PBI Domains – Ep. 229

Fabric domains sound simple—organize content, align teams, move faster. In Ep. 229, Mike, Tommy, and Seth unpack what domains actually imply in the real world: ownership, governance, and the operating model you need when Fabric makes data engineering more accessible.

News & Announcements

• Fabric domains (Microsoft Learn) — Microsoft’s overview of what domains are, how they group Fabric items, and where they fit in the governance story.

• Data Wrangler in Fabric notebooks — A walkthrough of using Data Wrangler to shape data inside Fabric notebooks without living entirely in code.

• Explicit Measures Podcast — Browse the full backlog and share the show with your team.

• Tips+ Theme Gallery — Standardize report styling quickly with generated Power BI themes.

Main Discussion

Topic: Fabric domains, ownership, and “who owns what” in a multi-team tenant

Domains provide a way to organize people, data, and content—but they don’t solve governance by themselves. The real work is deciding who is responsible for shared assets, how trust is signaled, and how work moves from experimentation to something the business can rely on.

Key takeaways:

• Treat domains as an operating model decision (accountability + boundaries), not just a navigation feature.
• Be explicit about what stays centralized (standards, templates, certification) versus what is domain-owned (local reporting and iteration).
• Don’t hand-wave workspace roles—define admins/members/contributors so ownership is visible and support paths are clear.
• Manage shared semantic models like products: documentation, versioning, and change management for downstream consumers.
• Define a promotion workflow (dev → test → prod) and pair it with certification so people know what is trusted.
• Fabric lowers barriers to data engineering; that increases the need for training, guardrails, and repeatable patterns.
• Adoption should match maturity: not every org (or business unit) is ready for Fabric-wide enablement on day one.

Looking Forward

Pick one domain and run a small pilot: define ownership, set a promotion/certification process, and measure how it improves trust before scaling.

Episode Transcript

0:28 welcome back to the explicit Masters podcast with Tommy Seth and Mike hello everybody everybody welcome back happy 4th of July happy Tuesday with the bang bangs and the few pews up in the air all right if you weren’t if you weren’t sleeping well enough already you’re gonna have all the fireworks late at night there’s someone there’s always someone always someone who’s launching fireworks at the end of Fourth of July and it’s like oh 10 o’clock I heard a couple okay 11 o’clock there’s another couple far who’s launching fireworks at 11 at

0:59 far who’s launching fireworks at 11 at night like midnight like okay people it’s been dark for hours now we can stop already I want to go to bed breaking up my kids go back to bed I know I know like having kids completely changes every late night shooting of fireworks yes like this is the coolest thing ever yeah and when you have kids you’re like I swear to God I’m gonna walk down to the neighbor’s house exactly exactly let’s start off with a quick intro so

1:29 let’s start off with a quick intro so for those who are listening this is a pre-recorded episode we did not get up early on July 4th to run the episode but we will play it for you so just FYI this is a recorded episode yeah if you are watching live right now feel free and you win a prize you could submit a mailbag and that’s now available yeah we’re all still sleeping in because you we’re all still sleeping in because Fourth of July is for us start on know Fourth of July is for us start on July 3rd so because it was a weekend so we’ve been partying hard from the third all the way on so that’s probably we’re still all sleeping somewhere excellent

1:59 excellent somewhere hopefully not the gutter hopefully not the gutter hopefully in our own beds those levels but I hope so I like to think that you can always have a bad day I don’t know all right so moving in there’s an article here from formu which will be in the description of the video and Gilbert talks about some a feature that I’m very much a fan of if you write python if you like this graphical interface of power query

2:29 graphical interface of power query when you start working things in fabric you have this ability to create some notebooks a notebook is basically a you could think of it like a cells of commands that you can stitch together like command after command after command you can kind after command after command you can use data transformation so this is a of use data transformation so this is a very common thing when you do data science or when you work with python there’s these Jupiter notebooks that you would use would use but the really neat option here is if you’re in one of these jupyter notebooks there is this thing called Data Wrangler and data Wrangler feels a lot like power

3:00 and data Wrangler feels a lot like power query query you have a lot of actions that you can do in the left hand side like formatting adding a filter providing like aggregation on certain parts of data and then the python what you do is you basically apply these steps and the same way you build in power query you now build in Python and then it basically outputs for you the answer or the results of those data engineering steps so you don’t have to be a python wizard to know oh which function was I

3:30 wizard to know oh which function was I using I think on one hand of this I’d like I love the fact that there’s a UI that goes along with this on the other side I feel like this is the best way to learn languages have a bit of a UI an editor that does a lot of what you want and then behind it there’s code so once you hit go or inject code or load the code back to my notebook it literally takes these steps and bundles them together and then produces actual python code that does exactly what you were thinking and I think this is great I love I absolutely

4:00 think this is great I love I absolutely love this stuff this is I when I saw this one this feature this data Wrangler I was like yeah people this is an underrated feature that people are not talking enough about it is so cool and what’s great about the Article 2 is what Foreman was doing is he’s actually applying this to a fabric workspace a fabric one notebook yes correct which is really neat which obviously makes sense because that’s what I’m really been impressed with Fabric and at least the the notebooks is it’s not limit it’s not like a limited type of

