of fabric now too because there’s going to be new areas

51:12 because there’s going to be new areas yeah there’s going to be new areas that are going to be exposed Now by using this tool now a lot of the fabric a lot of the roadmap addresses three things IT addresses your people technology and your process and so you’re gonna you’re gonna continue to invest in those areas and figure out what’s the best mix of those things for your organization so I think this was a really good conversation I know we’re at time now so I want to just wrap here and say thank you all very much for hanging in there in the chat while we’ve been fervently arguing around these

51:42 been fervently arguing around these roles and where they belong and where who’s gonna spend what what’s the Central and a central team going to look like or not look like in the future but regardless I think we can all agree agree everything is going to adjust there’s going to be new patterns evolving here and so we all need to stay in tune and and connected to what’s happening either from Microsoft or what the technology does and we’re going to figure it out there may be different answers to this question as we move forward which I think we definitely had different opinions about that here Tommy where else can we find the podcast

52:13 Tommy where else can we find the podcast you can find it anywhere it’s available apple and Spotify make sure if you want us if you have a topic that you want us to talk about or argue about make sure to go to pbi. tips and leave it Slash the podcast and you can leave a mailbag question also join us live Tuesdays and Thursdays usually on the same channel on the same screen so we don’t have any technical issues all right borrowing no technical issues for next time we’ll see you next week thank you all you all yes

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.