From Fragmented Models to Unified Semantics – Ep. 481

Every team eventually hits the same wall: you can’t scale analytics on a pile of one-off reports and duplicated models. In Episode 481, Mike and Tommy talk through what it looks like to rebuild your environment with lessons learned—moving from fragmented datasets and inconsistent definitions toward a unified semantic layer.

News & Announcements

• Microsoft Named Leader and Outperformer in the 2025 GigaOm Radar for Semantic Layers & Metric Stores — Microsoft highlights the growing importance of semantic models as the backbone of trustworthy analytics and AI-ready context. The post reinforces a key theme of the episode: a semantic layer isn’t “extra metadata,” it’s the shared business meaning that prevents metric drift and makes self-service (and AI) reliable.

Main Discussion: Rebuilding with Lessons Learned

Why Environments Fragment

Mike and Tommy describe the common drift pattern:

• New report requests lead to new datasets
• Teams copy/paste measures and tweak definitions
• “Temporary” shortcuts become permanent
• Eventually, the organization has multiple versions of the truth

The Semantic Layer as the Fix

Their argument: semantic models and metric definitions must be treated like product assets.

• Shared dimensions and definitions reduce debate and rework
• Reusable measures prevent KPI drift
• Model-first thinking makes downstream experiences (reports, notebooks, AI agents) easier to build

Governance Without Killing Velocity

They discuss the balance between governance and speed:

• Too little governance → chaos and conflicting numbers
• Too much governance → nobody ships anything

The practical approach is incremental: pick the metrics and dimensions that matter most, unify those first, then expand.

Looking Forward

The episode sets up a bigger 2026 theme: as Fabric and AI make it easier to create artifacts, the real differentiator becomes semantic discipline. The teams that invest in unified models and shared meaning will move faster long-term—because they stop paying the “rebuild tax” every quarter.

Episode Transcript

Full verbatim transcript — click any timestamp to jump to that moment:

0:01 What you do? Good morning and welcome back to the

0:34 Explicit Measures podcast with Tommy and Mike. Hello everyone and good morning. Hey Tommy. Good morning Mike. How you doing? 481 in the books. Not too bad. We’re getting pretty close. We’re creeping up on 500. Oh, I know. What are we going to do for the episode? I don’t even know. Send send your ideas if you want us to talk about something special for episode 500. , I think it’s safe to say, Tommy, we are the largest podcast around all things PowerBI and data related. We’ve got regular users. We’ve got lots of consumption on all the social media platforms. It’s been

1:06 Fun to see this thing grow a little bit. Yeah. , I remember early on in the in the very early days, I was like, Tommy, we just got to produce more. We just got to go faster. We just got to get more out the door. And it will grow if it’s if it’s good. And it seems like it’s doing pretty well. Anyways, we’re still doing this. Yeah. And honestly, like I said, I I just enjoy having being able to talk it out, too. But it’s amazing. Every time we go I feel like at a conference, someone comes up to us and says something. People know when I forget to upload one of the audio episodes, they let us know right away. Oh, yeah.

1:37 And yeah, we get to see the numbers, too. But honestly, if you told me that we’d be the things that we’re talking about now, , fabric wasn’t a thing when we started. It was literally just a PowerBI podcast. That’s true. Actually, that’s very true, Tommy. The first two years, I guess. , , fabric’s only been out for two years. So, I think remember I remember halfway through going I think we were like debating, should we should we go into fabric? Should we talk about fabric? And now here we are. I think if you look back at our most recent episodes, I think the majority of what we talk about now is just speaking of fabric and bringing now speakers from

2:12 Across the community into the fabric, , story and and hearing what other people are building and learning from people from Microsoft. This is extremely fun. I thoroughly enjoy this. Absolutely love it. Jumping in, our main topic today is a mailbag item. So, we’re going to have someone talking from a mail-in letter. I guess it’s not really a mail-in, it’s like a form-in letter, right? We don’t really mail things anymore. If Tommy, little side note, well, I’ll say the topic. I’ll come back here to like a little rant I have. mailbag from I don’t know if actually I don’t know who

2:43 It’s from, but it talks about fragmented models, how to build unified semantics layers, and what does this look like in lie of a company that just started, built things haphazardly and got different results. What does it look like to to rethink your your design of the system? Do you incorporate if you’re a PowerBI shop only, do you start incorporating fabric into your designs? Is that something worthwhile? , and we’re going to unpack that a little bit here. Now, anyways, , very excited about this topic. I think this is a good topic and

3:14 I think very relevant for many organizations who took a first pass at PowerBI and now are rebuilding or rethinking slightly. So, anyways, just interesting here. All right, Tommy, a bit of a rant. We say mailbag. Have you ever had an interaction where you are, for whatever reason, something at home, you have to fill out a form, you have to send in some money for the kids at school. I don’t know, I don’t know what it is, or something out there, and they’re requiring you to write a check or Oh, yeah. send something in from the mail. And I’m thinking to myself, this is so archaic. We are in We have digital everythings.

3:50 Get on board, people. I’m so frustrated. I don’t even own a checkbook anymore. And I had to go to the bank and like get like sheets of checks because I’m too lazy to go get like a whole book of checks. I’m like, “What’s the point? I don’t really want to write checks anymore.” There’s occasionally like every four or five months we have to write a check, which is incredibly annoying to me. It’s so dumb. Yeah. , honestly, I switched to a bank that was really online only because I realized that the in-person experience is probably the least important thing of a bank.

4:21 Yes. especially now. Yes. , and they I have to order checks like every so often because Yeah. especially for school stuff. And then the worst part is my wife’s friends like she does some stuff for like for the kids and there’s other events and they’re all sending checks. I’m like my wife used to go to Apple Pay or to any of those pay. Oh, are you kidding me? I wish I was lying. Oh, why? What’s the resistance to Apple Pay? I absolutely love Apple Pay,

4:53 Honestly. Like I know any anytime in our family someone buys something for someone else or like we buy something and we need to like reimburse somebody like my immediate go-to is Apple Pay. It is so stinking easy. It is the best. I absolutely love it. I think it’s so cool. Yeah, you would you would say so. , she doesn’t even have autofill for information on her phone. She just because every time she needs to get something she has to she comes down, she’s like, “Can I borrow your card?” I’m like, “Why?” Oh my goodness. Yeah. Oh yeah. Oh my Oh my goodness. Tommy.