4:31 limit it’s not like a limited type of notebook it’s full featured Jupiter notebooks yes and I don’t think it’s really lacking any of the features or the capabilities so any other tool that you want to use it in you can use it and with with data Wrangler what a great way if you’re learning fabric for the first time yes and if you’re also like hey I’m comfortable writing M or I’m in the power query editor all I need to start making notebooks well what does that look like there’s a there’s a one button press on a data table coming from the data lake so right click this file or

5:02 data lake so right click this file or table say load to notebook boom a bit of code is written now you have a data frame from your one like awesome then you jump into Data Wrangler make some more code there boom couple transforms a couple simple transforms done notebooks complete and I’ve heard actually some people some feedback from some individuals saying that the jupyter notebooks have been running very reliably and pretty fast as opposed to District doing dataflow’s Gen 2 which has some bugs it sounds like it’s like they’re still working out some Kinks

5:32 they’re still working out some Kinks there’s been some inconsistencies inside the dataflows Gen 2. anyways great recommendation Gilbert any other thoughts too no it’s a great walk too that’s a good walkthrough so I highly recommend check out that blog from formu Gilbert does an excellent job he’s Gilbert I would say is one of the the it’s like a sleeping Dragon I guess what would you say here’s a hidden giant What’s the phrase I don’t know what’s he he has really good stuff and like you

6:02 he he has really good stuff and like you think oh and every so often I start googling for something like oh look there’s Gilbert oh yeah wow okay this article is really good like he does a great job it’s like oh I’m at a Hidden Gem I think that’s what I’m going for it’s like a hit or go that’s what I’m looking for that sounds like a bad July 5th to me yeah exactly exactly what that sounds like sparkly sparkly Dragon fireworks

6:36 like sparkly sparkly Dragon fireworks what what was I drinking yesterday oh sparkling Dragon again box 35 volume by alcohol broadcast now sponsored by the sparkling Dragon sparkling Dragon what they need to create or what Marco Russo forget Dax Optimizer what he’s gonna do notebooks for Dax would be incredible because you think about if you’re trying to see like the evaluation context for each of the sub notes that you’re doing with the

7:07 notes that you’re doing with the calculated like a nested calculation imagine having that as a notebook so the first thing you want to do is a filter table what comes out of that and then you could write your next part of the dash query and then at the very end it’s just one function so if that’s possible Marco if you’re listening which you’re not Alex foreign [Applause] well I would also argue too like they just came out with a brand new tool that’s going to help you do some Dax Optimizer so I think that’s a they’re I

7:39 Optimizer so I think that’s a they’re It’s it’s not far off at this point mean it’s it’s not far off at this point it’s not quite a notebook but it’s it’s doing things to help you make and write better Dax as you go and I think I think it’s possible because in tablet or three with the debugger where it it goes through the steps and it’s showing you the query so you could technically technically take your nested filter or your nested calculation forcing people to write every component of a tax calculator if they want to but

8:09 of a tax calculator if they want to but if they want to see the output why would you want to when there’s shortcuts and oh did you say oh my God no were you the neighbor that was going to come over right now in your kids yes you want to I know I there are five really solid reasons I was going to say 100 I was gonna say five people in the world who would really want that feature yeah Tommy being one of them Tom being one of them yeah all right I’m moving on

8:41 yeah step away slowly look away slowly yeah yeah all right let’s jump into our main topic

8:44 all right let’s jump into our main topic for today so our main topic for today is talking about roles and domains as it relates to workspaces now we we had a previous episode called Rip workspaces so in our previous episode we’re talking about about this this concept of like now that we have fabric now that we have this data engineering exercise creating tables making more mature data for our our company inside these workspaces we technically have a proliferation of of lots more workspaces

9:16 of lots more workspaces maybe right depending on how we design things like based on what access we want to provide to generate engineered data inside our one Lake so that being said when we look at this ecosystem of okay well how how do we now rationalize workspaces and roles and people and what should they be doing so we started we started on the topic and Seth you very wisely rained us back in so we didn’t go on a large Rabbit Trail the entire time but this is the topic to go take that Rabbit

9:46 the topic to go take that Rabbit Trail back up and we’re going to talk more about those details any other things set that I missed there that you want to do for setup well I would just recommend that if if folks are the the precursor to this conversation is our episode 225. right where where we’re talking specifically about how these elements are the same like workspaces is the same name but how do they transition and how are they different in the fabric ecosystem with domains and we set the Baseline for some of the pieces that are significantly different and and how it

10:18 significantly different and and how it like reshapes but part of that conversation instantly delved into this topic which is very relevant which is how how do the people within these new structures work because it is it going to be the same are the roles within workspace can just remain the same or do we have to rethink about how how this is organized and what are people doing within the ecosystem yes so I guess we should talk about very

10:49 so I guess we should talk about very quickly let’s let’s go through some initial thoughts here right just jump into the beginning let’s talk about maybe workspaces and what roles are or as the roles are defined today right today there’s four different roles inside of workspace we have the admin the member the contributor and the viewer and I can’t tell you how many times I’ve switched up contributor and member member every time I had to keep remembering that the member role is is higher than the contributor role same thing yeah I’m I’m

