PowerBI.tips

Let's Talk Power BI Templates – Ep. 306

March 29, 2024 By Mike Carlo
Podcast Power BI
Let's Talk Power BI Templates – Ep. 306

In this episode of the Explicit Measures podcast, Mike discuss Let’s Talk Power BI Templates - Ep.306 - Power BI tips from the Real World.

News & Announcements

Main Discussion

This episode focuses on Power BI templates and why they’re valuable when you’re trying to scale reporting across a team. The conversation covers what templates should include, what they should not include, and how to keep them maintainable as your organization grows.

Key points from the conversation:

  • When templates help: consistent layout, navigation, and visual standards—especially when multiple report authors contribute.
  • When templates hurt: overly rigid “one template for everything” approaches that fight the business process instead of supporting it.
  • What to standardize: branding/themes, accessibility defaults, core page patterns, and common measures/semantic model expectations.
  • What to keep flexible: business-specific storytelling, domain-driven visuals, and edge-case analysis pages.
  • Operational reality: templates work best when paired with documentation and a clear publishing/review process.

Looking Forward

Keep an eye on upcoming updates in Fabric and Power BI, and let us know what topics you want us to cover next.

Episode Transcript

0:29 good morning everyone and welcome welcome back to the explicit M podcast with Tommy Seth and Mike good morning everybody good morning Mike got a number of really good intros today just a couple articles have been passed out recently looks like everyone’s getting a little bit excited about all these new things coming out from RBI desktop so we’ll jump into those in our new section but before we do that we’re going to talk today about templates we at powerbi tips are big fans of templates and working smarter not harder trying to build a

0:59 not harder trying to build a more efficient workflow is the thing that we’re all about and how we started the website and a lot of the things that we do were just trying to teach people there’s a better way than clicking all the buttons so we’re going to get in di Deep dive with some templates data mircat has a great article Stephen is talking about powerbi templates and talks to the different areas or the different parts of a powerbi template there’s number of different formats now that you can go use and create and build things with so how does that work and

1:29 things with so how does that work and how can you leverage these to make you look and work smarter and faster all right with that let’s go into some in introductions or topics Tommy what’s the news today what do we got we got three very interesting topics no necessarily breaking news just three good nuggets of information first one is by our good friend Eric SE Seth you want to you want to say I’m terrible with Swedish names Eric yes and the V have always thrown me

2:01 yes and the V have always thrown me off but he’s phen he I love his articles when they come out he just came out with combining visual calculations and numeric parameters to create a dynamic moving average this is hitting home for me because I’ve done this in Dax and where basically you select hey show me a three-month rolling average a six month or nine month just by selecting a slicer on a line chart I’ve done that in Dax it’s no fun to get that get that set up it’s a lot of lines

2:32 that get that set up it’s a lot of lines and we just talked about two visual calculations so it’s like oh maybe here’s a good use case because that may not that may not be a measure that’s going to be ever part of like that certification process right that sounds like more from an analytical thing more often than not not what someone’s looking at the company’s gold at so another interesting twist that he does on this one is he actually makes a parameter as well that allows you to

3:03 parameter as well that allows you to adjust the the the amount of smoothing or the moving average that you want so the length of the smooth right so that you what you’ll do for smoothing typically is I’m going to look at the last three data points and those last three data points is how many like if you’re doing months I’m going to do a moving average over three months maybe I want it over six months maybe I want it over 12 months so he actually does a really interesting addition to this not only is he using visual calculations in the visual but he’s also allowing you to adjust the amount of smoothing you

3:33 to adjust the amount of smoothing you want based on a slicer which I think is actually I really like that I thought that was a very slick neat example there as well how he’s incorporating the parameter in addition to letting it use the the visual itself so now I at first I was very confused on how how did you get the moving average to change inside a visual calc where the visual calc was interacting with other data points because my understanding was

4:03 data points because my understanding was a visual C can only work on the data that’s supplied to the visual does that make sense yeah so that was really interesting that it was allowing you to adjust that using on the this the the moving average Portion by generating a different series it’s so simple this way way too well this is why visual CS started to exist yeah yeah it’s a whole new world A Whole New World yeah like this this is I think one of many