5:25 Tommy, you’re living in the stone age. Well, I am not doing that. [laughter] Your family’s living. Let’s be clear. Wow. I I can’t even fathom. I think I have all the all the things stored on my phone and try to make it as least amount of memory as I need to. I Tommy, I got to be honest. I still can’t even memorize the the codes on the back of the cards for like the security codes. I still don’t have those memorized. like you, oh, there’s too many passwords and secrets and things rolling around anymore. I can’t even keep it all straight.

5:56 Well, finally Microsoft Edge is now keeping pass keys in like synced up with your Microsoft account, which is [clears throat] great because I dude Yeah, from a security point of view, that’s oh my goodness, I can’t stand it. I to be honest, Tommy, I have found I really struggle. So, this is another security identity thing. They’re talking like passwords and autofailing things out. I have really struggled with the Microsoft Authenticator app. It was really nice when it first came out. I thought this is so cool. I liked how it integrated with Entra ID so easily and I could just

6:30 Send it a message and then the phone would just show me a number. Dude, I hate it. And now I can’t. I have done one phone migration with the Microsoft password app and all of those token security things that they put on the app, they get stuck in that phone and you have to I don’t know how to do it and someone in the chat may be able to like explain to me how to transition from one phone to the other phone with the Microsoft authenticator app, but I have to physically go through every single account and disassociate the MFA and [snorts] then reassociate it

7:05 With the new phone. And I think to myself, dude, I’ve got like 25, 30 different emails that I’m trying to sign in across clients and other things that are all fade together. I’m like, this is insane. I don’t want to do that. Well, once you do set it up though, it is pretty nifty. I will agree, but it’s it’s the the immense amount of pain to transfer from one device to the other device really turned me off from the application. I was like, I don’t even like this. And the password app used to be able to save passwords. Can’t do that anymore. They turned that off.

7:36 Go to Microsoft Edge. I know. So then, so now I’m like, okay, what’s the really what is the point of the app now if it only does the MFA portion only and I can’t even store passwords and secrets in it. And now it’s like I don’t even want to use it anymore. So I’ve basically ditched the whole thing. I am for better or worse, Tommy. I don’t know some people have opinions about this. I am an iPhone user and I’ve gone all in on iCloud passwords, iPass. Really, it’s way better. I have extensions I can use in my browsers. It synchronizes across my different devices. If I have

8:09 My MacBook Pro or my iPad, it just works in all situations. You make it once, you can make groups of passwords. It’s just it’s the way it should have been all along. It should just have been this system. This is how it should work. So, what Apple’s got it dialed on the password management app side of things. Doesn’t cost me a dime and it works really well across all my Apple devices. So, I’m like, I’m done. I , but what about your computer though? It just there’s there’s an extension for the browser

8:40 And I can just have it right there. Or there’s actually an iCloud passwords app on my computer and it makes me authenticate with Face ID to the app to get the passwords out of there. I can see everything clear as day, no problem, right in the app. So if I nearly need to go to the passwords or get those security codes or something like that, it’s [snorts] easy. It’s and the autofill on the extension and or my phone works way better because when you did with Microsoft the you had to leave your website to go over to the app to punch in the number or select the

9:14 Number, you don’t have to do that anymore. It just autofills it out for you. So if you have like the email address you’re putting in for the password for the website email address and then you put in your password and it says okay we need the authentication code can you give us that? So instead of leaving the app having to go to a different app punching in the code and coming back to it hoping that it refreshes and reloads the screen. You don’t do that now. You just click the little cell and the little thing says here want me to put the code in for you that we already have stored on your device? Yes, please. And you just push It’s so good. It’s so good. I like it a lot better.

9:48 Well, now that we’ve gotten to be an IT podcast, I will I might have to look into it again. I’ve been fine with I just like things being simple and I never remember my password. The one that I the one I hate the most actually is having to sign in like a passcode to your email and just waiting for a dumb email to come. But anyways, okay, last last pitch on this time. I know we’re I know you’re trying to transition off this news eventually. One other note, I’ll just riff on this one as well. Do you

10:20 Share passwords with your wife? Like do you need to like give her passwords for the same accounts to the same things like the DMV or like the iPass? I know you’re around Chicago. There’s like an iPass thing that you have to go through. Absolutely not. So you make her get our own accounts for everything? Yeah. Oh yeah. Okay. So my wife and I share some passwords on some things. It’s just like, , we both log into Amazon with the same account. Like, we don’t have two separate Amazon accounts. So, she’s buying things or whatever off of Amazon. I log in with the same login that she does. With the iCloud password, you can have a

10:53 Family member or family shared group of passwords inside the app and I can just share her a group. It says these people in this family unit can have these same passwords. So, I can directly give her access to the shared passwords. And so I am IT support for our family. I I assume you are too also, Tommy. Oh, yeah. Oh, any little thing that comes up. Okay. So, there are occasions where my wife will be like, I don’t remember the password. I can’t get in. I can’t log

11:24 In. Something happened. So, I had to come in and troubleshoot what’s going on. And usually this the story is, okay, we’ll just go in and reset your password. She’s like, well, I don’t know what to change it to. You don’t need to change it to anything. Change it to whatever the phone tells you to put it in as, cuz it’s secret. like put the gobbledygook password in there so you never know what it is and just use the phone to authenticate and you should be good. Well, she doesn’t want to do that. So, I become password support and so I am now going through both of our devices and making sure all the standard phrases or all the little warning symbols that come up next to make sure all those

11:56 Are gone. So, I’m support and so now I’ve been slowly going through all of our passwords and updating all of them. So that’s like whatever the the phone decides is a 16digit character, letters, numbers, and dashes that it decides to pick out for the day. So you could torture me now and I I can’t I couldn’t tell you the password or anything anymore. So Oh, I tried that once and it was not a good experience, but I know exactly what you’re talking about. Yep. Yep. So, I’m since since I’m tech support, I’m like, I’m going to set the pace of like what we’re going to use for security and we’re going to do this and

12:29 We’re going to store everything in the iCloud password thing. And it so far, Tommy, I will say I’ve been very happy with it. There’s been very little friction with me going this direction, especially since I play it support for all the devices in our household and even like our localized family that lives here. I basically IT support for all of them. Honestly, I’ve been trying to do Microsoft’s p everything doing it passwordless and like just a lot of things are through like passwordless. Yeah, that’s what they say. But basically with their account like with

13:02 Pass keys everything I set up pass key with a with a pass key now. Yeah, but that depends on the website that you’re going to like or or site that you’re going to it depends on if they actually have that set up. Most are going that direction now but some don’t have that yet. M but that pass key and that pass key is only stuck on that device that you’re on. So like if you’re on your PC, it’s a pass key for that one. I know. I created on my PC on my phone. Yes. Correct. And but you have to But now again, you’re going to face the same problem, Tommy. If any one of your devices go down and you need to rebuild the device or get a new device like

13:35 Let’s say you get a new phone. Yeah. I I literally had to get a new PC over this weekend. So then none of your pass keys are going to work. You got delete them all. Yeah. So, you got to reset them all up now. Nope. All I had to do was since I already had a pass key on my phone, because you can create a pass key per device, was when I was signing into GitHub, I just scanned my phone on the screen. Then I was logged in. Then I said create a pass key on this machine and then I had it on my desktop and that was great, . But but you’re I don’t think you understand what I’m saying though.

