Microsoft Fabric & Tableau – Ep. 271

If you’re in a mixed-tool reality (Tableau in one corner, Power BI/Fabric in another), the question isn’t “can I connect them?”—it’s what problem am I solving by connecting them.

Ep. 271 digs into a pattern that’s suddenly showing up in real projects: treat the Power BI/Fabric semantic model as the central definition of measures + logic, then let other tools (including Tableau) query it when that’s the fastest path to value. The catch: that “simple” connection has licensing, performance, and governance implications you should understand before you roll it out broadly.

News & Announcements

• Connect Tableau to Power BI — Data Goblins’ walkthrough of connecting Tableau to a Power BI semantic model (including prerequisites like endpoint access and drivers) plus a good checklist of considerations before you standardize on it.
• Semantic Link overview — Background on Semantic Link and how Fabric + Python can interact with semantic models.
• Submit a topic idea (Explicit Measures Podcast) — Send a scenario or question for a future episode.
• PowerBI.tips Podcast — Subscribe and browse the full back catalog.
• Power BI Theme Generator (Tips+) — Generate and manage report themes so formatting stays consistent across your org.

Main Discussion

The headline idea: a Power BI/Fabric semantic model is increasingly becoming a shared “business logic layer”—not just something that exists to serve one Power BI report.

So if one part of your company is fluent in Tableau, and another is standardizing on Power BI/Fabric, it’s tempting to ask: What if Tableau could read from the same centrally governed semantic model?

The episode walks through the mechanics (and the costs) behind that idea. In practice, Tableau-to-semantic-model scenarios often require XMLA endpoint access (capacity/licensing considerations) and/or connecting to a SQL endpoint depending on what you’re actually trying to query.

Key takeaways:

• Know the connection path you’re choosing. XMLA endpoints vs. SQL endpoints change the user experience, the performance profile, and what parts of your model/logic you’re actually reusing.
• Budget for licensing/capacity, not just “a connector.” The moment you need XMLA, you’re in capacity territory—so run the math (including how many people still need Tableau licenses).
• A shared semantic layer is the real win. Centralizing measures/definitions reduces business logic hiding in report files and improves governance/auditability.
• Mixed BI stacks aren’t “free optionality.” You pay in enablement, admin overhead, and inconsistent user experience (often long before you see a clean ROI).
• This shines as a migration bridge. If one team is deeply invested in Tableau, this pattern can buy time while you standardize definitions and gradually rebuild experiences.
• Start with the business problem. Don’t sell this as “how to migrate Tableau” by default—anchor it to the pain you’re solving (data sprawl, inconsistent definitions, time-to-deliver, or organizational politics).
• Fabric changes the posture. When the platform includes pipelines/notebooks/lake + semantic models, Power BI reports become one consumer among many—so integration questions become inevitable.

Looking Forward

As Fabric matures, expect more “mix and match” scenarios—but the teams that stay sane will standardize definitions in the semantic layer first and treat visualization tools as interchangeable consumers.

Episode Transcript

0:28 good morning everyone welcome back to the explicit measures podcast with Tommy Seth and Mike hello good morning good morning and a happy Tuesday gentlemen for Tuesday never never forgets so Tommy was just telling us about some riveting things he was his his morning routine it is that time of fall what are we gonna talk about first let’s let’s let’s go through our main topic before we get into our news let’s go into main topic for today thank you Seth for keeping me on track our main topic for today is connecting Tableau

0:58 topic for today is connecting Tableau into Microsoft or particularly the semantic models is a is a new thing that’s Microsoft is rebranding the data sets but we’re going to talk about where does Tableau play in this game does it make sense or are there other tools or other ways you’d like to connect into your semantic model in general so we’ll probably open the conversation up more than just Tableau but we’ll start with Tableau initially and then we’ll go on from there let’s transition into some news jumping into some news right away Seth you found an interesting article on

1:29 you found an interesting article on gateways that was recently published what what is this new feature you found and what should we know about here from this I I did indeed and and I’m I’m one of those guys that clicks just cuz right because everybody I I think of all of the blog articles that get posted like new release of Gateway is probably not top of the list 100% however however this time this time rewarded rewarded with a little nugget of wisdom the new new desktop

2:00 nugget of wisdom the new new desktop Gateway supports supports the new query quering of the November desktop which is good but the larger announcement within there is now you can add a new tenant registration allow list setting and it’s mostly on the machine that you’re installing the Gateway but it is something that hasn’t been there before so if you’re not familiar you can literally put in a UPN or your email from any

2:30 from any onto a Gateway and just plug into it which has always been fantastically easy for for those of you who are popping around and and doing different things I I I don’t I don’t know if they were trying to like throw this into the r but it’s like I always love where we’re seven years into a product and part of the part of the article is please be advised that this is a good security measure to take but does not guarantee total data exfiltration but it’s a good security measure we just haven’t had it

3:01 security measure we just haven’t had it until now and that’s concerning it it is what it is it’s like you’re not going to get access to a virtual machine without the appropriate authorizations etc etc so it’s another layer in here that it’s another layer but because before it all felt like Voodoo and Magic before because you would literally log in with a UPN a user email address on the Gateway and be like hey I’m I’m going to log into this Gateway and they’re great like great well auth you cool then you go into power. com the service and it’s

3:32 go into power. com the service and it’s like oh look it just magically appears like how does it how does it know like what is it doing it’ll never know oh it’ll know it’ll know yeah I never I never thought that you could log into your tenants or domains like virtual machine and then log in with a completely different tenant yes and be if you gave me access to your machine I could I could say Seth Seth power a tips or whatever whatever yep and even though it’s in the my and be like cool go out

