Building and BI Team – Ep. 321

This episode is about the people side of analytics: building a BI team (especially in a Fabric world) and how to structure roles so the organization can deliver end-to-end—from source to report.

News & Announcements

• PowerBI.tips Podcast — Subscribe and listen to the Explicit Measures podcast episodes and related content.
• Power BI Theme Generator — Create and download Power BI report themes using the PowerBI.tips theme generator.

Main Discussion

The discussion focuses on what a healthy BI team needs, especially as Fabric brings more “enterprise” tools closer to report builders.

Key points from the conversation:

• Start with data engineering capability: getting data in, shaped, and reliable is foundational—even before you perfect report design.
• Bridge data + reporting: you can’t get to great reports without data, and you can’t build the right data without understanding reporting needs.
• Evolving roles in Fabric: why modeling, engineering, and reporting are increasingly connected in day-to-day delivery.
• Hiring/upskilling strategy: how to add the first key hire(s) and then grow the team responsibly as adoption increases.
• Operating model matters: admin portal, workspace structure, and governance tasks don’t do themselves—someone has to own them.

Looking Forward

If you’re building the team from scratch, hire (or contract) for data engineering + modeling early, then scale report development and self-service enablement once the foundation is stable.

What this means for Power BI practitioners—and what to keep an eye on next.

Episode Transcript

0:30 hello and welcome back to the explicit measures podcast with Tommy Seth and Mike you you may notice my voice is a little mourning and as well as my background is slightly different so I’m calling in live from the Microsoft build conference out here in Seattle so we have an opportunity to speak the conference I’m going to be a Microsoft expert ask the experts apparently for things around fabric so we’ll see how long that mileage runs it may be good it may be good they may stump me a lot I don’t know there’s a couple other

1:01 don’t know there’s a couple other MVPs from the data platform space here as well and if nothing else we’ll at least be pointing fingers at each other trying to get everyone to answer other people’s questions so we’ll see what happens should be fun well that should be a fun Thursday episode to hear about all the questions Mike but a happy early early Tuesday to you then oh yeah it is early for us and I’m I’m not going to lie with your background this reminds me of our early episodes because I used to do it from office too with with

1:31 to do it from office too with with trying to blow this a little this is this is a little reminiscent so here the hotel room doesn’t really come with like colored lights or backgrounds or I can’t really I was think I didn’t think about it till like midnight I was like what I probably should have thought about like some Lights I could have put around the room so it wasn’t quite just so Stark and white I don’t know anyways it is what it is walk in there say hey I’m a podcaster so what rooms do you have available exact well that would be I think honestly well so to be

2:02 I think honestly well so to be frankly honest right I I was thinking I would like to do this from the actual event however we are two hours earlier so nothing’s open like literally the building’s entirely shut down so even if I wanted to go even just sit in front of a sign out front or something that was build related to be like hey we’re here and sitting in the couch and talking to people about things you couldn’t even do it it wouldn’t even work because there’s there’s nothing open at this point so we we should run with this idea right like from now on right you need to

2:33 right like from now on right you need to bring the big mic or show show like hey you’re an influencer Mike so like as part of that being that big influencer you could you big influencer you could call these hotels and be like hey know call these hotels and be like hey what would what do you what do you guys want to work with me here do you have a do you have a room that would be fantastic in the background as I’m talking for a full hour to my

3:03 whatever the hotel name is yeah right put that in the background I think I think I think if you gave me the suite right that would be that would be great right behind like the double big couch and all this kind exactly right I like that okay for next for next conference I’ll to think about that one Expedia exactly there you go Tommy looks like you’re in Maui it looks

3:35 yeah right Seth you’re in Bora Bora oh interesting how much you guys really making that’s funny excellent anyways we’re excited for the the event it should be really interesting I did get a chance to get there early so I already have like my Microsoft build shirt that they giving you and they like they’re doing this really interesting thing now they’re actually creating the shirts on demand at conferences now so they have like a screen printer that’s like mobile setup and so you basically pick the kind setup and so you basically pick the the shirt you want they take of the shirt you want they take your shirt over to a screen printer they

4:06 your shirt over to a screen printer they screen it and then they immediately go to a a heat press I it’s I don’t really know how it works but then they steam the shirt or something they heat it and then it sets the ink apparently and then you’re here’s your shirt I was like that’s a really eco-friendly way of doing like shirt building because then you’re not making hundreds of shirts that people may or may not use and you can just take the gray shirt to the next event and just print the next logo and do the next things like that’s a really smart way of doing it the only problem is some people become regional directors hey we’re not gonna talk about

4:36 hey we’re not gonna talk about that so a slight story well Tommy’s since Tommy picked it up I have to put down the story now so we went to the Microsoft MVP Summit and they had the same experience pick out a shirt it had like Microsoft MVPs and they had this little tiny icon on the right hand shoulder for regional directors which is another level I guess it would be like I don’t if it’s level up it’s just a different person than the Microsoft MVP program and I was like I’ll see if I can get one of those shirts so I have a little Regional director shirt even though I’m not a

5:07 director shirt even though I’m not a regional director I was like oh well I guess I’ll let anyone be a regional director these days so and is this is his lawnmowing shirt so don’t wor no one knows I have it except all three of you now who are on the podcast so anyways it was super fun event really really neat event and again it was it’s a really interesting experience it was very fast did it very quickly very well done so superb experience so anyways the the venue is amazing I’ve never been to build before it’s the first time I’ve been it’s a it’s a brand new