11:20 same thing yeah I’m I’m I go from viewing to I’ve I’ve a membership to I’m a contributor oh but I membership to I’m a contributor oh but you’re a contributor right like mean you’re a contributor right like that yeah that yeah I do things like it feels like if it’s a contributor here yeah I’m not just here I’m doing stuff right right I’m contributing to the workspace every single time I gotta look it up everything every time I am a member of the Sons and Daughters of Italy but I do not contribute to them

11:59 contributing yeah not only are you a member you’re contributing that’s just like Seth and I being members of the Milwaukee power bi user group and we just don’t do anything yeah wait wait wait wait wait wait wait yeah awesome so okay so thinking through those different levels right so the admin can basically do everything you can add other admins you can remove people it gives you full control to publish the app all those fun things when we talk about the member basically the member can do everything in admin except add other admins that

12:30 in admin except add other admins that seems like that’s the threshold for the member level items then when you get down to contributor you start actually limiting what contributors can and cannot do and I think at the contributor level you stop having the ability of being able to create an app but you can republish an app or is that the memo I think maybe it’s the member level the members can only republish the app or update apps the contributors can only make content inside the workspace and then the viewer role basically you can’t edit anything you can’t add stuff you can’t make

13:00 you can’t add stuff you can’t make things in the workspace you you are a viewer you can look at the content that is there but you can’t do anything my opinion on base of those on those four roles the ones that should be used the most are the contributor level most everyone should live there and the member and administrator roles admin roles should be reserved for the central it team if you have like a a central team that needs to govern everything in power bi the commember roles should be the data

13:31 the commember roles should be the data stewards or the owners of the workspace like for example some organizations have the pattern of you can’t make it you can’t make your own workspace you have to request it and then they’ll be given to you well the requester becomes the member right they have control to do what they want to do in that workspace but to me that that aligns with that role what do you guys think is that your understanding of those different roles is there something I missed there no that that’s pretty spot on because honestly the viewer role is the the rarest one probably I don’t I don’t

14:02 the rarest one probably I don’t I don’t know why would you ever use that I would use an app a c-level ask for it and go okay thing so but that’s still what but why would I confuse people like a sea level is looking for a single report where I can just show up and just look at stuff there are some people who are more tech savvy and they just want to see what’s going on and you just there’s really no pushback there can’t you use you use like back in the day Tommy didn’t you report to a finance guy at one point yeah right so this would be a perfect opportunity where like somebody who has no idea what you do

14:33 somebody who has no idea what you do it’s like you need access to the workspace yeah yeah and then at that point if you’re a viewer I’ve seen that I’ve honestly seen them once with someone higher up who just wants to see the content but that’s the only reason for the viewer there’s no other reason to have a viewer because again if you’re doing things right you’re sharing things through a power bi app okay so that this becomes so when I think about the governance of those artifacts in the workspace this is where this becomes incredibly important

15:03 this becomes incredibly important because because when you talk about an app you have this and again I think apps were before you had the idea of a single app and multiple audiences right so that now I can have groups of people that can belong to the same app and then those groups of people can see different objects that are published with the app before that it might have made sense to actually have a viewer role because I only had you a viewer role because I only had I had maybe a lot of reports or a know I had maybe a lot of reports or a couple data sets in the same workspace and I only wanted to publish a couple of

15:34 and I only wanted to publish a couple of them via the app because I needed to have another app from another workspace that I’d have to man then it’s convoluted yeah it was so convoluted right and then the other two rules I think you’re on where in an Ideal World or in a perfect hierarchy you have report authors or even the self-service who can publish to the workspace that the gone through the the Coe they’ve gone through they at least know what they’re doing so to speak to create a report yes and then

16:04 to speak to create a report yes and then basically getting some basic training exactly to get to that level yes where you have workspaces that are you commissioned out there but you always still have that from a governance point of view and from a culture point of view have a workspace admin who maybe is not necessarily the power bi admin someone who they’re responsible for the content in the workspace yeah and I think this is one of the things that’s very important to note especially when you’re talking about Enterprise bi right versus reporting and intelligence that

16:34 reporting and intelligence that comes from the business unit because if you think about all the different artifacts that we have in this thing I think this is where I start thinking about where does domains and where does fabric include in this space if we have we have a report that people can look at that’s supporting that report is a data set so you can have access to a report via an app but you may not have access to share reshare or build build permissions basically on the data set what that means is you can just get the report and you can just leverage or use

17:05 report and you can just leverage or use the report as is but you can’t build your own reports off that data set I think that’s very relevant for things that are coming to that Central bi team you don’t you maybe don’t want a lot of people initially using your data sets at some point in time I think there are certain teams or other parts of the organization that you can throw them into okay we’ve done the training people understand what’s going on they’re now

17:28 understand what’s going on they’re now becoming power bi Champions they’re more comfortable okay now we’re going to give them more access into the data set okay so then let’s keep walking further back Upstream here right we make tables somewhere else right and that’s where I think fabric starts picking up the conversation here around fabric is now creating a lake house with tables in it and potentially it did a warehouse and there’s SQL based things so this is one layer further back closer to the source of the data