4:02 it’s in the my and be like cool go out to my service and now I have access to to that Gateway so like I said it’s not a huge hole because you’d still have to have permissions to all the data sources underneath Etc but good and bad the one thing I’m confused about this article and I’m I’m skimming through it now as you as you mentioned it here where is the where is the tenant registration allow lless allow list setting is that inside the powerbi portal no on the machine in the H key

4:32 portal no on the machine in the H key the registry in the registry yeah I saw that I got I’m looking at going wow that’s why it’s like it’s like well you already need admin you need you need to be a virtual Mach like a box admin to be able to get into the registry so it it’s not like typically if you’re if you’ve got VM set up you’re not going to have access to the the registry so this is this is a so let me just make sure I understand the feature here just to get my head around it this is on the machine that you are

5:02 is on the machine that you are installing the Gateway into you’re saying inside the registry as an admin I’m saying this machine this VM can only talk to these list of tenants correct so this would be one of those things that you’d want to pass up to your it teams yes and say when you spool up a virtual machine for a powerbi Gateway or if you’re going to apply a powerbi gateway to a machine you would want to ensure that only your 10 IDs are in the registry so that as anybody’s setting

5:34 registry so that as anybody’s setting things up Etc just be part of your I think BM standup process into the registry strange oh I was gonna say it’s just strange because I’m surprised they didn’t do it in the config there’s a hidden file with a like probably 50 other set or at least 20 settings that you can change in the Gateway and it’s like an XML I think it’s like config Gateway I’m surprised they didn’t put it there I’m sure there’s a reason but it’s it’s crazy that to protect your tenant you have to go into the

6:04 your tenant you have to go into the registry well you would like that doesn’t I guess it doesn’t make sense to me because like the the Gateway is a service that you’re applying to the VM right so in order to restrict the service from doing something yeah you need to tell it can’t do that yeah you don’t want to go you don’t want to go into the service itself and say this is the only tenants I want you to talk to because then it’s like that defeats the purpose right to some degree it also It also says here to if you for one thing I was also worried about was well what if there is legitimately your company has multi- tenants for powerbi for whatever

6:35 multi- tenants for powerbi for whatever reason multiple yeah there is a comma separated goid list of tenant IDs you could use so you could have multiple tenants in this property you just use a comma to separate them and so you can have multiples too interesting great find on that one Seth will always find the security loopholes on everything that’s like his Seth is like security security Seth there you go that’s like your superpower is it is it just one that’s one of them one of

7:05 one of them one of them Walter White is your other superpower yeah yeah thank just the way you look Z just the way you look that’s your superpower your face is a superpower and M and Minecraft oh all right let’s move on to our main session today the main topic for today is talking around Microsoft Fabric in part it could be fabric could be powerbi what are we talking about here and we’re going an article written by data goblins Kurt is just amazing runs all these great articles super well written and I really

7:37 articles super well written and I really liked particularly this article as I was reading through it and the link is in the description for those who are going to try and follow along or want to read the article this is a really interesting article and I also like one Kurt you’re doing a great job of there is a lot of content in the article and he uses the accordion feature of his articles in a very elegant way I was like this is very well done cuz sometimes you read a Blog and you’re like oh my gosh I’m scrolling forever I don’t know what information is important here and there are certain things like he really does lean into the

8:09 things like he really does lean into the user reading the article and saying okay I’m going to tell you all the details but for those of you who don’t need to know the accordion steps here I’m going to give you like the too long didn’t read tldr of it and say Here’s the things you need to set up set up this this this and this I as an expert in PBI already understand those things so I just glaze over those four brittle points but they’re fully expandable windows that you could say well if you didn’t understand any one of those steps you could accordi in it out and see how the details which again very well

8:40 the details which again very well written really like your thought Kurt the the blog post is just incredible could agree more it was one of the things that stood out to me as well yeah and maybe he’s been doing this all along or refining the process as he goes but I feel like really it was very noticeable in this one for some reason right agreed and and yeah definitely thinking about the end user how how to compact or throw in all

9:05 user how how to compact or throw in all of the details without cheero mic just being this long scroll list of like oh man where’s the where’s the heavy stuff but yeah Kurt Kurt’s just fantastic there used to be an external tool I think it’s actually in business Ops of connecting Tableau to powerbi correct there is an external tool that would and that’s that’s more around business Ops I think was connecting there’s a external tool that this basically handles a I think it’s a a small file that’s created like a connection file that’s made for Tableau so that is used

9:36 that’s made for Tableau so that is used for I have a data model I have a semantic model locally on my computer it’s open in parb desktop I click that button which then means Tableau can look at the cube that is local on your machine and can read it and then automatically you can build visuals on top of Tableau locally on your machine so let’s get into the article let’s maybe talk about first what does this mean let’s what is what are we talking about here when we’re talking about Tableau connecting to the semantic model in

10:06 model in powerbi powerbi I’ll start off on this one just kind I’ll start off on this one just just very quickly to give a of just very quickly to give a summary of a bit of the article and then I think there are a couple key points here to pull out here that hurt I think really illuminated very well the the concept of the semantic model is this is your relationship between your tables your Dax measures and your calculations it’s a cube that you can use and the neat part about the cube is it’s General you can use it anywhere it could be used for Excel it could be used for powerbi reports and in this case what the the setup is you publish a semantic