4:33 like this this is I think one of many where as the community really really gets their head around and adopts visual calculations to your point Tommy like this is It’s a this would be a very long very complicated thing to build and this is a step by step the fact that Eric went through and it’s like hey can I actually like how what other things will interact with or work with visual calculations and the netet result is a big benefit to business users right

5:03 a big benefit to business users right because it’s not big scary Dax that oh my gosh thanks for outlining how to build this all out in a Dax formula or series of them and then you have the inevitable question it doesn’t work in my model right where this this tutorial and stepbystep thing is where business users can leverage and get a a not an immense amount of value like this is a very complicated thing to do prior to visual CS it’s funny too

5:35 to visual CS it’s funny too because I’m thinking about all the scenarios that I’ve had back in the day we’re like hey the report’s great but we really like to see X Y and Z that’s the really important and that’s the usual Dax that takes the longest time it’s that very specific use case that’s apparently essential for them of that I’ve always been like okay we got to figure this out I have the three measures I have to clean it up again oh look two lines in visual CS done weekend

6:08 saved Bingo what else you got Tommy well one thing I didn’t put on agenda but Mike they just announced Windows surface 10 it has a co-pilot button Windows 10 surface 10 Windows Surface 10 I thought we’re talking about parb on this show I don’t really let’s keep that keep that one I don’t care about window Tommy Focus Tommy focus a little technical too the things that we use moving on number two is power query or what I believe we’re calling FX

6:42 or what I believe we’re calling FX GPT and this is from the power of business intelligence and all it is is a custom function in power query to basically go through go through your columns and you can add a prompt to it and then it will autogenerate from there so so cool I can’t imagine the tokens yes this is this is a bit wild in my opinion so they’re taking power query writing a custom function that you send text into which then uses it bundles it

7:14 text into which then uses it bundles it sends it to chat GPT and returns m is that what it’s doing yeah exactly you you got system a user and what model you’re going to use from chat GPT row by row but I have a few use cases I had right now a column with I think it’s like 761 unique values and I’m like how could I AI group this because they’re all very similar like oh this could work interesting and this is this is where I

7:45 interesting and this is this is where I don’t really know I I like the idea that we can do flexible things like this yeah in in my mind I go though is this really practical usable like it’s a it’s a cool example it’s a neat little demo like cool this is interesting but and I’m finding a lot of the chat GPT stuff or even just the AI things in general it has to be like people have to do a lot of testing to get it to do what it wants it doesn’t it isn’t quite how do I say it doesn’t quite Amaze the way I think it will but I think it’s definitely getting better so I it’ll be

8:15 getting better so I it’ll be interesting to see where the depths of this co-pilot stuff will be will allow you to go and where the the product team will take it and how they can shape the prompting in a way that makes sense to the users I think that’s going to be the hardest part is helping users be able to create the the correct prompts to get out what they want yeah what I’m finding across the Microsoft Suite in general every program has its own unique way of prompting like you think about it like the way I would want to prompt in

8:45 like the way I would want to prompt in PowerPoint is different than what I would have prompt in Excel which is different than what I would have prompt in vs code which is also different how I would prompt in powerbi either building reports or building Dax or M right right so you think about like the different context of what co-pilot needs to do and the teams have to develop a co-pilot for that specific type of persona it’s it’s challenging because as a user of multiple of those products you have to learn what works in each

9:16 have to learn what works in each tool which is interesting because like essentially what you’re saying is I need a co-pilot to find the right prompt pril it well but ultimately what that leads to is like what library of prompts that I then have to search for the right prompt then go execute against co-pilot to get the right result yes that’s it’s mindbending a bit just slightly and the value the value though and we talked about this last Thursday

9:46 and we talked about this last Thursday in the podcast a little bit just from the foundations of data right like all all this stuff still needs to be organized or at least put in in a sensible way otherwise you risk the output not being what you would expect it to be right so now we’re saying like not only does it does the data in some way shape or form have to tie out to a potential answer right in order for it to be accurate we’re also