18:00 to the source of the data and now we have to think about okay how do we control those artifacts in relation to everyone else so anyways that’s that’s where like we now have these multiple layers of things and so we’re using workspaces as a control mechanism to give people meteored access to various parts of our assets assets and this is where I would look at today we look at the guidance from Microsoft Microsoft recommends putting your data sets in a pipeline a deployment Pipeline and your reports in a separate Pipeline and then you can control the access to

18:30 and then you can control the access to reports and or data sets per audience or per group or whatever you want you get more control there and now we add fabric does that mean we need a third row of data sets we need a third row of engineering data that’s below this like a a fabric deployment pipeline you can’t deploy a lot of the artifacts in a pipeline can’t be deployed using deployment pipelines yet so what does that mean or they don’t work with Source control either that’s also true if you turn on GitHub or git it’s not GitHub

19:00 turn on GitHub or git it’s not GitHub get get integration which because you can only use Azure devops today but once you turn on git integration you can’t sync a warehouse you can’t sync the data flows more like sync a data set a report right now yeah it’s just data sets report which is which again I will say that is a big win I’m very happy that is working but I would hope in the future that all of the things would be able to be synced eventually so with domains it’s we’re considering a level up and for the remainder of the

19:31 level up and for the remainder of the conversation I’m open to being wrong about this and this is this is domains now talking about in the context of fabric because that came out at the same yes now our domains directly related to fabric or domains just released at the same time as fabric it doesn’t I don’t see any language in the documentation that says let’s just say what is fabric domain so maybe I’m wrong you need a fabric admin in order to assign a domain and create a domain domain so domains are a fabric enabled feature that’s what it looks like because in if

20:01 that’s what it looks like because in if you can’t a fabric admin or higher fabric admins can create an edit domains so if you don’t have fabric unless but didn’t that just replace the power bi admin like I thought a power bi admin now just turned into a fabric admin all of a sudden fabric you’re still a fabric admin yeah I think they’re changing the name of a power bi to fabric this is what’s confusing to me because it’s technically power bi like yeah it it’s it’s so interwoven now it’s like power bi fabric it’s like now one bigger

20:32 power bi fabric it’s like now one bigger word right as opposed to like fabric bi it’s Fabric or it’s it’s not just power bi anymore it’s it’s a bigger thing and you’ve that’s why we added a little logo on this on the podcast that’s because yeah because now we can now we don’t know what we are we don’t have an identity crisis we’re having an identity crisis and now everything’s green as opposed to Yellow before so now that’s another thing so anyways but the fabric admin can create an assigned or edit domains so that’s the first step and that’s in the admin portal

21:03 admin portal the domain admin can they’re consider the business owners or the designated experts and they say they should be familiar with the data in the areas of restrictions that are relevant they can assign workspaces however they also have to be the workspace admin to do so which I did not know know so a domain admin cannot assign a workspace if they are not an admin of that workspace already they can’t assign a workspace to their

21:34 they can’t assign a workspace to their domain if they are not already an admin of the workspace yes yeah but if you’re if you’re a domain admin you can just add yourself to the workspace but so I think that really talks about who’s the Persona of the domain admin then because it says they’re the the

22:05 then because it says they’re the the domain of the business owners or designated experts so that doesn’t sound doesn’t sound like I T to me it may be it may be and I I think that’s a big maybe talking point here in my head I initially thought of a domain admin as almost like a closer to a stakeholder closer to the business than it but I I think I’m I’m I may have to Pivot that because that role and especially the content in

22:36 that role and especially the content in a domain is not necessarily the data or the reports it’s really the more of the artifacts this the data sets it’s the lake house Warehouse I I agree and I I think if you think about the workspaces that so a domain is a collection of workspaces right is that is it yes yeah okay so in the workspaces we now have a collect a collection of reports data sets artifacts so call what you will and then

23:08 artifacts so call what you will and then an app likely right when you read the documentation I like their description of the domain is like a metadata tag so if something is attached to a domain it’s now quote unquote tagged with that type of information and the example they give is like Department based right so Finance HR sales right because those teams could be very large and there could be a lot of workspaces that various teams or groups of teams are working inside those domains which I do

23:40 working inside those domains which I do think is very relevant and to your point Tommy a lot earlier was a domain is related is one of the reasons of the domain is the need to have one Lake data grouped together right I can go to the sales domain and see all the artifacts or all the data that I have access to inside the context of sales now I think the wrong metaphor here or I guess the wrong infrastructure is thinking this as a folder and subfolder because as of right now it’s really not

24:10 because as of right now it’s really not that idea where I’m going to go to domain now I’m going to see all the workspace artifacts right now the only place the only ability to show data or artifacts in a domain is the one Lake data Hub the domain has nothing to do with apps and has nothing to do with reports well as right now you’re so you’re just saying the tagging is only really relevant for the data going to one Lake exactly