10:38 the the setup is you publish a semantic model to the service you then open up xmla endpoints so you have to be on something that is premium so there’s a premium feature to what this is looking like on how to connect to that Cube once you have that premium feature open you have to install some drivers on your machine I think he said o OA OA DB or something like that some drivers that were required there but once the olb drivers were installed you’re then able to connect Tableau directly to the cube and then you get your full data

11:10 cube and then you get your full data model as if you had built it so the idea is there’s one story of the truth there’s one Cube that has a central definition of measures and things you want there and then you can build reports off of that so that I think is the setup for that initial thoughts gentlemen around connecting tableau to powerbi or fabric I think this goes back to some of our conversations on what fabric opens up and more importantly that the data

11:40 up and more importantly that the data the semantic model and the powerbi report is no longer the end poost of what we do everything we’ve done up until Fabric and especially the Lakehouse Lakehouse powerbi was really the the last stop it was the there’s there’s no Outlet so to speak of what we’re doing what our Transformations were you’re either building you have to build this in powerbi it’s shown in powerbi and the only extensions the only turns were analyzing Excel or PowerPoint but really

12:11 analyzing Excel or PowerPoint but really that model really existed as the last stop what we’re seeing more and more and this is just another great example of it that because of the semantic model because of the semantic link because of what Fabric’s enabling for the business and and I’ll go back to what you see Mike it’s only it’s becoming just one now has just another route where we have endless possibilities of what we can do with a transformed structured

12:42 can do with a transformed structured model that we create in the same fashion which I don’t know if we know the potential of that yet I don’t know if we’ve we’ve really discovered all that’s going to do for the business and how that’s going to affect governance and how that’s going to affect users and how that’s going to affect the roles at a business this is that’s a good point I I do think there’s going to be some time here for people to understand well for one this is a a whole bunch of new features that just been thrown at business users it’s

13:12 just been thrown at business users it’s going to take them years to figure out what’s the best practices for their organization what’s the best way to push and pull through fabric I’m hearing a lot of value from many organizations that want to get people into fabric much sooner than they would be previously so this whole idea of direct Lake pieces I think there’s a lot of pieces there but particularly talking around this Tableau and powerbi integration I I think it’s interesting to me to list to think about the licensing that’s required in order

13:42 the licensing that’s required in order to get Tableau to bolt on to powerbi Microsoft has clearly been making it easy for a pro user to build reports on top of the cube as soon as you’re adding this other tooling you need to get after the xmla endpoint which cool but if you think about it that’s already a price barrier to a lot of organizations because other organizations may have a mix of pro users and some Tableau users mixed together well this instantly pushes you towards a premium capacity

14:14 pushes you towards a premium capacity however I will say this if your model this is where it’s a game that we play here because in Kurt’s section about halfway down he starts talking about how what Licensing do you need you need a powerbi pro license plus some Premium capacity or just a fabric capacity which goes all the way down to a pricing of an F2 so you can go down to $280 per month if you buy it from Azure then there is this concept of a premium per user so you could be a $20 a month per

14:44 a month per user workspace plus a user that has a premium per user license so the workspace so this this is where Microsoft gets really confusing I get a lot of questions around how does the powerbi licensing work and there’s like two layers of of Licensing there’s licensing at the user level and there’s licensing at the workspace level or the capacity level and I think this is very confusing to people especially when you’re talking about stuff like this any so so we so we bring it into the conversation just to confuse you f

15:15 the conversation just to confuse you f no I think it’s something worth talking about it is it is I I guess I have a question wasn’t the semantic model already available to Tableau what do you mean by that previously like so so there there’s two ways in which Kurt’s talking about connecting Tableau to semantic model one is well actually two ways of connecting to data in the Microsoft ecosystem one of those is thematic model one is fabric right which is really just a SQL

15:47 right which is really just a SQL endpoint for I see you’re I see you’re saying so you’re saying also okay so there’s also another SEL I was got it I was under the impression that the semantic model was already available for TBL TBL it it is it’s not this is not a net new thing this is just an article around like this is something that’s been existing for a while the context of the conversation threw me into a loop of like is this new no not new just something we’ve never really talked about yeah what strikes me as as a really like an organizational

16:19 really like an organizational move by Microsoft but also a brilliant one from the how it’s represented here is not being a a a a combatant combatant competitor with other companies and what that’s done is open the door to these really unique Partnerships even Partnerships to some degree across the technology ecosystem that they they play in and created really unique I think solutions for the larger populace

16:50 think solutions for the larger populace and and this is one of those things to me because if you think about a lot of the conversations Mike when you and I were Consulting together a long time ago that people had were it was hard hard lines of Tableau and versus Power VI and and I I don’t I’m not saying that that’s gone but I certainly the vehement between those two things is like lessened over time I think yeah to because I I think the conversation points that we kept hearing all the time were it’s you’re the business has to

17:21 were it’s you’re the business has to choose between one reporting platform and another if you were going to store your data and do it in one way like you you were stuck in that ecosystem completely and you couldn’t do anything and what I what I think is the brilliant move is is even with fabric and even I guess prior to that even with an advent of powerbi and the semantic model being the powerbi thing opening that up to a different reporting tool as opposed to just saying no man we’re not going to let you connect to our like our engine

17:53 let you connect to our like our engine this is what separates powerbi and it makes it what it is and we’re not going to let your visualization to no they open that door and now I think help facilitate for businesses when they pick and choose where their data sits what

18:07 and choose where their data sits what reporting tool they’re going to use like you’re giving them options and that’s always a good thing always a good thing from from the consumer perspective I think it’s new for sure and maybe it’s a a an the Advent of SAA in Microsoft but certainly I welcome it and I love the fact that Microsoft has adopted that in in the analytics space as well in opening up like yeah you can do whatever you want with our ecosystem like Plug and Play what works best for you if you