10:16 order for it to be accurate we’re also saying well you have to ask in the right way way still and you need and that’s going to be different per product so we’re introducing a different level of complexity here and you have to be really conscious when using the API of cost we I I’m curious with all the

10:33 cost we I I’m curious with all the things that we’ve talked about how much more expensive data flows Gen 2 are compared to notebooks from processing I don’t know if that has anything like if there’s any correlation here but because you can easily spend more than you’re willing in chat gbt if you’re using a lot of apis you’re like hey I’m just gonna go over the next 15, 000 rows with jabt that’s an interesting point because when you’re when you’re doing co-pilot co-pilot will let you look at co-pilot charg is based on the CU usage

11:03 co-pilot charg is based on the CU usage there’s a there’s a CU n the number of words you send in is how much CU you’re going to consume I I would imagine because you consume I I would imagine because chat GPT or co-pilot is running on know chat GPT or co-pilot is running on graphics processor units it’s very expensive and it’s costly for them to run that stuff so what in my mind when you say that Tommy I think I want a thing I want different things to fail when I run out of capacity units like I

11:34 when I run out of capacity units like I have an order of precedents that I may care about right for example during the middle of the day if I have a whole bunch of people hitting co-pilot and my reports stop rendering I went co-pilot to fail first and then fail my reports if I have data set refreshes working in the background and I have co-pilot running and I have reports being rendered I I would want the backend systems to continue to load and then I would I would want to

12:05 load and then I would I would want to make sure the reports are still in loading and then I would want chat GPT to run so I don’t I’m not exactly sure how all this is going to pan out but I’m also thinking here as well like again as I was going through my class recently and thinking about the different fabric skews that we have now now that fabric is out and they have announced the the retirement of the P SKS there’s really no reason why you need any other skew other than a fabric for premium but with this you can have a lot more catered experiences for different

12:36 catered experiences for different workloads that don’t allow chat GPT in them like let me give you an example right backend loading you’re building a lake house and maybe you’re loading your semantic models those may not need a full f64 all the time to run so maybe you can put those workloads on a smaller fscq that doesn’t even have chat GT chat GPT on it so below 64 so therefore you’re never at risk of people asking questions of your data or using that and then hurting that part of your process

13:06 then hurting that part of your process so you’re basically you’re literally isolating the features away from what you need like Mission critical stuff and then when you’re doing like front- facing reports that’s maybe an f64 and because now you’re dividing you you can divide the workspaces into different capacities now you can shift that load you think somewhere else you think by using different capacities that are not co-pilot enabled like lower that it’s going to remove the data from the

13:37 going to remove the data from the purview of a potential f64 Co copot you you have to attach like when you use copilot in desktop or anywhere you have to attach it to a workspace that has f64 on it so if you’re in desktop and you try and use co-pilot yeah yeah I get it so but you’re you’re man you’re you’re isolating you think your your thought here is you’re going to isolate data based on capacity I think you’re going to isolate workloads based on capacity I to isolate workloads based on capacity before it wasn’t that way it was mean before it wasn’t that way it was like it was everything was a

14:07 like it was everything was a nail and I have a hammer and just all on premium right that was the original thinking and maybe I have a division between Dev and prod now I’m thinking because now I have lake houses that may need to be their own lake house I may have semantic models that need to refresh in their own workspace so now I’m seeing the the the distribution of security and workspace so now talking workloads we’re talking teams of people that need to be to segmented I think there’s more opportunity again I’m say I’m saying you have to do this say this

14:37 I’m saying you have to do this say this is an opportunity for you to more comprehensively think about your environment and it I think also it provides you much more opportunity to automate right so in the morning you turn on your f64 for back-end loading you let it run all your data sets and then you turn it back down to an F4 during the day because you don’t you’re not loading any more data and then then now you saved yourself a bunch of money all your semantic models have now can go back to the lake house and pull out what they want there’s no compute need needed for direct Lake because it’s on the the

15:08 for direct Lake because it’s on the the semantic model side so then you’re saving yourself a whole bunch of money there so I I I feel like there’s a workload to skew relationship happening now or going to be able to happen so in I’m confused are we talking about two different things or I misunderstood what you were trying to communicate cuz I heard you talking about manipulating capacities for cost correct and but I thought we started in manipulating capacities to only allow co-pilot access