5:37 been it’s a it’s a brand new conference cent They have a Seattle Conference Center and this is the new version of their building so it’s you’re gonna media content from this new Seattle Conference Center which is gorgeous very well done a lot of glass very modern huge ceilings and stuff so it’s definitely very impressive looking so looking forward look forward to seeing a lot of content coming out from that as well super exciting to see that as well and then I’m I’m excited for the announcements there’s going to be a lot of big announcements we’re going to be able to see main stage and watch Saia come out and you

6:07 stage and watch Saia come out and you stage and watch Saia come out and present some things and see some know present some things and see some big content coming from there so I’m excited to see what event pieces we’re going to have what are the major New pieces of software for fabric or other parts across Azure and other things this is a very developer Centric conference so it should be pretty technical in nature so see what what what announcements are being made yeah and on behalf of Seth and I I don’t know if Seth losing this but I’m sure he does we’re pumped for you for being there you’re our correspondent but to be there to be able to be able to do the helping out but also be I’m basically

6:38 out but also be I’m basically you you can do everything I’m explicit measured podcast speaking from Live Events yeah we did it Bo we did it boys we’ve made it to the I want some I want some live live shorts going Mike oh yeah this in that corresonding corresponding thing like this is a good idea that Tommy’s having in real time this is Mike from the explicit measur podcast hey with me today right like have you heard here’s Saia he’s you [Laughter] know our best favorite

7:13 fan so well I actually a great idea I’ll see if I can capture some fun of moments and throw some shorts on YouTube and some other event things from the event specifically so I’m I’m going to experiment with a bit bear with me while I try and figure out my filming skills I don’t have a film director it’s like literally Me by myself so if if the camera’s a bit shaky it just may have to be that way because my arm only goes so far out to be able to capture that view or that shot so anyways should be fun looking forward to it so Tommy let’s jump in for our topic today I think we’ve got a topic just general

7:43 we’ve got a topic just general talking about some this is I this might would probably be more on the the softer side of the skills not technical more people and process potentially here and we’re going to probably jump with a topic around building a bi team so give us some intro points here Tommy and we’ll we’ll take it and run yeah I think this comes from previous episodes too where we’ve already talked about what are the roles in fabric now what how are we shifting with roles and I think the next logical question is okay what does that actually make up the

8:13 okay what does that actually make up the team we’ve talked about the individual career we’ve talked about what an organization is looking for but then what does the team look like and how do we actually build up a bi team both from the like you said from the skill point of view but from the different perspective point of view what makes a bi team successful what helps it grow what helps it have an impact on an organization or a department yeah this is going to be interesting because I think this is going to mean a lot of different things for a lot of different companies some ever evolving yeah it’s and I’ve been

8:46 ever evolving yeah it’s and I’ve been having lots of discussions around what does it look like with fabric now so we’ve had powerbi for this whole time and we had the fabric conference which was very focused on fabric things but yet powerbi is still a part of fabric so it really does feel like not that power barar is being pushed by the wayside but a lot of things in fabric are starting to impact and influence like how we think about doing business intelligence and how does that impact going to work on the central bi team because

9:16 on the central bi team because traditionally a lot of these roles data science and data engineering that would be handled more on the on the the it side and so now we’re bringing more of those roles closer to where we actually have reports and models so we kind to talk through that as well can I go ahead go ahead I was I want to really touch on the point that Mike just said because I think that’s really permanent to the fabric and powerbi just because we have fabric now doesn’t mean we have less powerbi or less need for

9:43 we have less powerbi or less need for powerbi if fabric wasn’t a thing right now we still would have this conversation on what a bi team is we would still have the whole intricacies of MSS and adoption and all the things that go through it just because we have fabric doesn’t mean that we’re pushing less powerbi if anything it’s the way around yeah I think and and that’s kind around yeah I think and and that’s where I I was going to lean into of where I I was going to lean into right like I think the powerbi story of individuals that know how to connect to sources of data transform it

10:13 connect to sources of data transform it and create something out of it in a model and visualization is the same things that that you can do on an Enterprise level fabric just brings those Enterprise tools closer to those people so we’ve talked about that a little bit in that roles and shifts fabric where like our our episode think 318 or something where I I kind 318 or something where I I leaning towards a slightly different of leaning towards a slightly different direction where the the the Fantastic thing about fabric is all of these tools

10:44 thing about fabric is all of these tools and Technologies are much closer together now but in terms of the powerbi super user right and somebody learning fabric it’s it’s the concepts are the same the tools are different and how you go about structuring things has a little bit more rigor behind it right as as well as technical knowhow you got to you you learn some languages you got to understand how the flow of these things work and operate through the different

11:15 work and operate through the different areas but that different area is now just fabric instead of I’ve got to set up a separate service and then add add credentials and get access methods and then I’m going to use data Factory and I’ve got to set up all these L Services I’ve got to do these now do you still have to do a lot of that yeah but it’s all in the same interface so it’s it’s less I think jumpy right I’m I’m going through an engagement right now we’re walking through in detail how to build data from Source system and move it all the way

11:46 Source system and move it all the way through to reports and and so we’re working with a team that understands like they’ve been through like Fabric in a day the high level and they’re like okay this is good but like what do I need to do where where are the gachas in this system because if you haven’t to your point St like if you haven’t gone through the build the ADF build the patterns outside of powerbi you don’t know to bring those patterns into powerbi and so there for example you have a a data source of Salesforce well how do you

12:18 source of Salesforce well how do you organize that content against your internal data sources do you have a separate lake house do you put it all in one lake house where does the naming live does the naming live at tables does it live on the lak houses so there’s all these I think considerations and I think the the community is starting to get their head around okay what does this infrastructure look like and it’s going to probably adjust and adapt as the tool change like metric sets could present another workspace we need to hold on to so we may have so we have like two