18:38 and Play what works best for you if you want to run both these tools at the same time you can like all of it revolves around the data itself and we’re not going to restrict you from accessing the data in any way so the I I if I if I’m strategic around Microsoft right if we’re thinking strategic about things like okay fine the output layer should be very be very generalized the the inner layer here of like the data whoever owns the like this is one thing that I’ve found very core especially even when you’re building applications or web apps whoever owns

19:08 applications or web apps whoever owns the data is King of the Hill you want if there’s there there’s something you want to own okay fine the visualization tool can change you can make different kinds of visual tools cool but if you own the cube where all the data lives if you own the design experience around how to create the things that you need in that initial area area that that changes things a lot like like that really puts the power in Microsoft’s shop go ahead can I can I let let me not I was going to say challenge but I

19:39 not I was going to say challenge but I don’t think it’s a challenge I I think I would tweak that statement okay I don’t think well the The Tweak would be do we I I agree with the sentiment but is it actually who owns the data or is it actually who owns the business logic that transforms it data because what we’ve seen here even with fabric is we don’t care where your Delta table lives we we care how people are are are

20:10 lives we we care how people are are are managing that data and where where all of that that logic or where the pipelines are going maybe it’s maybe it’s the whose computer you using I was gonna say who controls the flow of the DAT yeah yeah I I think yeah I think I think I I would say maybe my view of that word that statement was too narrow and what you’re thinking about now aligns very well with where I was going with that like it’s it’s more important for you to say I own the

20:40 more important for you to say I own the business logic I own the tooling I own the cube where all the stuff lives whether it’s a lot of business logic or a little bit or just tables and relationships I don’t know but whoever owns that becomes like the expert in that space or or strategically you become way more embedded in that area than area than okay what if Microsoft says well screw this this this parbi reporting side thing doesn’t really make much difference let’s go back to the Excel team and just invest a lot more in there I literally heard of of a I

21:11 there I literally heard of of a I literally heard a interview I was listening to some webinars recently and I think AO had made a comment inside the webinar that said there is over a billion users using the Microsoft M365 product Suite so this is word and PowerPoint you’re talking about a billion user opportunity this is why they’re so heavily bent on let’s get cubes and models and powerbi in the hands of those users so it is in their

21:42 hands of those users so it is in their best advantage that there’s already I don’t know how many hundreds of millions of users inside Excel now because that’s part of the office you don’t buy just Excel you buy the whole Suite it’s a it’s a package deal yep so when you get it now you get this powerbi thing in there as well and comes out for free this is a huge opportunity for them to continue to grow this thing and now there’s going to be a lot more cubes rolling around I I I think I need to interject because I think you’re both right because this a very interesting topic like that who controls the

22:14 topic like that who controls the data it I think you’re completely right but I think it’s the context of it doesn’t have to be a database it’s the data that people trust because it could very well could be Excel where all the data is the source of Truth and guess what those people control they really do control a lot even if you have everything set up in a structured database if the team if a business is relying on this Excel file yep or this flow of the data that’s where people are going to go so I I think you’re right

22:44 going to go so I I think you’re right but I don’t think it’s necessarily always having to be in that structured pipeline because and I think that’s kind pipeline because and I think that’s been our goal too is we’re enabling of been our goal too is we’re enabling the users we’re enabling the business to have more control of it but at the same time we’re trying to build up that why do we always say migrate away from Excel for database or for analytics it’s not because we just don’t like Excel we loved Excel but we’re trying to create some guidelines around we’re trying to create some structure around it but there’s a lot of organizations

23:15 it but there’s a lot of organizations still to this day you had the conversation a few weeks ago about JP Morgan and about they’re using Excel and guess what their business is relying on Excel so the tooling and is just more about where the source of the truth is at that point yeah I of the truth is at that point yeah to your point though like and the mean to your point though like and the examples you’re giving are are the automation that those tools allow you to do right and that’s the important part and why we lean into businesses and say yeah you’re functioning nobody’s

23:47 yeah you’re functioning nobody’s questioning the numbers that you’re putting together what we’re questioning is why would you want to be doing that manually right like it’s opening it to error right so set the things it work that way every single time data comes through and I think like you’re you’re going to have decision points right like even within the Microsoft ecosystem you have multiple different ways in which to create ETL and El pipelines and all these different things so like I I agree that like the the data ownership the data movement having data accessible in

24:19 data movement having data accessible in many ways outside of powerbi is extremely important for a data culture as well it’s just serving up those that information to to users in many different ways as easily accessible as possible so I think that that leads to another question or a really good question we talked about the downgrade of the powerbi tool as just another Cog in the fabric machine it’s just one of it’s pipelines it’s notebooks it’s powerbi reports do we want to then if we’re utilizing fabric

24:50 want to then if we’re utilizing fabric and fabric is again now just another Avenue for all these other ways we can use the data sure are we trying do we want to en courage then less I don’t want to say less use of powerbi but more use of other tooling or other Avenue of reporting this is a great question and this is one that was addressed inside Kurt’s article that I think we should talk about is because on so the question is should we should we look about other tools or other tools that may serve some purposes here I think

25:21 may serve some purposes here I think it’s for me it’s a lot about use the right tool for the right situation kind right tool for the right situation thing and one of the notes he makes of thing and one of the notes he makes here is is just be aware if you’re trying to mix Tableau and powerbi together based on what you’re trying to do there is some additional licensing costs because now you’re not just paying for just powerbi you’re now paying for powerbi and another tool so organiz like I’m looking I’m literally Googling like what other things could connect to a cube you could go to click click also has a program that goes right into the