24:40 going to one Lake exactly why wouldn’t you wanna I would think you’d want to see if I’m looking at all like so hold on if you have on my mind that was just my mind crashing right now yeah I belong to multiple workspaces that cross domains domains you’re you’re saying we don’t have a tag or an understanding of which workspace that which domain faces belong the only place you can actually in a sense see a domain in action outside of like assigning a domain is the one like data

25:12 assigning a domain is the one like data Hub right now and that will only show you data sets it will show you it will show you like the data flows or a warehouse but it doesn’t apply from this doesn’t apply for anything from a consumer point of view or like here here’s all the apps now you have access to if you’re part of this domain that’s not part of the part of the gig so to speak speak that’s interesting yeah you’re right so I’m looking I’m just hitting my head around this the tagging at this point only lives at

25:43 the tagging at this point only lives at the data set low and again this makes also sense I think too when you start having like okay we have a lake we now have a warehouse we now have a SQL endpoint there’s all these things now that come out of the the build inside the workspaces okay that makes sense but I could also see a need here to say what are all the reports I have access to that are in like there’s probably another Hub showing up at some point where we we would say here’s the report Hub by domain here’s

26:12 here’s the report Hub by domain here’s all the reports I have access to irrespective of or here’s all the apps I have access to based on domains based on what’s been built from my organization that would also make sense to me yeah and you can see the data set so again you saw that ability if you choose a data set you can see the reports in there so if I I’m looking at a data set right now and I I should be able to see I can see the report and it shows the downtrend Upstream but I don’t I still don’t know the one Lake data Hub because

26:43 don’t know the one Lake data Hub because that’s right now again the only place that in in a in a workflow or really just using domains actually applies there’s really no other place that you can a domain actually has an effect on something something in the power bi service this is simply in the one like data hub for now foreign

27:17 and what do you mean by that you feel like it should be more inclusive than just the data sets well it should be more than just the one like data how about I think there’s a there’s a hope here or the idea because we’ve talked about this I don’t know how many times we’ve talked about having folders or having some better hierarchy with all we need it now more than ever yeah and you go to go to the data Hub now I have in a tenant that is just a bunch of I wouldn’t say trash but like the power bi tips tenant we just have a lot of junk hanging around and so I go to the data Hub and there’s like pages and pages and pages of like sample data sets

27:47 pages and pages of like sample data sets sample data set sample data set like they’re all over the place think about your clients or the people you work with now how does anyone use the data set Hub or do you still navigate to the workspace and that’s that’s been my workflow like sometimes I’ll go to the one Lake data Hub but that’s not in a sense part of the process right now and I don’t know if that’s going to be part of the process I don’t know so like what I I don’t know so like what because if this is well you are you mean because if this is well you are you are also a very heavy power bi user as well so I think I think things fairly

28:17 well so I think I think things fairly much like if it changes when I’m if I’m purely a report consumer and I just get links to apps that I’m consuming the content from I think the story is slightly different and again I think if you’re if you’re talking about you’re simplifying some of that experience down for report consumers they may want to start moving into a space where they could actually start consuming content from existing data sets so they wanted they instead of just

28:47 sets so they wanted they instead of just being a report consumer I want to start building my own reports all right well how do you start that progress well then maybe you start more oh here’s the one like data Hub here’s the data sets that I have access to I can start building from them and maybe that’s over that self-service bi training that has to go along with that I agree with Tommy you along with that I agree with Tommy especially if we’re gonna know especially if we’re gonna especially if workspaces Traverse the ecosystem and workspaces are the permission levels and sets and groups like within a workspace I have data sets

29:17 like within a workspace I have data sets I have data objects I have reports I have whatever like just give me the top the same way you gave me like this really linear breakdown view right of how everything interconnects I have the full list give folders yeah main workspace data sets reports whatever the case may be it’d be nice folders would be very helpful here and I think that too okay oh no like we’re talking ironically enough in this one’s structures instead of roles I

29:48 one’s structures instead of roles I guess to some degree but I I guess the the question I have is when when I look at workspaces as they existed in power bi these yes to some degree these were elements of consumption as we deployed a new data set and report but it was a very singular to a data set and report I could always spool up to a certain size with a one data set and as long as I stayed within those parameters all of my P skew like if I

30:18 parameters all of my P skew like if I had an Enterprise goo all of it just continues to work I’m not incurring more cost cost where I’m where this is interesting for me is is like now we’re opening up these same levels of permission specifically around members and contributors and obviously admins that they can now have access to all of these other areas that are going to directly influence how much capacity I’m using in a different

30:48 much capacity I’m using in a different different way because if I max out capacities in a workspace now I have to increase because I’m doing ETL loads and processing data and querying data and I’m just like to me there’s a shift here of here of roles and responsibilities for these people that they now have across this ecosystem ecosystem and I don’t see like in particular areas I see like in the SQL Warehouse oh it’s

31:18 I see like in the SQL Warehouse oh it’s like oh you can lock down a user to viewer or a contributor or something like that so there’s like an add-on perm after the fact but if you grant somebody permissions they can go start consuming resources in a different way and I don’t to me that’s the fundamental shift here and I don’t know if it requires different levels of permissions but maybe it like does it offer should it now offer different like elements of control within each of those permission groups right and think