12:49 so we may have so we have like two workspaces now semantic models and reports we may want to add a third workspace for all of your data sets pipelines lake houses things like that you may want to have a fourth for metric sets maybe that’s something you want to share differently so it now starts talking about okay what what do we want to do it but the the possibilities are endless you can do whatever you want which is a good Reas a good thing and it’s also a double-edged sword it’s go no so random random

13:21 sword it’s go no so random random question you’re you’re going to build you’re going to build a fabric team a bi team what what do you need first o it’s a great question I’m going to say you start with at least some data engineering people because without that you don’t get this it’s it’s like building a bridge from both sides you have to start with some data to get yourself going but you also need the business person to talk to the business build models that other side as well so if you could get the one person

13:51 if you could get the one person that could do the data engineering and could bring in the other the other Island in here that would be great but if not there’s some so ideally you’d have one person that could do both okay fromp right now right now one I know all right if it’s one person I’m probably starting with the data engineering side of things because I think you have to get some of the data into the fabric ecosystem first and then we can like so if how to data engineer you may not be building

14:22 to data engineer you may not be building the best acts or the most pretty reports but at least you can get the data in and you can structure it in a way that gets out to the business at least initially I think over time you hire more people that are more focused on models and and front end side of things so I think yeah I think I’m leaning a little on the other end where I’m having these what can you get away with scenarios with right right nowe and and but the reason I say that is I think we’ve we’ve said this a lot like yes there’s a lot of principles of data engineering data science data

14:53 data engineering data science data Factory and power and data modeling that all goes into fabric but we’ve also established too this is a very everchanging spear where it’s not a copy and paste necessarily with the scenarios and the methods some of them will apply but they’re not going to apply you but they’re not going to apply copy and paste completely so for me know copy and paste completely so for me I think there’s still a lot of applications a lot of the Tooling in fabric where I can take a powerbi developer or someone who’s been heavy already in data modeling understands those principles and capture them with

15:25 those principles and capture them with hey there’s data flows Gen 2 there’s data pipelines which is is not necessarily the biggest jump compared to like learn Jupiter notebooks now if you’ve never learned python so there’s two out of the three ways to do data engineering we can start with now do they need to have full principles of The Medallion approach again this goes into them what can you get away with scenarios if we need a lake house or we need a staging lake house and just where our production data is that’s actually a good way to start again I don’t in my

15:56 good way to start again I don’t in my head right now with the way Fabric’s going going it’s not necessarily going to be everything we’ve done in the previous Azure services that we’re doing exactly in fabric so let’s at least get us started where I don’t have to necessarily hire the data engineer no because it’s not an exact copying pace of the same skill sets there’s a lot of skills that overlap but it may not be the same exact person because I think things are much more integrated so if I want to get quicker adoption if I want

16:27 want to get quicker adoption if I want to get quicker into f fabric I’m going to look at my data engineer or my data modelers who are more on the developer side they’ve understand the purpose and not just the skill of data flows the purpose of a gold semantic model and I think we can now bring them a step back to understanding what a lake house is and the purposes of of that we don’t have to do the full implementation so you said a lot of things in there I I had to bite my tongue on I saw I notic I notic fabric

16:58 tongue on I saw I notic I notic fabric fabric isn’t this brand new ecosystem of things that don’t exist it it is the tool sets brought together does that mean that they’re going they’re they’re not going to be enhanced or offer up new features or offer Simplicity no of course not I would imagine that’s the direction in many cases but to assume that we’re in this Fantastical new world that didn’t exist before I think is wrong just flat out I agree that’s the way my Tommy’s talking Seth interpret his fantasy world okay sure so so rather than like diving into

17:31 sure so so rather than like diving into all the other areas I don’t agree with what I’m hearing you say is your first person is the modeler and and front-end person more so I would say it’s closer if you were to consider powerbi the realm of just powerbi it’s a powerbi developer who’s focusing on semantic models powerbi developer now where does that power is it important where that powerbi developer comes from

18:04 Finance always Finance I Finance I know know yeah more more technology side more business side I would probably lean towards more of the technology side but again I don’t think that that to me is not as permanent as understanding to again to my point to reiterate understanding that they’re already building the purpose of a semantic model or data flow that’s been their focus because they understand it’s for the users or for other people to build that they

18:35 for other people to build that they understand that we’re not just building a report but I’m building something that can be replicated yeah I want that methodology or that mindset okay so I’m struggling here a little bit on this part I’m I’m super glad that both of you chose something different well you you got to choose one now Seth like so what are you choosing I I do scien I do do okay so this is the first person right first person you’re hiring you’re you’re hiring on I I I think it’s a a bi

19:08 hiring on I I I think it’s a a bi developer with a data engineering background so that’s that’s where I’m cheating to some you are cheating right but that’s okay well but but I I am I am and I’m not right like I would pick me so but but think about this think

19:23 me so but but think about this think about this right like data engineer obviously background DBA some somebody in technical spaces that understands structures of data and or now in in Realms of probably albeit harder to find would be somebody who understands like big data sure because those structures are slightly different not slight vastly different but also having an understanding of you but also having an understanding of how to develop something that the know how to develop something that the the business can Can

19:54 the business can Can digest digest the there there’s two parts here one is that person I think is for both of the reasons you guys stated Tommy I I disagree that we can go very front end because with the fabric ecosystem if we’re talking about a business intelligence SE fabric you have to have somebody who understands these structures to some degree on the back end in order to develop a solution

20:24 end in order to develop a solution that’s going to scale as opposed to just the solution I’m building just for this report report and to to to Mike’s Point right if it’s just the data engineer Tommy I think you have relev a relevant argument from the standpoint that well we’re going to hit a blocker where we’re pro we’re delivering what the business needs but the business isn’t going to understand the value behind everything because it’s not presented in such a way that can make like bring insights or like to the extent that somebody who