15:39 capacities to only allow co-pilot access to certain data so the answer is yes to both you you can manipulate SKS for cost but you can also manipulate SKS for features as well one of is testing out but because if I have a workspace that is on an f64 capacity that is copilot enabled y I can connect to data from other workspaces right I can have shortcuts or sure but that’s a shortcut is not using the capacity from a different workspace

16:09 the capacity from a different workspace understood so your presumption is that any what you say go ahead I don’t think it’s a presumption I think it’s what happens but okay well I I would I would love to test it out you’re saying that you would think that the data in a shortcut from a non f64 capacity would not be available to copilot doesn’t matter if it’s from cop it’s where the computes being used from it’s the data is still available blob storage is just flat storage there’s no

16:39 storage is just flat storage there’s no there’s no cost for you to have storage in a different I thought I thought your your point to co-pilot was to manipulate what what data had had access to oh no no not at all I don’t I’m not talking anything about data access I’m talking more about cost okay and management of cost pieces because where the workspace the co-pilot lives in that’s the one that’s going to get charged for the cus of the co-pilot but okay so let’s take my idea my idea okay okay which I I don’t know if it resonates in

17:09 which I I don’t know if it resonates in anywhere but like that would be interesting right where you’re you’re playing scaling but I still want to use co-pilot but I’m going to run typically on a much lower skew but still have but still have access to all the data by shortcutting from workspace that’s not64 that’s exactly right rather than everyone like that’s the thing is is that data is that data available mhm it is I’m pretty sure I’m I’m almost 100% short if it’s in and again what the

17:39 100% short if it’s in and again what the thing I don’t know is are they hard blocking like even imagine this can I haven’t tested this one this is one I actually physically need to test turn on turn on a workspace put it in fabric run some stuff hydrate the data into the lake then put it back on a pro license so the lake the one Lake L has the ability to see the data in the in the workspace but there is no fabric capacity attached to the workspace let me ask this question so that’s one thing I don’t know will work I don’t sure if

18:09 I don’t know will work I don’t sure if they’ve actually blocked that path at this point I have to test that one what what would you what would you prefer would you prefer the high Jinks that you’re talking about where we’re manually manipulating capacities and using different things to well I don’t know what you would be turning it up and down outside of manually like you is there some automation that that you think is available because we’re in an ecosystem as opposed to Independent Services correct so right now so this is this is

18:39 correct so right now so this is this is what I think let let me ask let me say this way I just ran my class and I brought up this exact topic and they love the idea and they’re like you build it Michael we’ll we’ll pay for it like multiple large customers were asking for this exact feature because today there is no easy automation button of just saying like so think about data bricks and how it auto scales up and down you basically said I here’s my minimum workload here’s my maximum workload the tools that we’re given today are very

19:10 tools that we’re given today are very close for us to be able to actually scale up and down and have this all automatic what Microsoft will either Microsoft’s going to have to build this or someone’s going to build this yours truly will build it and therefore from that it’ll actually Force Microsoft’s hand to start thinking about autoscaling the f skes so what should be happening here is you should turn on an F skew it should start at an F2 no no matter what it should always start F2 and what you should do is you should start throwing workloads at it I’m going to load this

19:40 workloads at it I’m going to load this data set I’m going to load these things Microsoft should be able to approximate what that data set load will look like and say uhoh we’re g to we’re we’ve already maxed out the F2 or let it run for just a hot second and it maxes out the F2 look I’m hitting at my threshold at 100 and then when it says okay I’m going to bump up this looks like a workload that I need to bump up to an f8 so then it would then scale up to the f8 which would exactly what data bricks does scale up to the level it needs run the job and then within 15 minutes it scales itself back down to the smaller amount like that level of automation

20:12 amount like that level of automation should be built into the product right now yeah but at the same time at the same time here I and I provided very similar feedback to certain individuals but but I would say no and but the your description of even how data breaks works is not is altruistic as you you consume there are a ton of workloads or things you can kick off where it just goes hey I’m going to max out everything I’m going to max out your your whatever you selected and I’m going to go full bore and then it’s going to fail and I’m