31:48 those permission groups right and think about right now the domain contributor which I could see a role for they have one ability they have literally one ability if you give someone domain contributor they can from the workspace area assign a domain to them that they have access to from a workspace that’s the only thing they can do there’s no other in a sense feature or permissions that they have which okay cool but that’s not I I agree if there’s not really a role

32:20 I I agree if there’s not really a role for that right now or I don’t really see that being part of a process just with that being the only ability for a domain contributor to do also I’m always concerned when with documentation if you dedicate an entire section to how to add an image to your domain domain which is one of the sections here I think we’re still in a pretty preview mode of what’s available well okay two thoughts around this one one the image thing do you shut you ask me to provide you an image but you don’t tell me the

32:50 image but you don’t tell me the dimensions of the image You’re Expecting come on this is like so rudimentary like any any platform so okay come on Microsoft this is preview I get it but like someone do their homework here like what size image do I need to provide should it be an SVG do you accept jpeg like nothing in there in the description this is more similar to me trying to do my two-page image on the image just pick an image like it’s the Pepe meme we’re gonna just throw it on there and it’s the boom done so fix that that’s one thing to fix on the other hand though what when you start

33:21 hand though what when you start assigning workspaces right so as you start thinking about the domain owners here’s something that’s interesting that I like what’s going on here when you go into the experience around workspaces in this domain you can assign them and there’s a little UI flow screen that says assign by workspace name so I can literally pick the names of the workspaces and group them or I can assign my workspace admin and I think this was talking to what we were saying earlier right if there’s a name of someone who’s an admin in all

33:52 name of someone who’s an admin in all these workspaces well Tommy’s the head of sales every workspace he’s made is part of his domain yeah so Microsoft themselves are pushing more towards this concept of there should be a hand like a handful of admins and the admins really are controlling like a group of workspaces and they’re trying to govern them and this is what I would see very well clearly called out in the adoption roadmap as well this concept of really data stewardship data ownership

34:24 really data stewardship data ownership it starts with the planning of workspace admins and workspace members and so I really I like that feature I really like the fact that they’re actually leveraging the names of the admins on the workspaces to start grouping them together because that’s that is a maybe not a backhanded way but it’s like a a discovery way of figuring out who owns that content I really like that that’s actually and I think that’s the story here that yes we’re each trying to aim or own your content yes who’s really owning it right

34:54 content yes who’s really owning it right yeah and I think this is also to it by

34:57 yeah and I think this is also to it by seeing that here it also it reinforces my belief that there should only be one or two admins on workspaces like it should we should really be focusing on when something goes wrong if a data set’s not refreshing if there’s a problem with the workspace we really need to focus on a few people at the top at that admin level and if you have like if everyone in the workspace is always the admin the admin what good is that doing you you’re you’re basically just giving everyone all the access and I don’t think that’s a very

35:27 think that’s a very robust Enterprise strategy that you should be applying to that and I think this goes bigger into the roles too where and again in in a perfect governance world I see that workspace admin or at least some role there where they when we say they’re responsible for the content they’re responsible for the data I I’ve seen this as from a centralized bi when like it’s all blank it’s like well no one’s inputting information or yes and there’s that part that governance too where your sales well you have to make sure that you have

35:58 well you have to make sure that you have the stewards in place and that’s part of that governance side of things but and that’s where domains make they do make a lot of sense because you may have you lot of sense because you may have sales operations the know sales operations the sales department this customer B2B b2c whatever the your organization is but having that in terms of who owns all that and not not just who owns what reports and if the reports have nice features in it yep but okay who’s controlling the flow of that information yes exactly right yeah I keep waffling back and forth

36:28 yeah I keep waffling back and forth right like even as I think about waffles waffles waffles with syrup right and butter butter strawberries oh strawberries and whipped cream you are a dangerous man like you think about contributor right yeah contributor roll across like permission across all the the interface and the the first thing I think is like okay cost and then you then I think about okay skill sets of these people and then should you have a report

36:58 and then should you have a report contributor and an ETL contributor and an ml contributor and then I and then I listen to you guys as you’re talking through who’s who’s the owner right of workspaces of different elements yeah and I think what’s being challenged here is is like are we going to find ourselves in a position where we do need finer grain control or is this an exercise in like listen we need to stop with like segmenting people and allowing them to

37:28 to solve problems across the Spectrum if if they’re going to engage and obviously they have the responsibility of capacity usage obviously they should understand what they’re doing before they do something but that’s part of the organization’s responsibility to teach these people what to do within that ecosystem that’s the center of excellence the center of excellence is expected to teach people these things so right so but they need to find that policy yeah but taking that that same challenge I have in thinking about what what has been very segmented roles

38:01 what what has been very segmented roles very segmented permissions to different areas of of the data organization based on those roles fabric it challenges that because now it’s an ecosystem that they can easily Traverse the other the other point to dovetail into what you guys were talking is as as you were discussing ownership of workspaces ownership of domains like the first notes I started to like typing were man all this is going to allow is this going to pigeonhole