25:52 has a program that goes right into the cube so you could have a fabric Cube where click connects to it or and something I didn’t bring up earlier that Seth you brought up was I was thinking about connecting to the XML endpoint the XL mpoint is the the analysis Services partition of fabric there is a SQL endpoint as well so you could also connect Tau directly to the SQL endpoint and say I’m going to go access just the SQL Server portion of it and go request data directly from there all of those implications change how much compute is

26:24 implications change how much compute is used it changes their pricing on because now I’m paying RBI and potentially I’m paying for something else and the other the other thing I think is a weakness here is now when you start relying on other tools you’re now beholden to how efficient is that other tool in adding business rules and or logic around what’s happening from that semantic model so I think there’s risks there from a cost perspective yeah I I don’t

26:54 from a cost perspective yeah I I don’t know if from a reporting perspective it makes sense to me to have both both powerbi and Tableau running at the same time because to your to your point Mike licensing there’s also a knowledge Factor like but your teams would have to know one well unless it’s like Tableau

27:10 know one well unless it’s like Tableau in one part of the business and powerbi in the other the challenge with that is you’re change you’re mixing up the user experience depending on where people are interacting with things and you’re you also have to deal with multiple infrastructures right multiple different ways in which to manage or which you ways in which to manage or which how to deploy things so know how to deploy things so there it’s not without cost from a lot of different levels if you’re mixing in where I think it’s valuable is if you’re in one and like you you pick it allows

27:40 in one and like you you pick it allows you to choose which reporting one you want I guess is what I’m saying or transition from Tableau into powerbi and there’s a bit of reality here too as as cool as this is there’s no way you could roll this out to an Enterprise this this workflow with fabric with without to your point incurring a lot of cost and a lot of fin angling the model there’s a lot of things that also don’t cross over like calculation groups I’m I’m not gonna I’m not gonna say no to your answer Tommy I think I

28:10 say no to your answer Tommy I think I agree with you it’s at the core however I think Kurt gets some really good examples of like why you would want to mix the two and and just some considerations around what would be the thought process around why we want to mix things right I think he gives some good examples here inside inside here talks about a little bit of you have a team like you have multiple teams in the company one portion of the company has just to your point SE earlier the knowledge base is all built around Tableau it’s a tableau thing so since

28:41 Tableau it’s a tableau thing so since their knowledge base is all around Tableau already instead of throwing away all their existing work and you can say look you guys keep your tool we’d like you to migrate onto powerbi but in the meantime we will build these QB things that the rest of the organization is going to use you are more than welcome to continue to connect to them and use them so I think there’s there’s this idea of like you’re in the process of migration and you are not quite fully ready to turn everything on for powerbi because it does take some time Tableau is a really robust visualization tool it’s not a good data engineering tool

29:12 it’s not a good data engineering tool and I also would agree it’s not a really good place to store business logic so that’s something that’s as a as an admin of powerbi or an admin of just di in general if there’s business logic being incorporated into those Tableau files I’m notare again this is probably me just being naive I’m not aware of how would we scrape through all those Tableau files and extract out all the measures or things that we’re doing that are business-based logic pieces that are inside the Tableau file itself this is a risk however I feel much more confident

29:43 risk however I feel much more confident now with what powerbi is coming out today I can go rip through all the pbip files that we make and I can go find all the measures that were created in the thin reports so I have a I think I feel like I have a better reporting back to the organization surface area of all the things that come out of powerbi where are they being used I have better metrics around that because it’s all in the same ecosystem so I think I lose as an admin some visibility when I let Tableau run there Kurt also brings up the point of you have you’re in the middle of you’re in

30:14 have you’re in the middle of you’re in the middle of a migration and you’re building net new table ports because you just don’t have time to rebuild them in inside parbi so there’s like a the stop Gap appearance that going in there as well would you agree with those reasonings Seth is there anything else I missed there that you would see why you’d want to have Tableau existing versus just all powerbi well I I guess no I don’t have anything to add on that I guess the the initial reaction I had with Tommy’s statement though is I Tommy you said why would you deploy

30:45 would you deploy fabric why would you I don’t think this is realistic if you’re trying to scale and say okay all of our Tableau or the majority of what we’re doing in Tableau we’re now going to point to fabric I think it works in circumstances I don’t think it works from a we can scale this up for all of our current Tableau that we’re trying to migrate well and and because you’re right though you’re you have a point there’s a lot of organizations that are using powerbi it’s but Tableau has this method of like

31:17 it’s but Tableau has this method of like land and land and spread they they hit a team they make that team really happy they provide a lot of support to that one person or that one team and their goal is once you land in one person the goal is everyone’s going to love this tool continue spreading it across the organization so it’s it’s that is their approach to get it in and move it Microsoft’s approach is we’re just going to give it away for free with desktop to everyone and because of that people like well I’ll just figure it out and just go from there so I think there’s

31:47 go from there so I think there’s two different strategies in how both companies are going to Market yeah and I don’t know like the the how do you migrate question is is a harder one because I’m not 100% familiar with like all of the tools around Tableau but like alter like all the the the the push back I would have there is it depends on the data sources and all of the stuff that you’re doing to curate data for Tableau reports because it there’s a lot of it that has to happen because it’s not a model driven front end right there’s a lot of