14:06 I don’t think I am either. I’m saying now that you have a new device, your pass keys just don’t come with it. You have to now every site you go to, it’s going to be this little extra hook or step to sign in and you have to set up all the pass keys again for for my PC. Yes. Correct. And so that’s so that that’s my friction point here, which is like I I understand why that’s maybe faster and okay, fine. Like for some things it makes sense. On the other hand, I’m like I don’t want that headache. I just want it to let me in. Like scan my face. Let me in. Just let’s

14:40 Just go. Like it should not be this very difficult thing and I understand it needs to be secure. I get that but like it should not be so much friction that it’s difficult to run. No, we completely agree. Okay, I think it’s time we actually get to discussion. Anyways, total ranting. Yeah, Tommy, I totally agree with you. All right, let’s let’s talk about our news. You have a news item that you threw down here, Tommy. Let’s go through the news item. What are you talking about here? I’ll put this in this will be in the description of the video. So this is on the mig PowerBI blog, the official blog. And yeah, the name of this is an

15:14 Interesting name. Microsoft named leader and outperformer in the 2025 GigaM RA radar for semantic layers and metric stores. Okay. I’ve never heard of giga and I think that you’re supposed to say it. And also I didn’t know that was a category. Well, and also Tommy, when I look at this graph, like [laughter] these these if that’s what you want to call it,

15:45 Graph visual. I don’t know what you want to call it here. The visual is it takes me it took me a hot second to like what are we trying to show here? I didn’t really even understand like so we have like the magic quadrant from Gartner. We have the Forester wave which again it’s like a wave of things and they highlight the better or the worst sections of the graph. And so it’s it makes sense. This one is using a radar plot and I’m like looking at this one going interesting. I’m not sure if I really

16:18 Get their visual. I believe it’s supposed to be representing a bullseye. And the closer you are to the center, the better you are with innovation, feature play, maturity, and platform play. I don’t understand what I don’t know how to read this. Radar plots vendor solutions across a series of rings with those set closer to the center. Judge be of higher overall value. Okay, this makes this is like so unintuitive. Okay, I’m okay. Fine. Happy that Microsoft made it to the center

16:49 Here. This chart is trash. You don’t like this chart? No, I get that. I get you’re trying to make a bullseye, a perfect bullseye of something. I get it, but it’s so difficult to think the center of something, it’s literally got an x-axis, like a crosshair on it, the center is the highest number point value. So, and and you would think normally in graphs, when you do graphs on things in general, the further away you are from the x y coordinate 0, that’s 0 0. So, right. Exactly. We’re basically playing golf here. like

17:21 How close can you get to zero? What’s the lowest score you can get in this example here? And I’m like look at this go. I just it’s dumb. Anyways, and then they have to highlight like colors on each of the items and then you have to go look up the colors because the colors in addition to the position on the graph mean something as well. It’s not like an easy glance. Like what are you doing? I like I don’t even know what the metric sets are. This is insane. So anyways, they were became I guess based on that they’re a leader and outperformer. I can’t

17:53 I’m not sure if I’m going to trust the guys who can’t really make a graph, but let’s I’m happy that they reviewed all these other tools, but they did compare other tools as well. They they were looking at Google Cloud, SAP, Oracle, DBT Labs, which is interesting. I’ve been hearing a lot of language around DBT at scale, which I’ve heard a number of times about cube. I’m not sure if I’ve heard of Cube before. What What is Cube? , they look like they’re an outperformer. I’ve never heard of Cube either. Yeah, I’m not sure who Cube is.

18:26 So, that’s interesting. Anyways, so what this I and I had actually put this into Notebook LM because I had no idea what this whole thing was. But basically what this whole judging contest was is looking at the semantic layer which is not the semantic layer that we’re thinking of and it is in PowerBI but enterprise data stacks from traditional business intelligence and the most significant catalyst that they

18:59 Say for going over the smithic layer is generative AI yada yada yada. yeah so the they’re basically defining this semantic layer as standardization complexity reduction efficiency collaborations now how they do that semantic layer each person not sure and I have no idea what a metric store is that just creating measures [clears throat] okay so I’m I’m going to go a little bit on a tangent here potentially somewhat tummy a metric store is basically the semantic model in my in my understanding

19:33 There may be some other nuances around what the metric store is doing. I think this comes from more of like a SQL standpoint like I have tables, I have relationships and we need to define a measure and the measure is some of this column or it could be something more complex where you’re doing like a case statement or some function that’s driving a calculation that you’re going to use. Right? So there’s there’s this concept of like you’re you’re able to run functions that do some metric or aggregation logic and you’re

20:07 Just storing all those things in the same place. So in in my mind when I look at these companies and again I’m maybe going to be poking a bit more on like at scale. At scale does a lot of this. that says look you don’t really need to define like the actual table or you you can define a table but hey this column has a number this column should have some aggregations on it and this column should be grouped by these various things in my opinion it feels a lot like metric sets that we no longer have anymore [laughter] so

20:39 I know right this here’s the dimensions here’s the measure that this thing would relate to right and these dimensions may go with this group of measure Right? That that’s what it’s saying here. It almost feels like a perspective some to some degree. But this is these semantic layers are what we’ve been I guess in my opinion very comfortable and familiar with Tommy. We’ve been doing this for years now with 10 years now with semantic modeling and that and that’s a definition of how these tables relate what calculations I want. It’s all stored in a single semantic model.