20:56 like to the extent that somebody who understands the modeling the visualization ETA or can do so I think it is I think it is pivotal that organizations try to find the expert that is going to lead these efforts that has the full spectrum that is really truly a technology person that understands the development from the end to end and just chooses one area or another because they like it better either I love building pipelines and doing data analysis and finding data

21:27 doing data analysis and finding data problems or I’m on the front where I love interacting with the business and I love like solving the the business problems and bringing them insights with data and as a byproduct of that I build models that extend and and and visualization so I’m probably cheating the most here between everybody and for the audience no we didn’t plan this right like I pick the back end I pi r and Seth Seth tries to cheat and pick everything but at the same time I think

21:58 everything but at the same time I think you make a good point where we can’t you can’t say if you’re going to draft someone into the NFL draft well I want someone who throws like Patrick mahome tackles like Ray Lewis it’s like that person I don’t want to say doesn’t exist but we have to understand too even in our space with powerbi we know that type of person with that ideal for us is limited there’s limited resources for even in just the powerbi world who can do it all and now we’re actually to asking to expand that I like well this is this is

22:28 that I like well this is this is why I started with my phrase here was like this say build a bridge from both sides right you can’t get to a report without having data and you can’t get to data without understanding what the report should do so like you you’re meet need to meet middle way there I would say your first hire should be very well versed in if you can both data engineering and powerbi so to your point Seth like if you your first hire should be strong regardless doesn’t really matter like you want someone

22:58 really matter like you want someone who’s had some runtime in Big Data stuff Delta tables understands paret because you can I’ve seen it happen where people don’t quite understand the technology enough and this is where I think some of these the dp600 tests don’t really do us a good service of really testing knowledge of people because it doesn’t I haven’t taken the test yet so I can’t speak 100% to what the questions are on the test but from what I’ve heard from people is it’s a good test of knowledge of general knowledge around the like very there’s some very specific

23:29 the like very there’s some very specific things there’s a lot of different touching points across the entire fabric ecosystem but it’s not telling you like how to what is the optimized activity doing what is what is the structure of a Delta table how do I keep costs down where do I look for information around capacity reporting and find that compare it to storage costs right so to me there’s like some really we real world technical pieces that I want my person to understand so I want someone to be able to speak to those have some experience around that

24:00 those have some experience around that and then again if you’re talking about your first hire they’re the ones administering the admin portal they’re the ones probably building your workspaces the ones planning which teams get powerbi and do we roll it out all at once do we take a business Le approach or we an itl approach right so that first person I think is very important to have a lot of skills on them and what I think I find a lot in organizations is they already have someone who’s doing powerbi where we need more skills is we need that expert in powerbi to a little

24:31 expert in powerbi to a little bit more educate on the data engineering side and I fly as this thing grows there’s definitely going to need to be some story around okay well we now have the core of what we’re trying to produce like working where do we plug in a data scientist now do you do you go use semantic link to go grab things out of models and figure how that works like so there’s going to be I think additional roles in here as well that are going to potentially pop up and you’re going to be able to need to have these these teams to work together

25:02 have these these teams to work together so your first High regardless I think to your point Seth needs to be strong they’ve got to understand DBA they got to understand Delta they got to be able to understand models and Reporting and that’s a hard person to find yeah I I would say a a way to get around that is is to hire two people not one like to like if you’re all in on fabric or it’s a starting point right well but Mike what to your point like if you’re on a

25:32 what to your point like if you’re on a business side and you already have somebody in powerbi a perfect a perfect way to implement or dive into fabric would be okay you hire the data engineer corre and you pull that powerbi person into Federated model yeah and this is what this talk about in the road map this Federated approach these two these two key roles work together towards what Adrien and the chat points out which is all of this should be business-driven based on the objectives that the organization

26:03 the objectives that the organization needs and we both know or all of us know and including our audience right powerbi offers us that that quick win so you need that you need to keep continually continually delivering things for the business to engage their interest their solving their problems Etc and you need the data engineer on the back end to build the the go forward strategy depending on the business needs are going to be because they ever they’re ever increasing yeah right and that’s where that key part of data engineer

26:35 where that key part of data engineer comes in where okay great there’s this new ecosystem we’re going to consolidate everything around this here’s the strategy by which we’re going to execute and by separating out those two things you can definitely like not look for the the purple unicorn out there from a hiring perspective but ultimately like those those two roles I think get to play very closely together now much more so than they did in the past with all these other services like we’re in different environments completely and I’m just handing you data

27:06 completely and I’m just handing you data no now we’re all in the same place hey I need you to like create this object move that forward share it over to this workspace or I’m developing this and it’s a a much tighter knit back and forth I think that will accelerate a lot of the projects But ultimately it’s it’s about being able to support both of those things in order to Accel at what you need to from a deliverables perspective because if you choose one or the other each one of those roles is going to hit hit a block in fabric right

27:38 going to hit hit a block in fabric right or any ecosystem which is like oh well now I’m in I’m a data engineer I’ve I’ve done the whole pipeline now I’m in powerbi how do I build a visual right like how do I create this thing and the same thing for the business person or somebody who knows powerbi well everything’s done in power query out would love that that just continues but like like I want to Branch out in what is what is this Lakehouse what is a data warehouse like I what does that mean

28:09 warehouse like I what does that mean right and then you’ve got to do a lot of ramp up and learning and all those things are slowdowns for the business so opportunities to learn absolutely but you’re setting things up in a in an optimal way and then depending on how the team builds and grows then you have a lot of different choices and I think we delve into other aspects of what’s important when building a team I love what you just said because that’s two things I was I going to mention I’m asking myself through this conversation two fundamental questions about who I’m hiring number one what is the final