32:19 driven front end right there’s a lot of things you can do in power VII and in the ecosystem so when we talk about fabric it’s much more than just reporting it’s the ENT Enterprise tooling that just enhances what you could already do from an ETL perspective and modeling perspective in powerbi and because of that a lot of those connection points like Kurt’s article points out in the second half are SQL based so if like that’s your basis for a lot of the Tableau reports could you could you make an argument that you would be simplifying your life if you

32:50 would be simplifying your life if you went the fabric route because all those services are combined like that’s the whole pitch of fabric we’re taking all of even in the Microsoft ecosystem we’re taking all of these services that were these independent things that you had to connect to get permissions have access to Etc and we’re combining them together where everybody can play you you have that and more in in Tableau implementations because it’s third party third party third party

33:20 third party third party third party third party and they have these you’re just creating your own web based on whatever person is saying like oh yep this one this is the tool we want right and then you plug it in oh this is the tool we want and then you plug it in so it depends on like the sources of information but if I’m a company and I’m looking at like my data is going to scale and I need to grow anyway could I choose fabric as a backbone for that and then have options in the in the future sounds pretty compelling to me

33:51 future sounds pretty compelling to me you just said cannot be understated I I’m working with a client right now where it’s Tableau but it’s also alrex for one team or for another department this where I was going to go we thought we had the wild west at least it was just powerbi yes right but how do you how do you report on that like to me is like Tableau isn’t just a standalone product that does everything it does like it’s purely the visualization side of things and I think par gets a bad rap because it gets lumped into this whole like oh it’s just a visualization tool no it’s not it’s it’s a semantic

34:22 no it’s not it’s it’s a semantic modeling engine and a visualization tool which you don’t even get with like tab you have to go buy other features of Tableau or you buy an altrix or you buy something else third party and this is what I find with a lot of these other tools a lot of tools don’t have the entire story they have portions of the story that they’ve done really well with but not all of it and also if you try and bed so one of the things we’re talking about here is this is just buying you even just to get desktop for for Tableau you got to pay for it the desktop version is a paid feature and if

34:53 desktop version is a paid feature and if you want all the connectors you pay more so it’s not a cheap program to begin with just to get it on your PC in addition to that when you want to send those reports somewhere Tableau server that sucker is expensive I was just talking with someone looking at getting Tableau server and embedding reports for an external application they were talking $3 million per year to be able just to run Tableau for an embedded application that needed lots of lots of users but you’re talking scale up things like you’re buying machines you’re putting it on VMS you’re doing

35:24 you’re putting it on VMS you’re doing there’s a whole bunch of of extra yeah price that goes with that yeah and I think about oh my gosh how easy is it for me to then get to powerbi to embed into applications like when you’re starting it’s so much faster see but that’s so that that is always well and I’ve always worked in Microsoft tools so obviously I’m biased speaking of bias segue I think Kurt does a fantastic job of like Towing the line and he calls it out multiple times he’s like I don’t want to get into a tableau versus powerbi think that’s true that’s very true I’m trying to stay in the middle

35:54 true I’m trying to stay in the middle here yes one of the things that I think has been compelling and why I was a huge Microsoft fan and powerbi fan right out of the gate with this one is it was modelbased it’s it was solving all of these other underlying problems that are still massive problems for a tool like

36:11 still massive problems for a tool like Tableau all of the curated tool how do you connect all these things how do you merge and transform and like it’s a precursor to all of the other stuff I want to do right so to to me that modelbased approach versus the report based tool approach I think is like Kurt lays out really well the significant difference between the two but something I’ve always thought powerbi had it would because powerbi had that it was a much

36:41 because powerbi had that it was a much stronger tool like to to scale within organization and leverage in the ways we see Microsoft taking advantage of it and I think I as we’re going through this and I’m I’m looking at the documentation I’m looking at the the capabilities all well and good that you can do this but if I was a consultant and I was like this is the migration factor I don’t think this would be my recommendation because the cost would be probably the biggest one oh you’re already using Tableau now we’re going to build on top of this fabric licensing

37:12 build on top of this fabric licensing just to Showcase what fabric can do that’s not a really great selling point because you’re already incurring a lot more cost I think from the Mig if I if I want to migrate eventually people from Tableau to powerbi and one of the ways we’re doing that is well we don’t have time to change our reports from Tableau to powerbi why don’t they just connect to the semantic model and fabric we’re going to build that and showcase what fabric can do from a semantic modeling point of view time time out out what in this scenario

37:44 time time out out what in this scenario what is the business problem that you’re trying to solve if we’re trying to migrate from Tableau to migrate that’s what we do no no that’s not a business problem what is the business problem cost is is solution dver is your is your data all over the place the driver and nobody has any control or understanding of what data is going into any of those reports like what is the there has to be a key

38:16 like what is the there has to be a key driver for you to take on a migration or a PC or for you to move anything into a different platform you could take one of the 10 reasons why people are migrating today cost a but okay AB no I’ll name all 10 if you would like so and you can pick and choose from the menu the cost regardless though I think your your view on things we don’t need to go in all the 10 reasons why but I think your your high level you’re looking at it from an Enterprise centralized team

38:46 from an Enterprise centralized team perspective I think you’re undervaluing the amount of political horsepower that comes with getting everyone to move on to the same thing and also there have been in larger organiz ganizations there are teams dedicated or have been solely focused on getting wins around things that have been solely built in Tableau so I’m not I’m not about to discredit those teams and those that amount of work and effort that has been there there is physically a gap in knowledge between what we’ve been doing in Tableau and what we can do in powerbi because the the team skill is not there yet plus