21:13 Now what I would argue that PowerBI and Microsoft fabric does not have the strength of is how Tommy we have like a single model right we think about the the single semantic model and there’s data there’s a there’s this mix of like the data in the semantic model and then there’s a semantic layer the metric side of it what I think data bricks and snowflake are doing now when they’re building they’re now building their own semantic layers on top of stuff let’s remove if imagine a powerbi file with no data Let’s just imagine the definition of the

21:46 Tables, the relationships, and the measures, right? And not having the PowerBI file contain the data in it. Does that make sense what I’m describing there? It’s almost like a PBR is just the metadata. Yeah. So, if you just had just the metadata of everything and not the actual data, that’s what I feel like a lot of these semantic solutions are trying to produce. Look, let’s just define the relationships, the column types, descriptions, information about the data that we’re going to be using. That way, Aentic Things can pick them up. You can use chat bots against them,

22:19 Right? It it gives more context to other tools that you’re going to use to go build stuff with. So, that’s my impression of like what’s what’s going on here with this semantic layers. Does that make [clears throat] sense? Yeah, I guess so. I would be interested in looking at the other competitors there like especially for I want to know who at scale is and I know what DBT labs does but that’s usually like there that’s probably the most crossover from pure development to PowerBI what we usually hear. DBT Labs is interesting I believe they just recently got acquired I think

22:52 By somebody I think they just got bought bought out by someone. So, DBT Labs was open source. There is a paid level of it, but they just got acquired by somebody. , and I’m not sure who I can’t remember who it was, but there’s some other big D. Oh, for Fiverr. Fiverr just recently acquired DBT Labs. So, the open- source standard that is DBT Labs. I’m going to assume there’s going to be some lesser amount of DBT Labs will be less open- source in the future. But my understanding with DP DB DBT Labs is again it’s a very SQLbased language.

23:25 It’s helping you build like semantic layers of things. It’s almost like a in my what I liken it to is it feels to me a lot like a materialized views thing, right? Hey, define the structure of what you want and then you can then read that structure and then produce tables or things off of it based on a pattern, right? And you could have parameters and you can parameterize dev test prod connection strings. It’s like parameterized SQL that you can redeploy over and over again and it just rebuilds

23:58 These tables for you automatically. I’m not a big fan of dbt. I feel like it over complicates things way too much and you have to have a really strong team who understands what dbt is and it’s a lot more engineering than I think it needs to be in place there. So, I think I think you can get away with solutions that are cheaper and less complicated than DBT Labs. But, , it’s a commitment and and teams that I’ve seen bring on DBT, it’s quite the commitment to make sure that you really understand what’s going on. And in my opinion, if you just use data bicks,

24:32 Just use data bricks. You don’t need DBT in concert with data bricks. Data bricks has all the things that DBT does and more. And I think M data bricks is a very mature space in this as well. on this GigaM report. I don’t know when it came out here, Tommy. It said it seems fairly recently. Yeah. But I also would say I’d like to see where did Snowflake fit on this chart. Where does Data Bricks fit on this chart? Both of those solutions they may be in preview right now, but both of those solutions are building semantic layer things. I’d like I’d be very curious to see how those are going to be

25:05 Lid listed on the radar. I’m I’m actually surprised you’re right that data bricks is not even on there at all. , they’re not they’re not a semantics layer right now, but with their Unity catalog and their positioning around Unity catalog being the single secure object and having data being distributed from there, I see that as being a major point here as well. All right. Anyways, I think this might actually play very well, Tommy, into our next conversation, which is actually the main topic today. So, let’s What do you think you think it’s a good time to transition to the Yeah, I do. That’s perfect.

25:37 Okay. So, I’m going to kick it over to you, Tommy. Go ahead and read out the mail question of what we got today. Let’s unpack a little bit of what this is doing. I think this whole semantic layer designing things is this conversation. So, kick it away, Tommy. Oh, perfect. All right, here’s another great mailbag. Starting out with PowerBI. Our company’s reporting team has been using import mode exclusively and we have not been reusing models very commonly. There’s a lot of redundant data storage and I’m sure there are many

26:09 Conflicting DAX definitions and model relationships. I’ve started learning about fabric and I’m realizing we will need to rebuild some semantic models from the ground up. Thing is, I’m not sure our team will have the patience to do things right. I suspect they will dive in and start creating more redundancies and our work will continue to be fragmented. How can I kick things off on the right foot and influence the team culture to work together with a single source of truth rather than individual sources?

26:43 Oh, this is a great question. Good. Yeah, I I was thinking the same thing. What a great question. The first half of the question feels all technical like, hey, let’s talk about technical things. What can we do in a technical space to make this work? The second half of this is, I think, really the crux of the issue here, which is the culture. What does the culture of the team look like? And as you evolve and mature as a company using more PowerBI, how do you not only roll out PowerBI or continue to adopt and and develop PowerBI, but how do you do that while

27:15 Adjusting the culture as well? And this is interesting. Okay, so initial reactions, Tommy, what do you think about this? Woo. Well, one of my first things I thought was, , you’re learning this with fabric to have a single source of truth. And but then the next thing I was thinking was, okay, great opportunity. This is the something that we’ve been I think putting on the soundboard , pretty consistently with fabric is the ability just to quickly create things. But this is a I think a

27:49 More common than we think this situation where you’re already having the redundancies. If you already don’t have the right data culture or process and then you just get fabric well where do you actually start? I think this to me though leads this is a great great opportunity. I agree this I think this is a huge opportunity. So my initial my initial thought here is you’re starting to think about fabric and you’re doing lots of single imports. So these two points make a lot a

28:22 Really it sticks in my mind. You’re probably right. You’re probably importing the same stuff multiple times and you don’t have master tables of that data. So as I think about this solution and again I’ve gone through this transition multiple times with companies where you start with PowerBI and then you’re trying to say okay we’ve done it enough that it’s adding value. We actually need to be a bit more strategic with what’s going on here. So my first thought is think of the PowerBI file the semantic models you have today. You’re importing data inside that import

28:56 Process. You’re doing some transforms. You’re shaping the data the way you want. And then the semantic model is then building the relationships between those tables and the measures and calculations for the reports that you’re going to be building. So it’s like data loading step. Then there’s modeling step and then there’s servicing step. That same principle does not change. There’s no difference between that when you’re only in PowerBI or whether you move over to fabric. When you move into fabric, you can take a lot there’s a lot more tooling at your disposal to build bigger, more complex,

29:30 More centralized versions of those original tables that you had previously. And you can then direct link to them getting access to multiple semantic models. So this is an opportunity because if you look at the landscape and again I think this is a neat opportunity for them. Go get go go go after your analytics reporting. Go find where your data like go find the workspaces. Go collect the data on who’s using what reports. go use the scanner API. Go get a list of all the objects in your

30:02 Tenant. Figure out how many times each event has been occurring for a view of a report. You’ll immediately be able to say here’s the number of events by report that are being rendered. The more events you have on a particular report, the more important that report becomes. So the fact that you already have PowerBI established in your organization means you already have a huge wealth of data that you’re not probably looking at that can tell you what is effective for your business teams to use. I would start my effort there. And so look at the most used items

30:36 And then look at those elements and say, “Okay, how do I is is there redundancy in the top 15 items that are being used across my tenant?” Let’s look at that first because there’s I don’t think you can boil the ocean. You’re only one person unless you have really big leadership buyin, which is another one of my points here is if you’re trying to steer culture, make sure you go find that executive sponsor that’s trying to get this rebuild of the system or this this update of things. It’s very difficult for tail to wag the dog, right?