28:42 hiring number one what is the final destination right now for the workflow in fabric if you looked at building the lake house what the UI gives you it’s a semantic model right now that final destination is really building a semantic model so that’s really the should be the core Focus according to the current workflow and then the second question I’m asking myself is how much

29:03 question I’m asking myself is how much how much a lotted time do I have for rampup if whichever person I’m bringing in whichever background they’re coming from I think we’re both we’ve already established or we’re all are greeing that person’s going to have to have some overlap in the other skill set so how much time am I giving that person to learn X Y and Z and to be familiar and much more importantly comfortable with but my really fundamental question here is I’m going through this is what is the final destination for a lake house or

29:35 final destination for a lake house or for the Sol deliverable in fabric right right now say the question again what is the final destination for a deliverable in fabric what’s the final solution based on the UI in fabric I’m calling it The semantic model everything leads up to building a semantic model right now I think that’s your story today I’m not sure if that’s your story forever right so me again I’m

30:05 forever right so me again I’m I’m talking with other MVPs right now about hey we can use semantic link to go pull out some data from a model and we can go run some machine learning on top of it somewhere else or do there’s there’s other opportunities to start pulling this data out and leverage other places I don’t know so yes I think for now I think the semantic model makes the most sense but it it feels like there’s this concept of if you read some of the

30:35 if you read some of the Kimble documentation or where theology on like how to build warehouses and stuff Donald’s talking about this in the chat right now which I think is really good building the warehouse is conform dimensions and facts to what you want to be calculating so it really does make you take a considered effort of like what are we really measuring what are our main kpis that we care about but if you think about the semantic model the semantic model can sometimes be little tiny smaller versions of like what the actual data warehouse is doing and so when you

31:05 data warehouse is doing and so when you look at look at architectures of what you’re building you you you you have potentially a larger Warehouse where all the things live with all the the dimensions all the facts that you want in a larger pool and then you have potentially these semantic models that might be just carving pieces of that off for specific business function so again I think it gives you an ultimate amount of flexibility to be able to design what you need to go back to what we’re trying to get to is right business can’t make

31:35 to get to is right business can’t make decision without numbers they need some numbers they got to trust our data so that’s where we land yeah I don’t understand the question or the statement you’re saying what is the end goal of fabric what is the primary deliverable right now according to the uine fabric and to me everything all roads lead to building a semantic model when I build a warehouse or when I build a lake house say hey your next step is build a semantic model I’m not I’m struggling with I’m struggling with the point of saying like

32:06 struggling with the point of saying like what’s the end deliverable the end deliverable is insights and kpis for the business to make business value it’s business value so like whether that’s an Excel sheet whether that’s a p report whether yeah and I’m saying it’s gray it’s gray yeah because I’m saying the actual technological deliverable because I can build a there is no technology deliverable the business doesn’t care Tommy the business doesn’t care that you’re building the Taj Mahal in Fabric or and you call it a semantic a semantic what they don’t care they don’t

32:39 semantic what they don’t care they don’t like like I I do want to lean into this okay right and a less non-aggressive tone Oh no no I like the aggressive build otherwise my Carlo is GNA drop in joke here to ease the move I got chat GPT go hit up some jokes here pretty soon cuz that’s getting a little I think this good like no my my point being I agree with Mike 100% oh the the only reason they’re spending money on business intelligence is because we can produce insights that

33:09 is because we can produce insights that the business does not have the value to the business is in the deliverable itself and what actions they can take on it the only reason they’re investing anything is because we’re producing efficiencies or Revenue generating IDE is or things that help them make decisions better and faster that’s the end goal it has nothing to they could care less if you

33:39 nothing to they could care less if you built an individual model for every single report as long as it was fast if you could do it the same way every single time instead of going o now it’s going to take us a month the reason we build these this infrastructure behind the scene is so that we can produce faster results to the business has nothing to do with the end goal being a semantic semantic model and and I and the reason why I’m hesitating here a bit Tommy on your on your point around I think the semantic

34:09 your point around I think the semantic model is a very pivotal piece of fabric it’s very used it’s very common it’s it’s a piece of technology that accelerates or speeds up this ability to get get from things I want to know about data into actually getting answers about the information right it’s it’s great at aggregating it’s great at filtering but as I think about about fabric now and the possibilities of doing other things with it like I can build shortcuts to my Warehouse and give those shortcuts to other teams in other workspaces so I

34:40 other teams in other workspaces so I I think I think to your point mean I I think I think to your point Tommy like semantic model when you lick at powerbi has been a very pivotal element of it what we get now is a lot more flexibility to give people other aspects like for example go to the SQL analytics endpoint you could go in there and just say hey write SQL against tables that are in the lake house that may be what some teams need I don’t know but the ability the fact that we have the ability to do this and I think the to me the aha moment in a lot of this is

35:11 to me the aha moment in a lot of this is the friction to do that the friction to get users into to use SQL endpoints semantic models powerbi report has been greatly reduced and we’re now at a place where oh you need access let me type in your name now you have the workspace or let me create you a new workspace and I’ll just send some data to you the ease of doing that doing that in in Azure with all the permissions and everything was much more difficult previously this is this is a way faster

35:42 is this is a way faster speed there’s I’m not worried about network security it’s all built into it’s it’s built into the ecosystem it’s that stuff that I’m like man this is really helpful and that I think is what’s adding a lot of value to the business and and let me get I’ll give the reply to that’s maybe less hot takish and number one the semantic model to me with powerbi again I feel that it’s being elevated in terms of where it lies in terms of a data solution it’s no longer the end goal semantic model report semantic link I