39:19 the the team skill is not there yet plus some organizations don’t care about centralized cost looking at all their bi tools they’re focused more on each business unit has level of spend and so my argument to some degree here is if there’s a team that’s generating revenue and they have a they’re they’ve decided as a team they’re going to be their revenue is part of that will pay for Tableau because that’s how their team operates yes I would agree we probably should all think about how could we plan to get towards powerbi will help us out

39:49 to get towards powerbi will help us out but that team will have to integrate potentially with other teams across the organization get data from other things if I’m the C team I’m going to try and build semantic models that help me distribute that well allow allow me then to refine my declaration EXC so to speak which which may help in a lot the many of the scenarios the cost does not seem to outweigh the value of approaching things where we’re connecting to a semantic

40:19 where we’re connecting to a semantic model where we’re trying if we’re toing this particular situation outlined in the in the article I don’t know where the value outweighs the cost both from time any switching and and just monetary cost that’s what that’s trying what I’m wanting to State State here there are great so if it’s sure I guess if it’s an open qu like I I if we’re in an all or nothing

40:52 like I I if we’re in an all or nothing right I I think the value is you could have a few data sources that you’re you’re going to create semantic models around because you need to shape and combine the data understanding and knowing having the skill set to be able to do that on the powerbi side and or work with a tableau team there’s value there I I think the main drivers for cut over would be something like okay your organization is starting to play in the lake you have big pipelines that you’re starting to run because you’re you’re pushing through a lot more data

41:22 you’re pushing through a lot more data so you’re already looking at platforms that can support that and there’s many of them many different services that you can plug into many different service providers Microsoft Amazon Etc right so world is your oyster in ways where you would select something like fabric is oh okay as an option like and and I’m not talking fabric right now I’m talking fabric like six months like when it goes GA when it’s fully baked Etc there’s a that’s a compelling offering from the standpoint

41:53 compelling offering from the standpoint that wait a minute you’re telling me I can get all these Services Allin one and I I don’t have to migrate off of Tableau right away long-term strategy we may but I can get my pipelines I can store my data in the right way I can connect to this like okay like that to me is something that gets thrown in the mix as a potential solution yeah because without that then like in the Microsoft world and I’m I’m sure there’s the the AWS world as well but Microsoft world

42:23 AWS world as well but Microsoft world it’s okay I don’t need to go grab ADF service I don’t need to stand up a lake I don’t need a third party Lake data bricks I don’t need all of these other things that are problematic for large organizations or small to to just manage Implement correctly connect blah blah blah you get all that in one one one package and there is and and the cost is and this was my other point Tableau and

42:53 and this was my other point Tableau and the costs of Tableau are the report and what I can say is powerbi and the costs associ associated to it are the ecosystem of data it’s the ETL it’s the models that are supported all the time it’s the reporting it’s the how many other Enterprise Services under here that I can also leverage in different ways and all of the sudden that makes a lot more sense especially

43:23 that makes a lot more sense especially if I’m undercutting my competitors by X right yeah because now I’m on now I would be moving to a platform that is accessible to other tools and gives me options versus what those current Solutions are is I’ve picked this thing and it’s the only thing I can use which eh is is speaking from a a a naive view of not not plugging into all those systems so there I’m sure there’s capabilities moving and pushing data in different ways however you

43:54 data in different ways however you data in different ways however semantic models are a thing that know semantic models are a thing that that I I think have been proven to be the the industry leading standard as far as like speed efficiency and serving up data in the ways that we would need to from an analytics perspective so one thing I’ll just I’ll just note here from a a tableau perspective that I’ve observed when I’ve seen Tableau show up from companies and i’ I’d be conf willing to confirm this with you guys as well when I see Tableau appear for users typically there’s a another team

44:24 users typically there’s a another team that is sourcing or providing the data or the Tableau team themselves have purchased another tool in addition to Tableau like an alrick or something else that is shaping data so in my opinion you’re already Tableau is already primed in a way to cut themselves out of the the ecosystem because they don’t have so what what fabric is giving us is more control further Upstream to do data engineering things that we may want that’s stuff that we’re now giving to the business

44:54 that we’re now giving to the business potentially to have them do their own data engineering to some degree and it’s all part of the same pricing the same tool it all bolts together so that’s just my observation have you seen the same thing with Tableau when you’ve seen it in out in companies no I’ve seen this the mix of the people who are building Tableau are also skilled in a tool like altrx okay but they are Le another tool

45:17 altrx okay but they are Le another tool yeah they they’re learning another tool because it’s because it’s very visual based it’s very drag and drop it’s very I see my pipeline it’s whatever I think that where I think in the very near future that changes is is how fast and how well Microsoft can Implement co-pilot cross fabric oh that’s a good point because why would I have business like and this is this is the prep of all that right like I think you compile all

45:47 that right like I think you compile all these services and and they’ve even talked about copil here copilot there copilot here and we already know what chat PT can do on llms etc etc it’s true like I want to create a pipeline that sources data from this table here here I want you to transform it I want you to do do this there’s your pipeline wow tell me that’s not like it’s absolutely because it’s already done in data bricks like there always been there there have already been demos where people are like I thought Tommy you did one where like somebody was no no no no it was a

46:18 like somebody was no no no no it was a different MVP posted on LinkedIn watch I can transform like watch this you can transform like watch this engine and I thought it was the know engine and I thought it was the data bricks one can transform data I just keep talking to it like this and it’s it’s building my own pipeline so I I think what we all know about AI is the barriers for technical entry are going to be lower and who are who are the subject matter Experts of a business area that understand how you need to