31:11 It it’s a bit more effective if you get leadership buyin that says look we’ve been finding success with PowerBI let’s now build a better data structure and they come down with the initiative from top down we’re going to start this new proof of concept we’re going to have a bit more rigor we’re going to establish a center of excellence and we’re going to take what we have existing we’re going to take the best knowledge of what we’ve built and rebuild these things in lie of fabric so we’re not doing so many imports we’re going to reuse data there’s some data principles here that I think need to be led from the that enables a business to start small and slowly drip out access to the rest

31:46 Of the organization. So, it’s interesting that you went that way with looking at what we have and trying to remove redundancies that way. I’m I [clears throat] almost am thinking even starting a little more basic here. So, sure because odds are they’re not just going to just turn on fabric for the team and say like I hope not off right off to the races. Yeah. I would not do that. I would not do that either. I would make a security group and in that security group I would put a couple people who are going to be experimenting with fabric and then I would say we need to figure out what our strategy is on

32:18 How to take in our existing stuff and start migrating the data part over to fabric and see if that’s even useful. It may not even be useful, right? You may not want to go down that route. You may just need to stay with import models and just stay there. Well, I I think for me this is where the the cultural change is in a philosophical change or conceptual change on what you’re building, right? So, this is a great opportunity to to put the lakehouse as now the the central area for our data storage. It’s

32:50 Like, hey, this is what we’re going to be doing. This is a change for us. , , what we’re going to be treating these lake houses as, , the single source truth whether or not they do the medallion approach or not. Regardless, to me, my initial strategy wouldn’t necessarily even be starting with, hey, let’s look at all the redundancies and we’ll go from there. Is making sure everyone understands the importance, the positioning, , and where the, , lakehouse lives in terms of business intelligence. Now

33:25 Let’s start putting in and building out our most fundamental data into some initial lakehouses and this is where you see it from the that redundancy idea. It’s like hey everything’s going to pull from from these lakeouses to build our semantic models yada yada yada. So there’s you also wanting to give the incentive for people just not to say that I’ll do it myself because sometimes with PowerBI it’s like well there that’s slow I need to do some additional transformations like no no no this is storage this is central

33:59 Repository and everything we’ll be connecting to will be from the lakehouse. So why don’t we actually start on our most common data and we’re going to start building out this flow of this new flow of data and I would start there so that way they can get a grasp on probably that 60% of data that’s in most semantic models and again this a start because honestly because you have to teach too you have to give people conceptual change on on what’s going on rather than just the technology change. It depends on what

34:33 You’re doing, Tommy, though. Like I I would agree like to some degree, yes. But I would also argue I would want to ask the question to these to the business team is who’s in your company building semantic models, right? Is are are we is this a central BI team that’s building like a handful of semantic models that have a lot of repeated imports on them that is then being distributed to the broader part of the organization? And so are are is the main consumption method semantic model to reports? like is it are they consuming semantic models? Are they consuming reports? Or is the

35:04 Business building their own tables and imported data? That’s what I’m [clears throat] thinking about here because when I look at this, this is going to be terms that I’m just going to generally use here, but I’m going to try to explain them as best I can. The way he explains the business that he’s using right now is the semantic models are acting like data marts. Not the data mart that’s getting deprecated but or yeah like a very small like knowledge center that is segmented and what happens is when you have lots of

35:37 Semantic models that are all using import you’re basically building little knowledge centers little islands of knowledge and there may be redundancy and conflicting calculations across those different islands or data marts when you start thinking about fabric you can stand back and say let’s try to unify this and act more like a central data warehousing and again I’m using the word data warehouse because that’s a traditional term and fabric has adopted a data warehouse we don’t need to use the fabric data warehouse but I think of it like a data library right there here’s all the tables and all the things that we care

36:09 About as an organization let’s just put that stuff somewhere where we can use it no matter where it is so if you move away from this mentality of we’re doing these very small solutions to a centralized solution solution. My hope here would be you can entice the business to use this new fabric or centralized data storage solution because they get more data. It’s cleaned a little bit better and it’s easier to get access and put into semantic models. Right? That’s that’s the main vibe here.

36:44 That’s the story with why why not have just a bunch of little islands? Well, that’s just disorganized. The the principle here is pull that stuff together, find common elements and when you see there’s a table that’s being used three, four, five times with just slightly different calculations. Now you can have the conversation around centralizing to the same thing. Was just in a client recently where business requirements were given to the semantic model layer and it was a whole bunch of things around date calendars. So date calendars can get quite comp complex and confusing.

37:17 The issue that always seems to come up is I want to compare this year’s sales to last year’s sales and then the question becomes if I’m doing the 13th through the 15th in sales comparison should I be comparing the 13th to the 15th of last year or should I be comparing the same period of time last year right so if there’s a leap year like if it’s a Monday through Thursday week 10 of this year should I be comparing week 10 of last year the same Monday through Thursday right

37:51 So you have to really explain what is the comparison and what we were finding with this company was one department aligned with this definition and said yeah this makes sense a different department said no we don’t want that we actually want to compare the 13th to the 13th all the way to the 16th of the 16th and we said that doesn’t make sense because your dates are shifting you could have more weekends or less weekends. And so there was there was a fundamental challenge a misalignment between two different business teams around the

38:25 Definition of things. And so as soon as you go to centralizing again speaking to the point here around we might need to rebuild some semantic models. We have a lot of potentially redundant calculations. We probably have a lot of tables that are being loaded the same way but maybe different transformations in them. That’s where you can address these challenges. And so the whole goal of this is to get everyone more on the same page and without someone at the top or someone making the decision to say okay we’re going to do the calculation of the 15th to the 16th that’s our comparison and that’s how we will think