36:14 semantic model report semantic link I think what else they’re doing with semantic model being able to be in the service the semantic link is getting elevated in terms of the use cases for it it’s no longer just for a report the other point Mike you’re helping build shortcuts we’re building lake houses why well at the end you’re really help you’re helping those other teams know you’re helping those other teams probably build a semantic model you’re not just giving them a shortcut for the sake of a shortcut you’re helping or or it’s a shortcut so they can just access it in the SQL nlx endpoint it doesn’t have it doesn’t have to be a and

36:46 doesn’t have it doesn’t have to be a and it’s my point is like yes probably again semantic model is probably the most commonly used element because it’s caching data it’s super fast it gives you the answers quickly great but that may not be the final destination for a lot of your things and now that we’re adding data engineers and potentially even data scientists to this workload data scientists probably won’t use semantic models other than to get data out or export it but they’re going to be using their primary they’re going to be in notebooks they’re going to be looking at Lakes they’re going to be getting a bunch of tables so there’s a lot of other new workloads I think we’re

37:17 of other new workloads I think we’re we’re experiencing here that is going to change the Reliance on solely just the semantic model not saying it’s a bad technology or we’re going to move away from it but it’s it’s now there’s there’s the yes and a whole bunch of other things too speaking of yes and while while Tommy was talking or during this dialogue there a thought that struck me based on some recent work we’ve been doing that I wonder whether or not fabric is going to open the doors more on

37:49 on is the business or opportunities to engage the business in looking just for data if data if efficiencies and what by that if you think if you think about like especially in systems where data is harder to get right because it’s it is unstructured it’s in Lake it’s Etc there are so me like we one of the the biggest things we did even in Consulting together Mike where like you walk in with powerbi and it’s like guys you’re doing this like super manual Excel process right like powerbi can

38:20 Excel process right like powerbi can solve this for you yeah in that same vein though there there are those use cases but many where you have all of these third-party systems in organizations where sometimes business users just need listen I’m generating this report from here I’m generating this one from here I’m loading into Excel I’m doing all these processes and then I still need to

38:44 processes and then I still need to analyze the DAT the granular data output right could there be value of sucking it into business intelligence sure but that’s not their need I almost wonder if fabric now opens that door to have like an easier workflow for engaging with business and saying hey let’s let’s build an automated pipe for you on this and we’ll just we’ll we’ll connect to these things we’ll do the Transformations we’ll we’ll knock out all of that stuff that takes you a really long time and and we’ll put this CSV file in this location for you on a

39:15 CSV file in this location for you on a weekly basis or something I think this is a real I think this is a real solid use case I think to your point there Seth right there there’s going to be some experts or people that are hired and again this is I think there’s going to be a continu shift here in the business side I think the business is going to hire some more technical people or they’re going to find more people that are both business fa facing and Technical in nature because I think there’s a lot of value here so a lot of times when I think people export or need to see all the data there’s trust issue with the data the reason people are asking for

39:47 data the reason people are asking for the very fine grain detail either you need an answer around a very specific thing or you can’t trust the rollup of numbers to a certain level and so I think a lot of times what we get ourselves stuck into it’s like this is a potentially a trust increaser where I give them access to more of the raw information and teach them hey here’s here’s a Lakehouse here’s a here’s a SQL analytics endpoint go here and then here’s how you write some basic SQL to get some dat together they can start picking at the information and learning what it is but I in your in your comment

40:19 what it is but I in your in your comment of this Seth it makes me really feeling you need to be at least very a concerted effort to figure out who will own own what things and when you give the business time to experiment with stuff which I think we for sure should do what what is the mechanism the pattern the the the whatever the process is to say okay you’ve you’ve had time to play with this stuff business you’ve added some value to it you even have a report maybe you built what do we do do you want do

40:49 you built what do we do do you want do you need to retain ownership of this or do we need to pull it back into Central does it come back into a main process that the central business intelligence team will own or do we just leave it alone and let the business continue to manage that space so this is where I think a lot of the the process pieces is going to push around a little bit around more conversation should be held around we’ll let you run wild but you own it like our hands are not tied it’s if it’s bad data you can come back to us if it’s not bad data and

41:20 come back to us if it’s not bad data and that you’re shaping it and doing weird things with it and you’re reporting to a VP based on the output of that You’ changed it you don’t come back blame us so it’s it’s this by how do I say this by bringing more people together and having more access to data there’s going to need to be more conversations around who’s the stewards of the data and I don’t think a lot of companies have a super strong ex culture around data stewardship and this is accelerating that conversation I I I agree with you and I would almost I

41:51 agree with you and I would almost I would almost pose the question do you think fabric is going to cause more friction in organization initially yes but it should get better over time I would think right hopefully well the the precipitous of that question is you already you already have organizations that like have different departments that are like yep we’re going to be data driven we’re going to solve these problems for the business and like Finance great example right like we’re we’re they’re definitely broaching into business

42:22 definitely broaching into business intelligence experts on their team to solve y solve y problems in finance right so taking that a step further now I have fabric now I can build my own ecosystem like we now own fabric and all of the things are they going to stay just in finance or is that going to Branch into now there’s a central bi team and they’re trying to do something in Fabric and then there’s the marketing team and they’re building things in fabric because they nobody’s

42:53 things in fabric because they nobody’s talking to each other right like now all of the sudden what you do with three different departments that all have big fabric ecosystems like I yeah you’re just supporting that continually as an organization or what are you doing with those resources are they 100% dedicated because the reason I asked that question is like being a central bi guy right with a central data team I already know that a a a central bi team is a support team to to some degree yep but but not

43:24 team to to some degree yep but but not every single one of these departments needs 100% dedicated time all the time exactly right right so so like to get the most like value for dollar like in my mind it makes a ton of sense that that isn’t just dedicated to a singular area you would have those resources plugging around and doing things for other parts of the organization that also need assistance so that’s why I’m poking that question to see you poking that question to see what you guys think well and I I