20:43 bore and then it’s going to fail and I’m still going to run for a couple hours so there are inefficiencies even in the low like Min and Max cluster usage yeah but it does eventually ramp down turn off however I agree from the standpoint that a little bit better of the burst up burst down scale up scale

21:03 the burst up burst down scale up scale down type of activity would be extremely worthwhile where I think any of these companies win is where they build confidence with consumers that they’re going to manage that for you and they’re they’re looking at Cost optimization for you because if you if as a cons as a heavy consumer Microsoft Services and others and third party it is the

21:35 others and third party it is the largest pain in the rear for any organization to get stability or understanding around costs and if you stabilize the foundational data and realistically if we if we look in the long like in the Spectrum I think out five 10 15 20 years right like where we’re at now is this is the foundational data era right like we need to we have tons of data we’ve figured

22:06 to we have tons of data we’ve figured out how to store it cheaply and and compute of those workloads should be extremely well known to all these companies you build trust in optimizing that cost and that frees up money to go spend it on all your AI stuff 100% then you plug that in and that’s where all my CPU goes and I’m like yep I’m all in because it do like now the data processing and storage like the hey I

22:36 processing and storage like the hey I just need to form my data so I can leverage it thing is like I think we’re there I think we’re right in that realm where it’s like there’s an acceleration Point here and whoever makes the ease of use of storage and CPU of just just the data things right because I think that like we’re already moving past the conversation of like the foundations which is always interesting for us as data engineers and bi developers because we know how much

23:06 bi developers because we know how much work it takes to like get things through these processes but we also know that you could screw up in this environment and cost your organization $ thousand doll in a day instantly easily easily EAS so are there guard rails on like the capacity itself in fabric not allowing you to do that where you could in other tools yeah but at the same time you could probably also be saving a ton of money especially as you’re proliferating fabric throughout an organization that would need to be streamlined and Mike

23:37 would need to be streamlined and Mike that’s actually the the direction I thought you were going to go that’s actually one of the I think the approaches in documentation is that you have dedicated capacities for co-pilot because it can be a significant load on a capacity and if you just do everything co-pilot in every workspace if that’s part of your we know semantic models can take up an entire premium capacity we’ve seen that before so now you’re combining co-pilot so that’s from a much simplified approach

24:08 that’s from a much simplified approach one of the things Microsoft said I think that makes sense is there’s dedicated workspaces that are going to be meant for co-pilot for the right people because everyone’s prompting all the time and I hope your sales team has some things in the pipeline well pring the right things the name of the game here is make it easy really is what what this is right so all these services are azzure more Azure Based Services what I see fabric doing is fabric is is going to absorb more and more Azure Based Services and so they’re going to

24:38 Services and so they’re going to continue bringing those features to powerbi and again the easy way the easy button here is we need to be able to capture cost for Network traffic for cus all the things and so everything’s becoming this capacity unit based measurement which I think is good I think it’s I think it’s already it’s a good place to go all right Tommy you want to skip the last announcement go into the main topic all right so in in an effort of having a long conversation around these

25:08 having a long conversation around these news articles today which has been great also very great for talking about topic things good things to think about and consider moving forward then let’s talk about our main topic today today is talking about powerbi templates and again as we’ve said we’re big fans of templates we think this is super huge it helps out organizations a ton and adds a ma a massive amount of value to your workflows workflows so Steph gives us a great article around this one the link will be in the description from data mircat and it’ll be powerbi templates is the main article here so for the rest of the

25:38 article here so for the rest of the world outside of Mike’s close personal friend Stephen this this is Steen richel I think who’s from the Czech Republic so I’m going to run that past my wife and she’ll yell at me later if be like that I am so disappointed in you that you still didn’t get it right but hopefully that’s a little closer yeah true true so give us a a quick overview Tommy what’s what’s the the general gist of this article here so you may be aware already of a powerbi template oh yeah there’s the pbit but you may not