38:58 Of it over time or we need to say no that’s not that’s not true we need to be doing same period of time last year comparison to make it equivalent and so I think some of those decisions need to be talked about because people don’t necessarily are they’re not thinking about that level of depth at all times and every team is not thinking about that depth. Centralizing this makes those conversations happen. I’ll just pause there. A lot of things I said. So, it’s interesting you you’re bringing up already the analytical queries here where if I’m dealing with the lakehouse at least, right? And so, you’re talking

39:30 Metrics definitions the really the business rules. But I I think for the the mailback questionnaire there’s more frustration not just on the DAX metrics which you’re I think you’re dead on there getting that clearcut that’s through the center of excellence but to me I’m going to go back to I don’t know if that affects the lakehouse methodology or approach at all because you still need to get that data in. you still need to have it cleaned up and

40:01 Trimmed. That me scenario that you were bringing up is DAX and is going to be in the reporting. Well, also well also in both these situations if you’re doing two different imports, one import can be using different M than the second import. Mhm. So even during the import mode, you could have the source could be the same, but the resulting table could slightly be different based on what’s being filtered out or something something simple like that could actually throw off the numbers substantially between two different imports. And to be honest, Tommy, it’s very difficult unless you’re

40:33 Looking at the M code line by line and doing checks at the end of the data for those two semantic models. It makes it more difficult to vet that the data is correct in both import modes. So, back to this question’s credit. You really shouldn’t be doing two imports. And I’m going to go here a bit more on the licensing part here too, Tommy. If you’re in a PowerBI Pro or premium per user, it doesn’t cost you any more to do two imports at all. It’s simple. You just do them. There’s no more there’s no added cost to you, the user, to import

41:07 Those two models twice. It’s true. This is true. on the fabric side, you now are in charge of managing your amount of compute you’re using and so you don’t want to do two imports inside fabric because now that’s double charging you for processing the same data. So even from a licensing standpoint, the the whole mentality shifts away from, , just just re just redo it, just reimpport it, just copy the model, do another import, change what you need to change, boom, done, not a problem. Because there’s no cost impact on that. And it’s that

41:40 Mentality breeds bad behavior. It breeds I want to say like not dissent, but what’s the word here? It’s it breeds complacency. like I’m not going to work hard to fix it and make it the right thing that we need, the one source of truth. It doesn’t cost me anything. So therefore, it’s meaningless. I can therefore I can just import it again twice. Not a big deal. So I think fabric really forces you to start thinking about smart use of your money capacity consumption. And in this situation, I don’t want to process that

42:14 Data twice. If I’m already writing the data down to the lakehouse in a vorder table, I don’t want to again cause the model to import that data yet again and then spend more money doing an import that’s redundant. But I again that’s why I think the the conceptual shift for this team is especially if they do have access to fabric is well we’re not doing import the same way anymore. We have direct link. Yes. and and it’s like that is going that is the default way. That is how we’re building thing. Correct.

42:46 Yes. import has been relegated if you were going to discourage import you thing and discourage in the sense of you still do it at some level but I’m thinking to myself like you want to encourage more direct lake the default and what you should be striving for is through the lakehouse through direct lake any really import just becomes outstanding or I don’t want to say niche because that’s too extreme but different circ circumstances otherwise that way

43:18 You don’t all the things you were talking about are not going to be as not I don’t want to say as alarming but this conceptual shift we are building lakehouses and then we’re building semantic models off of that our data is coming from the lakehouse so that and honestly that way a few workspaces you’re going to be able to see redundancy very quickly because unlike PowerBI where all my sources maybe I was using data flows but I would , it was so hard to see, especially even with the lineage view to

43:50 See, okay, this semantic model’s connected to this SQL database. Great. That SQL database has 1500 tables in it. So, I don’t know which table they’re using. I don’t, , that doesn’t everyone’s using that database. So, yes. Yes. , with this with the lineage now, it’s like, hey, notice that you’re building these marketing reports. What lakehouse did you build? Well, this is the marketing lakehouse. this is what everyone’s using to grab their basic storage data. you can transform if you want but it’s coming from here

44:24 And to me that’s where the focus is and then you start looking at migration and where the how many reports overlap because that at that point too Mike it’s also a audiencecontent thing too where I I literally just said this at at my session the last thing is and I always love this hot take where if you have a report and you that report you cannot think of two to three questions that report answers. It does not need to exist. It has no reason to exist.

44:56 That’s true. So, but that happens after we’ve set up the new flow of data, the new stream of data coming through fabric. That doesn’t happen during how do you handle how do you handle things you don’t know, Tommy, yet like in that in that scenario, right? , there’s there’s like polished publishing reports, right? This report doesn’t answer some specific questions and actually I think honestly there should be an overview page on a lot of these reports stating the purpose of this report why here here are some questions you may

45:29 Want to find answers to in this report I think that’s a really good example Tommy like I like that like that idea but to me that’s a polished built report when I’m sometimes doing internal reporting or just things that are for me right I’m somewhat doing reporting builds around just discovery I don’t understand what data is there I’m trying to figure out the statistics of this table. Like sometimes I have reporting that doesn’t really answer specific questions. I’m just doing a discovery, an exploration. So I think that’s a different type of reporting. I wouldn’t limit you in like I think there’s there’s definitely

46:00 Reasons for having both answering questions and real insights to do day-to-day work. But there’s also this idea of discovery and exploration. I don’t really come in with a specific question. I just come in with curiosity and want to dig around. So, I might take a little bit of issue with your comment there about some things. Okay, that’s I think that’s fair. But I think at the very least is this you’re starting with where is our data actually come from and it’s not coming from the lakehouse. Yes, I think that’s that is the fundamental shift that I see here in

46:32 This situation is you have to really be come to terms with we’re going to move away from import power query data flows gen 2 right we’re going to move away from that and we’re going to start saying where do we land the data here I think comes a good question that we had we were talking with Brad about data warehousing and SQL endpoints right that was a really good conversation on the podcast I think this really piques my interest a lot more around yes we’re saying lakehouse Tommy in this regard I think lakeouses are good for

47:04 Batch loading nightly loads lots of data storing all the information I also think there’s a there’s a strong story here to say once you’ve loaded the data multiple times or storing all the information inside the lakehouse maybe silver and gold could live in the data warehouse and moving more towards a structured table layer. I have found immense success with like loading files to the storage account that are like