43:54 know what you guys think well and I I don’t know if I’m going to contradict myself but I think as a I consider what you’re saying it’s expansion of the bi team Services three of the mo major projects I worked on as a bi team probably should have been in a Lakehouse rather than in a semantic model the way we were pushing data together and integrating data together or mashing data together we didn’t have that ability at the time but it was going through CSV files and mergings and app pendings and multiple data flows but it would have made more sense for the business to be in a data engineering

44:25 business to be in a data engineering role and stored data in a lakeh house and I think that’s for me that to answer your question this is really talking about if I’m looking at the bi team this is just expanding their roles rather than help me build a report help me build a a model that we can build reports off of to your point I think it’s we have this data can we store it we just want it stored somewhere locally or we want it stored historically whether that’s in a simple lake housee or that’s their own environment that

44:55 or that’s their own environment that they can play in that goes perfectly into the idea of okay we’re going to put this data together in a lake house you can query it you can look at it any time it’s not only report level view it’s you can look at the raw data now know you can look at the raw data now you own it now you own it until it gets the report so it whether or not that’s friction or not it’s to me it’s the same thing when we build that first report for a team and they go that’s not right you’re like that’s your data and yeah I’ve had a number of conversations where

45:25 I’ve had a number of conversations where we show up and we build some reports or some models like I I didn’t expect it to be like that well we’ve never really actually had a full view of data without people trying to put their hands in the pie and like massage this number and massage that number and clean this over here like that’s the way the data comes in so I in so I was talking with somebody else mean I was talking with somebody else and and the there is years of data engineering that has been occurring in places and organizations and think I’m go back to I’m going to go back to finance saw a

45:55 I’m going to go back to finance saw a store procedure recently had well over 3, 000 lines of code drop replace tables all over the place really interesting stuff but this was like a five six year development plan of like figuring out what works and cleaning some data there’s a whole bunch of wear claes in here well wear Clauses means I’m cleaning data I’m I’m picking up data fields in the column and cleaning stuff great you have all this Rich business logic and then expect someone to come in from the fabric team or the powerbi team and be like hey rebuild

46:26 powerbi team and be like hey rebuild this entire store procedure in a week like I’m going to need a little bit more time to figure out what’s going on but the the idea here is like these things are already happening right there’s data engineering occurring likely somewhere in your business and people are picking up and figuring out how to get this stuff done because they got to deliver results they got to report out and make things happen what I’m what I’m saying now is you now potentially have a tool where you don’t have to do things in a vacuum or a silo and this is where we start talking about okay who really should own things what

46:57 okay who really should own things what does the stewardship look like and Jackie you’re spot on you’re talking about this this is where the center of excellence steps in and says hey look we as a company we need to decide data is valuable it’s an asset how are we going to get teams to play well with that with that data space and I agree with that when and if you have that and my my my what my perspective and and comment around that that fabric ecosystem is many organizations don’t have that yet true right that’s true I I agree with

47:27 true right that’s true I I agree with Christian like fabric is not self-service business intelligence that’s powerbi and that’s where that belongs fabric opens up this all other ecosystem that Mike I think drives directly into the conversation that an organization needs to have otherwise you may have factions within your organization that are building their own Solutions and that’s not what that’s not going to in the end solve your business

47:57 in the end solve your business intelligence needs right you’re it’s the same conversation we’ve been having right and and maybe maybe we need to like have an episode where we just talk about that nugget of how do you transition from nothing to fabric to Bringing people and organizational units together because you’re right Mike how are they solving that today probably in third party systems the vast majority of the time it’s in it’s in a separate area so is it a great opportunity for you to

48:26 so is it a great opportunity for you to like solidate all of your information into these areas that potentially can be shared with one another absolutely right but it requires those teams to work together at least somebody to give them the vision that hey there’s a real value here in that we don’t need all of our stuff in these third party systems you guys can actually still secure your area but those pivotal things that you guys always have problems with sharing numbers like here’s how easy it is to

48:56 numbers like here’s how easy it is to share and this ecosystem and that’s where I think fabric does take off it’s just how do you how do you get organizations to see that path and or get to the point where you’re pushing data strategy into the conversation because it is really valuable and I I think this is just to me making me think about who am I hiring or who am I looking for so but like the all the things we’re talking about I’m going leaning back on my own experience I’m looking for the central bi guy who’s F or person who’s frustrated at their role

49:27 or person who’s frustrated at their role because they’re probably already doing a lot of data engineering but they don’t know it or their Solutions are asking for it but they don’t have that capability like I said I on from user groups and again especially my own experience had I had this technology available to me at my roller company this would have made things a ton easier but it was still the same idea we have this raw data we have to clean it then eventually build a report so I I have never gone to any operational system system and had any data come out really

49:58 system and had any data come out really in a shape that RBI any reporting loves even if even if you go to the even if you go to the source system and pull out data and say I’m going to pull it in and do it in Excel there is no you’re always doing something to shape the data it’s always going to happen because it’s your row level detailing and now you’re like aggregating in column information like it’s a different structure so it’s going to need some level of touching that yeah and and a bit of exaggeration but if I’m hiring or interviewing like did you have an sta engineering team nope great then you’re my person

50:30 nope great then you’re my person basically because they had a s they had to think in this mindset and they already understand powerbi I think obviously they need a ramp up and I think I I do not want to downplay Mike and Seth what you’re both saying you have to have someone whether or not they learn it or there’s their ramp up time that is going to know that more than just the fundamentals of a medallion approach and the data engineering background I don’t think that’s something that can’t be learned from someone who who’s been in the back end of powerbi for a long time now all right