38:32 is this going to pigeonhole organizations because you’re gonna just dig deeper trenches in each of your business unit areas to develop your own ETL you develop your own pipelines develop your own reporting and what does that do for the Enterprise bi and then I was like well actually it simplifies it because we would just take those curated final tables of these different areas of the organizations and plug into them to combine data sets across the

39:02 to combine data sets across the organization in our pipelines right because our goal would be combining data sets from different areas of the organization and fabric actually I think enables that even more because those business units and those data professionals are the owners of the source information yes and we would be just sourcing from the particular areas to combine and and continue to extend or combine across the data sets

39:34 extend or combine across the data sets because like ultimately Enterprise bi is about taking disparate data sources regardless of where they are and a lot of times it’s in third-party systems right so in the in this case it’s like hey if Finance goes through the exercise of plugging into those third-party systems and pulling it into the fabric ecosystem so that they can report on it this is a really interesting challenge like collectively bringing the data together together but but ownership levels remain where they need

40:04 ownership levels remain where they need without removing access to the data itself which just Fosters this growth pattern that we’ve talked about of adoption I think it speeds it up to some degree well I I think I I think I see where you’re going here but I think maybe there’s a role missing with the from the domain from the domain side I think there’s a domain member missing because I think if if I see where you’re going with this let’s say I’m claiming creating a lake house and I have access I would have access to

40:36 and I have access I would have access to all the workspaces in a domain someone assigns me that this is not what it is now but see where I’m going and I want to shortcut from another workspace like from another content in another workspace maybe I didn’t know I necessarily had access to where maybe that’s where I can keep link together all the data that’s relevant to me because again right now the only roles are in a sense from an assigning point of view not not necessarily from the access point of view

41:07 the access point of view he’s signing an access so let me let me ask this question because you guys are closer to this if I is there a concept of an Uber admin like there’s an admin that sees across all the workspaces or could be assigned across all the workspaces and have access to all the data a fabric admin in the service principle that you assigned to that so I I think in most cases yes there’s a uber admin the only time I would say I would say you need to think or rethink that Uber

41:37 you need to think or rethink that Uber admin type Persona is when you’re talking about regulated data spaces Financial or I I hear you or like so there’s there’s like legal implications to some things about data that like there there needs to be like a there is a yes there’s an Uber admin about like everything else but like there’s like a handful of workspaces that are tightly controlled by process and we need to know exactly who access things and stuff like that so that’s the only exception I can see that would be in my

42:07 exception I can see that would be in my mind that would say yeah just there is an Uber and admin for everything look aside from regulation that’s the only piece yeah but it’s interesting because at the same time for as much as you would lock down those workspaces you would you would either need to plug into that workspace or plug into those data sources to use data for other purposes or reverse it’s such a lockdown workspace that it would be it would have to have permissions into other areas to pull

42:37 permissions into other areas to pull data into it I see so because you can’t you can’t Finance is not this Silo that can’t like doesn’t have to interact with other company data you have to you have to combine it and I get the I get like HR and finance are two areas where you have to be extremely diligent about who has access to some data but to to believe that those areas don’t need access to other information from other groups is not accurate so what

43:07 other groups is not accurate so what you’re saying is the ability to in a sense go Downstream where fat where I’m not saying this exists today but you’re saying having that ability to pull data from other workspaces but they can’t necessarily pull from you I I it’s it’s this theory of like how do you how does this ecosystem allow us to access the data we need to in whatever whatever use case because there are many right we’re diving into like more harder to Define ones but does it help with that or does it hurt because if all

43:38 that or does it hurt because if all we’re doing is digging deep deep

43:40 we’re doing is digging deep deep trenches this isn’t doing anything except just unifying an experience where people can spend more money yeah okay do their ETL I want to jump on your point there because I think this is a great point what I what I foresee happening here and as I was thinking about these different areas by bringing the synapse World by bringing pipelines data Factory like all these other features into fabric that can be now handled inside the workspace looking at this going okay the business unit can now do more of

44:12 the business unit can now do more of their own data engineering so when we look at like the spectrum of data in a company there there is some things that are coming from Enterprise systems and maybe the role of that Central bi team is now becoming less around building all of the cubes and the reports for the business unit and now what’s happening is we’re seeing a shift in finding capable Engineers or even looking the business and my my note I wrote down here was the business might actually start hiring their own

44:42 might actually start hiring their own data Engineers to build what they need to build in the data Engineering Process but when I think about that that note I think think looking at the cost of data the majority of the cost of data is the people the labor to get the pipelines built to maintain the the effort or the work you maintain the the effort or the work five thousand dollars a month for a know five thousand dollars a month for a P1 SKU to do a lot of things with data you can do a lot with that P1 SKU that’s that’s relatively cheap compared to the cost of an employee to design all

45:13 to the cost of an employee to design all the pipelines to make all that work so I feel like what we’re seeing here is we’re now Shifting the labor costs out of I. T and bringing it into the business units a bit more but like you you are and you aren’t right well could it help out businesses but it’s also it also is your view on what is enterprise bi because like I don’t say you do that say you throw these additional users into the particular business units great then those business units get the value in that work stream there’s still