47:36 JSON or API calls. If you make an API call, save the result of the API call to Lakehouse. I think it makes a lot of sense when you start getting into structure tables. I’m liking more and more the ability to use just SQL to do transformations, make tables, insert, update, merges. There’s there’s a lot of other really rich SQL level things that we can be using here inside fabric that I think make it much easier for us to consume things directly into PowerBI reports or semantic models for that matter. So

48:08 I I think I’m a bit more playing with more of the items now. I used to be totally to your point Tommy I used to be totally like everything lakehouse everything notebooks everything data flows gen 2 it has to get to the lakehouse as quickly as you can. I still think that’s a story, but I’m looking at more opportunities to find other places where I can start looking at, okay, well, maybe maybe we do put gold in a data warehouse. Maybe that does make sense. That’s what I’m thinking right now is I’m I’m blending more of the tools together because it’s so easy to use different tools for the different

48:41 Purposes. I I would completely agree. I think that’s part of the maturity and the journey, but would you Okay, let me ask you this then. when you in this this person’s shoes, you were just learning about fabric. You’ve been PowerBI. This is not Mike with the knowledge that he has now. Are you setting up or focusing on trying to get the a warehouse set up like in along with everything you’re doing with the lakehouse as well? Would that be one of your major initial focuses? I’m going to go back Tommy to what I

49:15 Said at the beginning of the podcast. , the very beginning here is let’s before I start jumping in and saying let’s go all in on Lakehouse, let’s go all in on these builds of things. I’m going to probably take a very detailed look at what is already out there and just really affirming these suspicions. And I think my opinion here, the suspicions this person writes out are probably exactly right. There’s importing the same tables multiple times. We’re looking at the same source multiple times. We’re doing slightly

49:48 Different logic in those imports to transform data multiple times. I think there’s a lot of redundancy that’s being built into the system that you just can’t observe. So I would take a really concerted effort to say let’s try to get a really good view of all the things that are being created. I would probably turn on fabric. I would use Semantic Link Labs. I would scrape all the workspaces. I would scrape all of the models. I would get all of the columns and definitions of the measures and column definitions.

50:20 And I would just put all that in one place. And then I would stand back and say, okay, now that we have semantic models and we have all the definitions of said semantic models, let’s go after some scanner APIs and let’s scan usage. So then let’s attach numbers back to semantic models. So I think I would need to say let’s just take a holistic look at our company in PowerBI. Here’s all the things we have. Here’s how many times they’re being used. That to me I think will indicate right away how bad

50:54 The problem is and how much redundant stuff we’re seeing. So that’s just my initial look at this is just to start there and say that’s where I’m going to start. I’m going to start exploring that stuff. Huh? So, we’re going to start with all the whole landscape and but I think you start with looking at everything that’s there. You don’t build it all. Right. So, this is just investigation is to understand what you have in your current solution because what I think you are going to find is there’s a like we’ve talked about this before Tommy like the pyramid

51:25 Of reporting, right? Oh, yeah. The bottom layer of the reporting is like minimal effort, minimal value, but there’s a lot of it. Like it’s a huge foundation. There’s going to be a wide there’s going to be tons of models, things that are not even refreshing anymore, stuff that’s no longer being updated. Like, why does it exist? Does it still need to be there? Yes, it imported a table a year ago, but it hasn’t refreshed in 6 months. Odds are that report’s not adding any value anymore to you. So, like those things are like, I’m looking for that stuff to say how much of this base layer

51:57 Of stuff is just trash that we don’t even need to look at. Then we start talking about layers that are up. And so we get further up the pyramid and as you look at again I’m going to Matthew Roach’s data warehousing modeling examples right he has this idea of when you have reports that have a large audience or very important audience that’s the s that’s where you should be putting your effort I don’t want to be spending a lot of time on reports that get just one or two views a month I want to be spending my time on reports that

52:30 Are getting hundreds of views or executives are looking at or people are making very like we’re hiring or firing people based on decisions that come out of this report. Like that’s an important report report. We got to spend time making sure that that thing is always working and has exactly the right numbers. So we’re not letting people go that we don’t need to and losing talent in our company. Talent is going to be so important inside our companies moving forward. We’re going to need to have people who know our business and we need to retain them and make them happy. So that to me, I’m looking to how do can I spend my time and serve the most amount

53:03 Of people, dude? I love it. And I guess I’m trying to look at one more thing in the mailbag in here. , and some of this too is just also about the people involved at the end of the day because honestly if you don’t have the right team like we we [clears throat] always talk a lot about the technology and obviously the new features but a lot of this too is your team culture is someone has to take a lead someone has to take initiative. Yeah. and and know what they’re doing or it doesn’t there’s no technology that’s

53:35 Going to be able to solve that, , human error, human behavior thing because that’s what you’re dealing with too is whichever I think whichever route someone goes is also get it written down, figure out a common process everyone can agree to and then make sure that you have a way to actually roll it out. Yeah, you’re you’re exactly right, Tommy. There’s going to be a huge not challenge. It’s going to be a huge effort for you to go back and forth between,

54:07 , making sure that you’ve got teams that know what they’re doing and then the data culture part of this and then actually then talking about the data. So the data part is part of the story. But to this last part of the question, I think you’re alluding to the very last statement. How do I kick things off on the right foot and influence teen culture? That’s the part that that I think I would like to give some, , final thoughts on this one. In order to impact team culture, you have to be leading by example. You have to get leadership buyin at the executive level. Someone at the top has

54:39 To be like this is a good thing for us to spend time and money on. We we will support it. there will be a bit of painfulness. There will be some friction and challenges. We’re going to have to do things differently. I really think the leadership needs to step in and say that to give confidence to the team that we’re going the right direction. So that’s something that has to happen. If you don’t already have a center of excellence where you’re describing what you’re doing, what you’re building, what’s next on the road map, I think that’s I think it’s a huge value for for communicating up the chain to

55:13 Leadership. Hey leadership, here’s what we’re going to be building. Here’s the first series of tasks that we think are going to be important to the business to build lakehouses of tables. This is our milestone. So having that stated and publicly available communicates to leadership, what you’re doing. On the way down, leadership down, it also communicates and gives confidence to people that hey, we’re looking at better ways of doing things. Here are the rules of engagement on how to get access to these new tables that are going to be more complete, longer series of time. it’s going to be easier to

55:47 Connect to right you start touting the values of why you’re doing these things but you go then you communicate that downwards so I can’t every company that I’ve stepped into and they ask me like hey can you build this thing I’m like sure but I always ask them for tell me where your center of excellence is and they some of them have them some of them don’t and so I request from them give me a confluence site give me a shareepoint site give me someplace where I can localize the information and knowledge that I’m producing into a common spot.