51:01 of powerbi for a long time now all right well we’re about 51 minutes into this episode and I would like to ask this next question so who’s your second hire no I’m just teasing because that would be like another hourong conversation as well that we’ve fired our first person but this this conversation I think is coming up very relevant and the only the only other thought I have that comes along with this when Tommy and and Seth I think you make very valid points here one other thought that that does trigger some mindset

51:31 some mindset here when you think about companies how they structure their company dictates a lot of how well we are able to communicate across departments so I’ve been in some organizations where there’s very distinct silos of product and that team has its own profit and loss it’s penl for that particular thing or you’re in a you’re in a company that have been acquiring other companies right and when you acquire companies you typically get other Data Systems someone has to make the decision to say look are we going to centralize all these systems to one

52:02 centralize all these systems to one system or are we going to let these other business units run themselves kind other business units run themselves independently for a period of time of independently for a period of time and I think how you structure your company directly impacts will an a will a center of excellence actually be effective because you now have made potentially multiple silos in your business because of the pay structure or what they’re allowed to buy or not allowed to buy business unit one says hey I’m going to go buy Salesforce business unit two says no I’m using

52:32 using Dynamics who’s who can override that decision and make sure that we’re having the same decision around what makes the most sense for the business I don’t really care what you pick just pick one or trying and figure out how to merge them down to one so I do think there’s like there is some level of like corporate oversight and structure that does play a factor in here as well it either encourages communication between departments or it potentially increases the siloing effect yeah I think final

53:02 effect yeah I think final thoughts for for me essentially business intelligence exists because it creates business value ultimately fabric is just a new method for us to solve a ton of technical problems that we’ve had in the back end with all of these services in different areas and it helps us streamline how we go and build those so these new methods still need to include quick winds while some of these data organizations processes and storage to maximize benefits in the future all get baked out this is one of those key

53:34 get baked out this is one of those key areas where organizations should lean in and understand how they can leverage this new ecosystem and have the organization aligned to those for long-term goals to really maximize the benefit because this is a unique platform that I’m not aware of any other competitor to bring all of these tools and services together in the manner that they have and it it doesn’t it doesn’t remove any of the same self-service abilities and value that powerbi offers it just

54:04 value that powerbi offers it just enhances and opens the doors for organizations to be much more I think cost efficient as well as create data ecosystems that they can leverage and share with each other in a much much more meaningful and faster way yeah and I think as I consider the question rather than the first hire there’s two diverging roads on how you structure the team am I solving for fabric now in this current state or am I solving for fabric or what I think it will be I think those

54:34 or what I think it will be I think those are very two fundamental ways of how we’re going to build this team and how we look forward to not just the one person but the group of people that are going to be part of this yeah I said it at first when Microsoft originally announced fabric out at build last year literally it’s been here about a year oh yeah crazy so I was very excited about it because a lot of Technologies I was using and building with at the time was being incorporated into powerbi and I think I think Microsoft is very good at

55:04 I think Microsoft is very good at thinking that way I definitely felt there’s a lot of rough edges but I definitely think because of their concerted effort to continue refining and perfecting and this whole encompassing solution of ease of use platform multiple tools that do exactly what you want or have have capabilities data engineering data science business business intelligence pieces like the there is a wealth of really smart people trying to figure this out and make this easier for businesses to use and so for that I’m

55:36 businesses to use and so for that I’m very excited about where things are going and I’m seeing a lot of improvement since it came out things are much easier to use I’m much happier using spark and notebooks now than I was when it first came out so major improvements in that area but Microsoft was new to that space they they didn’t have the architecture yet figured out and so what if one thing Microsoft’s really good it’s Microsoft as good they’re not first to market for anything really they’re like slower to Market because they’re big but when they come in they come in with great UI really costeffective Solutions

56:09 great UI really costeffective Solutions I think they deliver a lot of value for that for the spend of the money so I’m very excited to see where this is going to continue to go we’ll hopefully have some more key thoughts and things that we will’ll do for Thursday’s call because we’ll actually had some runtime and some announcements out of build and see what where they put their their money what what things in fabric are we going to be getting here that’s going to make this even easier or what new features are we going to be looking for so I’m looking forward to seeing what the the next phase of development looks like inside

56:39 phase of development looks like inside fabric Tommy what is your final thoughts that was my final thoughts before you talked good two roads I can’t wait Mike for all of your short hot takes and all these videos that are going to come directly out of the Learning Center either way Thursday should be a fun call don’t talk me up too much we’ll see how well I do I just want to know how many of our previous episodes are going to be irrelevant now well we’re we’re just gonna co-pilot everything anything co-pilot everything we just let co-pilot handle it all so

57:09 we just let co-pilot handle it all so we’ll see what happens anys with that thank you all very much we really appreciate your time and your listenership hopefully the audio was good and the video is still okay today we appreciate your time and we would love it if you would share this podcast with somebody else we we spent a lot of time thinking about these topics we do this every day hope there’s some nuggets of wisdom here that you found from our experiences in working in this space if you found that let somebody else know you enjoy the content as well and share somebody else Tommy where else can you find the podcast you can find us on Apple Spotify or wherever

57:39 can find us on Apple Spotify or wherever get your podcast make sure to subscribe and leave a rating it helps us out a ton do you have a question an idea or a topic that you want us to talk about in a future episode maybe it has to do with today head over to power. tips podcast and leave your name and a great question finally join us live every Tuesday and Thursday A. M Central and join the conversation on all of power bi. tips social media channels excellent and we’ll see you next time

Thank You

Thanks for listening to Explicit Measures. If you enjoyed the episode, please share it with a colleague and leave a review— it helps others find the show.

Subscribe and follow for more Power BI content from PowerBI.tips.