46:48 area that understand how you need to transform data it’s the business User it’s the business user dist Ste it’s the person who is the one cobbling stuff together right right now that could be the one that is determining like okay here are my data sets I want you to do these things to them do these things to them and and show me the output and then I’m going to analyze it oh yep that’s exactly what I need now I’m going to say go create me a report I want you to show me this this and this the whole time you’re talking I’m trying to figure out a way to argue with you or to

47:18 a way to argue with you or to Counterpoint but there is none my open AI just had their big conference and they had this thing called GPT man you are you’re right Tommy sent me a video and it’s it’s impressive what the thing’s doing now what what just what chat GPT can do now and what you can build as a business user you and no and it’s a really really great Point here that you’re right when co-pilot and not just co-pilot but when just what people are probably going to

47:48 just what people are probably going to build for these migrations too just not just necessarily Microsoft but what we could build our our and plus explicit measures podcast GPT that we could build that help with the migration what it can feed in we’re only moving in that direction don’t forget open AI has only been around for 12 months yeah right so this wow what yes it last year publicly available Rel releas a publicly

48:20 available Rel releas a publicly available version chap DPT has been around for year it’s been rolling for a while yeah yeah it’s not like they just developed the code last year but chat GPT was available this time last year or just became available so we are very very incredibly early in the stages of what AI is going to be able to do and how we can feed it and what the workflow looks like that’s a great Point Seth and I’m like I said I wanted

48:50 Point Seth and I’m like I said I wanted to argue so much like oh what about this or it’s like well it’s and again we know it’s not going to solve everything however it’s going to be a great at least framework for someone to build what they want and then yeah maybe having someone else look at it but I having someone else look at it but there there’s a lot of good mean there there’s a lot of good things coming so let’s try and do I think we’ve done a good job of in diving into Tableau thinking about where that fits there’s a lot of other

49:21 where that fits there’s a lot of other tools that built onto the semantic model piece I think Microsoft is making the right decision here random question my son asked me the other day and Tommy I’ll see if this because you you are a huge chat GPT fan what does GPT stand for it’s I’ve see it all the time I know the first one is generated I had to go Google it because I did not know what the word the moniker GPT stands for I I know the first word is generated I cannot tell you the other so

49:51 generated I cannot tell you the other so it’s I had to Google it again because I forgot the the phase generative pre-ra trained trained Transformers that’s that’s that’s what the GPT stands I’m like really I thought it was something else like yeah it’s they just abbreviated it and threw it together and like a little chat GPT that’s that’s the that’s the whole moniker anyways cool technology we’ll be interested to see where that goes with co-pilot I’m very excited to see what Microsoft announces I believe there’s a lot of announcements going to be happening here or have happened already at ignite so that will be if

50:23 at ignite so that will be if you haven’t watched it already go back and check it out ignite should be a very interesting event where Microsoft is going to go a lot deeper and what their plans are around things data and AI based things in general with that I just want to say thank you very much oh before we do that I forgot I’m skipping ahead here we should do final thoughts so this article on RBI semantic models any final thoughts let’s do a quick wack up here Tommy what what’s your final thought looking at this one I think we’re moving more and more Direction where other tooling besides just

50:53 where other tooling besides just Microsoft’s straight Tooling in around business and data business intelligence and data is going to become more prevalent and more accessible something that we should all keep our our minds open to and maybe even looking into that a general skill set cool Seth what are what are your final thoughts on this this topic final thoughts check out Kurt’s article he does a much better job than our hackery of discussion around it in describing how you can access and utilize Tableau walks has the tutorials has whatever

51:24 walks has the tutorials has whatever current you’re doing a fantastic job thank you for all the content and I think Microsoft is making All the Right Moves not only now in integrating and and working with other tools but I’m excited to see where things lead in the future I would say I think I I think the latter Point here on this one was really impactful to me so your point around what’s the business use case for this why would you want to migrate on or off of Tau it’s going to be a very important question for organizations as they try to think through what is the best and I think the main reason

51:55 best and I think the main reason organization start centralizing on a tool is cost I think that’s one of the main drivers of getting cost down there’s a lot of hidden cost that come along with that education of your users building it out start standing up new things right you’re you’re potentially rebuilding a couple things you’ve already previously built in other systems so I would say from a final thought standpoint think about why you want to migrate away or think about the cost implications of why you need to keep around other tools click top

52:25 keep around other tools click top whatever that looks like and think through strategically where does it make sense to best spend your effort around managing your bi space and I think you’ll find that there there’s some compelling lower cost Solutions out there than you have today potentially so that’s the reason why we all have jobs is to continue to optimize and and bring down costs right you don’t pay for Consultants well then set you your fulltime on this one but you don’t pay for people to come in and like Design Systems unless it’s at a lower

52:55 Design Systems unless it’s at a lower cost than if you didn’t do take that effort or take that time all right well that we’ll wrap up here thank you very much for your time your listenership your ears we thank you this is a a a fun 50 minutes of talking about random things we appreciate you engaging with the conversation please if you don’t mind if you like this conversation please make sure that you share this with somebody else let somebody else know that this was a Content that you liked engaging with Tommy where else can you find the podcast you can find us in apple Spotify or wherever you get your podcast make sure to subscribe and leave

53:26 podcast make sure to subscribe and leave a rating it helps us out a ton do you have a question idea or topic that you want us to talk about in a future episode head over to powerbi tip podcast 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 powerbi tips social media channels thank you all very much and we’ll see you next

54:16 [Music] n

Thank You

Thanks for listening to the Explicit Measures Podcast. If you have a topic you’d like us to cover, drop it in the suggestion link above, and we’ll add it to the queue.