56:20 And every single company that I’ve done this with, they like what we’ve done. They continue on building the center of excellence. They take the examples that we’ve given them like definitions of common terms, training videos that we’ve done, how to connect to semantic models. Like the economy, the attention commodity commodity around what you can do with video now is amazing. Like you can you can literally record a meeting showing people how to use a report, publish it on the site. Best tip. It’s it’s easy. Like it’s there’s no like you have to do it anyways. Why not

56:54 Record that meeting and put it somewhere where people can see it and that way if someone has questions about it, you could just push a link to it. Hey, we just released this new report on remember we were talking about that overview page, Tommy. Here’s the questions this report could answer. Here’s some visuals you may want to look at. Oh, by the way, if you want to see Tommy running and building things in the report, click this link, watch the video, and here we go. The video’s there. Microsoft has made it so easy to create, edit, snip, create videos with ClipChamp, now built into your Microsoft 365. It’s almost

57:26 You want it or not. Y it’s so easy. and teams will provide summaries and teams premium will let you aggregate the data like all this really rich stuff that you can easily just put this content down and people can get value from it immediately. So that that’s the stuff I’m thinking about, Tommy. It’s like I’m thinking about, okay, the the culture starts with you. The culture starts with leadership. Leadership says, “We’re going to buy into it. We’re going to let you build this space for you to invest your time and effort to maintain and build documents, put up

57:59 Your procedures, how do you request a workspace, how do you request to get access to fabric?” Those are the things you start with and then you can build on top of that. And then the next time someone says, “Hey, you’ve got this wonderful [snorts] thing that you’re building.” Great. That’s so cool. Where do I get information on getting fabric? Oh, right here. We already have a site built for you with some general information how to get access. Ask your manager, fill out this form, do these things, right? Here’s the process we’re going to follow. Clear as day. So, that stuff, I think, instills culture

58:31 And starts shifting the culture in the right direction. Dude, I I I love that. And even with the video thing, I remember you saying that years ago that if you ever had to tell someone more than twice if they asked a question, you made a fiveminute video. Yep. Those became one of the most popular things that we had. But honestly, man, it is right. Like this is just going back to who’s going to take an initiative again. And you and I were lucky enough to be those people that were like, what, no one’s asking for a champion, but I am going to be that champion. and just took

59:05 The took the lead. , and this is going to require it too because there’s just more more places and more, , features, things to build than ever. And just having an an approach, having a buy in that, that’s where this is going to be solved or lost. One, again, I’ll just pick on here Microsoft a little bit. They have this thing called purview and they have this thing called one lake catalog. , one of the things I think that Microsoft has done a little bit poorly compared to

59:37 Like other companies like Data Bricks, Data Bricks has reinvisioned this whole Unity catalog experience of hey, there’s there’s a discoverability thing. There’s a central place that has all the management. Microsoft does the same thing. It just managed slightly different, right? So, it’s not perview. It could be one link catalog, but what you have access to just automatically shows up in these cataloging areas. I think Microsoft could get a little bit better on adding more metadata to tables. I’d really like to see something integrated with one link catalog where we have like column names and descriptions and primary keys and

60:10 Foreign keys. There’s there’s some missing metadata that I think we need to have in the one link that would make that story a little bit more compelling. But at the end of the day, I think fabric is a really useful tool to you. Your company needs to figure out is this worth the extra spend because it will cost you more to put in fabric than staying with the current pattern that you have today currently. So you have to really evaluate as you’re rethinking this solution are the advantages you’re going to get for adding real time lakehouse common storage of information

60:44 Pipelines that reuse faster load times right get integration like th you have to really wrestle with those features and say are those features valuable enough to the organization for us to move away from pure PowerBI pro licensing and now additionally fabric as Well, and if you’re if you’re a bigger company, I think it does make sense because you get fabric at an F64 level, you can start sending free users to powerbay.com to go use the fabric environments. So, an F64 is like an

61:16 Unlock moment if you get to that scale. So, if you’re already in a big enough scale where you’re going to have to go get an F64, I think fabric is a really good opportunity to look at it because then you could remove a lot of PowerBI Pro licenses and stick with the F64. So, you’ve really got to do an evaluation like what’s the size of your company. Larger size companies, I think they’re going to definitely want to go with fabric and those larger end SKUs and then turn off those PowerBI Pro users and distribute content directly through the fabric licensing. Again, that’s another whole podcast in and of itself. Can of worms. Yeah, can anyways final thoughts here. Start with

61:49 The data culture. Start with executive leadership. Make sure you have buy in there. Focus on those areas first because once you have culture moving in the right direction, fabric and distribution and data handoff, I think will become much easier for your organization. You’ll have a much better story to spend the time and the effort to centralize and make this stuff more consistent. Tommy, any final thoughts for you? I think fabric’s a great opportunity especially in this scenario to think differently about how data has been working and almost start fresh but it’s a conceptual change where

62:22 Again what it what do we hold most dear where does data flow from it used to be the semantic model semantic model still absolutely essential but it allows people to finally think differently and but that has to come through education training and just plain old marketing so I would agree with There’s a lot of marketing that goes along with this new world. You want people to use the things that you’re going to be producing. That being said, thank you all very much for taking the time today. This was a great mailbag question. We really want to encourage you to use the mailbag. We have a link in the description here. If

62:55 You have a question, please use that link. Go to the website, fill out the form, ask your question. We would love to unpack your ideas and your questions as you’re going through migrations or PowerBI challenges or fabric challenges. We we like to unpack and think through these things. Most challenges Tommy and I probably have already seen. So, [laughter] you’re probably going to get at least an opinion about one of those challenges you’re facing. So, we’d love you for you to engage in that way and fill out the form. We know your time is valuable. Please make sure you like and subscribe to our channel if you want to hear more of this. Tommy, where else can you find the

63:27 Podcast? You can find us on Apple, Spotify, wherever your podcast. Make sure to subscribe and leave a rating. It helps us out a ton. Do you have a question, idea, or a topic that you want us to talk about on a future episode? Head over to PowerBIt tipsodcast, leave your name and a great question. And finally, join us live every Tuesday and Thursday, 7:30 a.m. Central, and join the conversation on all of PowerBI tips social media channels. Thank you all so much, and we’ll see you next time.

Thank You

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

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

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