45:44 value in that work stream there’s still two fundamental challenges that Enterprise bi like solves one is combining data sets from disparate parts of the system I don’t want disparate parts of the system as long as we can like and this is where does the fabric ecosystem bring it closer for me to access those things because one of the the things we’re very good at is combining data sets across the organization to produce the really deep insights that you would need across all those areas primarily for extended leadership for for the the

46:15 for extended leadership for for the the leadership dashboards for things that are driving company decisions and does that always involve combining data sets no which is why I segment those out like no agree elt reporting and complex the combination of data I’m gonna I’m gonna disagree with you here slightly I think fabric I so I agree I agree with the fact that there’s Central bi team understands those needs and and that might be a central role thing here but when I look at what’s happening now with fabric that doesn’t necessarily have to be a role that’s only serviced by the

46:46 be a role that’s only serviced by the central bi team the central bi team may be the team that says hey I’m gonna go get all the Salesforce data and my other data here’s the tables of it oh okay here you go business now you build the value from it so to me there’s like that the logic or the business logic of how we join internal data to Salesforce data that may that may live in it central bi but I see an opportunity here for the business to own that and now what happens is this is where I think it becomes really important in here is because of that business unit to be able to own that

47:16 business unit to be able to own that information there is a clear conversation that needs to happen now that says okay business unit you own it therefore you’re responsible for it so you have to give me an admin of like so you don’t just give this stuff away and say here you go have at it go have fun but you it comes with expectations like if there’s a problem business unit who do we who do we call who’s the prop who’s the solver of this issue when there’s data ownership though like my problem

47:46 data ownership though like my problem with this is this but that could be given to the business no you can’t because you’re like the the very things I Define for you as an Enterprise like problem that bi teams Enterprise bi teams today solve are they are the ones that understand all these areas and collectively create objects that are infinitely more valuable sorry just in general can be infinitely more valuable than what a particular business unit does on its own with on in its own third

48:17 does on its own with on in its own third party tool today with Excel with all these reporting what happens is as you matriculate all that up extended leadership teams have to go okay I’m talking to finance they have their own reporting I’m talking to sales they have their own reporting I’m talking to marketing they have their own reporting I’d love to see all of the how these things combine and come together to resolve these company objectives not your business goal objectives and the problem that you’re introducing here with your theory of complete ownership by these business units is those differing teams would have to come

48:47 differing teams would have to come together collectively to try to put together combined data sets and I will tell you right now it does not work it will not work if you have of completely independent teams that have to come together to combine data to have their combined data sets who’s going to manage it who owns that so so so so I another question is we got to wrap I understand we gotta rap but the question so I would also challenge your like when you say who owns that it won’t work

49:17 you say who owns that it won’t work what’s the main pushback for that is it is it the technology that’s that’s blocking you people people okay so to me that’s a data culture issue as opposed to the technology doesn’t allow for it it’s not it’s not I dude I will I will die on this hill I will I will tell you through extensive experience that unless you have a team that owns something these disparate like that’s what I’m saying lack of connected things

49:49 saying lack of connected things no no no no no no but that’s my point you’re saying I have 10 teams that own something and I’m saying bringing 10 teams together to combine the one thing without one team owning it does not work it does not work and I’m gonna add my two cents and then we’ll rap here because I’m just so I’ve got five more sends on this one but go ahead Tommy just just disagree on this one a week ago me and Seth were going out I completely agree with Seth here because not only are they putting

50:20 here because not only are they putting the investment of fabric which the cost of fabric but now according to you Mike you’re also doing an investment of each businesses and I’ll have to understand the technology it’s not yeah 100 agree yeah and that’s not but that’s it’s not their role they don’t care it’s not necessarily not yet it’s not just their role but yes you have to now bring to the table I’m telling you you have to bring the table to your CIO or your CEO saying okay you’re gonna invest all this in Fabric and we still don’t know how much we’re going to cost but also now

50:50 much we’re going to cost but also now you have to bring in all these people to understand the technology and make it work the right way seamlessly with each business units are already doing this like this this is what it’s like or not there’s there’s this Theory and there’s like what’s really happening business units are already building are already buying premium skus in their own department for just their needs and they are like business units are already Skilling up knowledging up oh you and I Mike Mike the business unit does this one person the one person who can make sense of it

51:27 for the record I sorely disagree with your statement there Seth I think this is other issues are going on here you and I have gotten into it let’s do it oh boy oh boy anyways well with this heated argument coming to an end we thank you very much for your time we appreciate everyone listening and we really appreciate your listenership if you have other disagreements with your it organization Central bi or other leaders in your organization please send them this podcast so you can also argue about this with your internal company as well well no no don’t do that that’s not a good idea

51:58 idea so our only request of this is if you don’t mind if you liked this content if you like thinking about these things we really appreciate your listenership please share with somebody else who may be thinking about this as well and also have a happy Fourth of July Tommy where else can you find the podcast you find the podcast since we’re available Apple Spotify join us live on power bi tips channels and we’ll see you whenever we see you we’ll see you again sometime in the future take care and thank you all very much

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.