26:09 yeah there’s the pbit but you may not know there’s now actually four other ways you can technically create a template in the service in your developer tools which you may have never either heard about or have no idea how to use and obviously we know we love the powerbi templus the theme generator really talking about report based themes Stefan or I’ll say Stephen as well yeah we Ste we’ll say Stephen Stephen Stephen starts with the visual

26:40 Stephen Stephen starts with the visual templates the report based templates some people also use data templates which is actually based filtering or configuring the model what data is coming in and that’s what we’re all aware of that’s what we probably all know however Microsoft is actually quietly released in over the last couple years and especially with fabric some a lot of additional ways to create a theme that all serve a very different purpose for example I wouldn’t say I wouldn’t say quietly I think I would say with a lot less Fanfare than like visual

27:12 with a lot less Fanfare than like visual CS right it they’ve released them but the community is like yeah I see the point but I don’t want to write Json like they’re released but the community doesn’t really maybe have the tooling they need to support it yet I think if I pulled 100 powerbi professionals I think 10 of them would know about putting template on a new report in the service at the suffix knowing that tool so let me explain so you can actually go into the semantic model in the power ba service and when it says Hey create a new report off the semantic

27:42 Hey create a new report off the semantic model if you add the suffix parenthesis template and parenthesis it actually creates a template file that when someone is going to create a new report off that model uses the template so that’s one way it lives in the service only new with developer tools well we also have the pi pbids file and we also have power query Gen 2 templates the model. Json and so very developer

28:12 the model. Json and so very developer tools as well so our options of like you said without a lot of Fan Fair expanded yeah I don’t I don’t even know honestly if there’s a even a Blog out that lets you see the model starting from template experience I think that was out there for a bit we might actually have to blog that one because I don’t know I can’t put my fingers on any articles that actually describe that exact feature and honestly I don’t know they if they’ve deprecated it or they’ve even kept it around I’m not even sure if

28:42 even kept it around I’m not even sure if that still exists which which is this one that we’re talking about like in the service so in the service if you publish a report connected to a model and you change the name of the report to actually contain like parentheses template parentheses in the name it knows that that that is the template file that they want you to use when you create a start from existing or start from new report in the service so the

29:13 from new report in the service so the idea would be is you could have like a page with some theming stuff in it so that way when you create on top of that data set it now automatically has a template to start from so I think that’s the feature that one that one that one seems like the oh that’s an interesting finder like the way you figured that out thing as opposed to a standard Microsoft like to me it would just be easy like hey do you want to create the explore explore the model create the create a template

29:43 the model create the create a template if they wanted you to do that like I don’t I I loved it I love the idea I just haven’t I haven’t seen it anywhere else but this blog which was interesting so I’ll like noted we’ll we’ll give it a give it a give it a crack so but yeah for for we we ripped you Tommy you ripped through a ton of the different types of templates for all users that may not be familiar with them should we walk through like what are the specific things you can do in each template I’ll let you drive for the sake

30:14 template I’ll let you drive for the sake of time I think you can yeah so yeah templates start so he starts outlining the different kinds of templates right there’s things at the visual level there’s things at the theme level there’s things at a report level so building like the templates for working with data models you have data templates like the Bim file like that’s that is a yeah and I guess what are the things we can do with them right yeah then you have the pbids file so a powerbi data source file

30:46 pbids file so a powerbi data source file which allows you to create a a common connection to like a SQL Server an Oracle database like you you can then share the connection to other people then he also talks about the model Json file so the model. Json file describes a data flow and how the data flow stores his information so you can actually manipulate a data flow using the model. Json file and add your own data now there’s a power query template file pqt file so that also now exists and I think those are all the templates that I’m

31:18 those are all the templates that I’m aware of that he’s talked about in the in the file I don’t did he actually speak on the pbip file I don’t think I saw that in there directly called out he did okay so the pbip format and also by the way just recently at SQL bits

31:34 the way just recently at SQL bits there’s now a PBI report Improvement as well so I believe ruy romano and Mohammad I believe just did a demo over at SQL bits and there were some there were some Twitter pictures that were recently sent out that were describing this new feature around hey you there’s now a new format for the report side the for the report side is now a breaking format but this new format now includes all Json

32:04 this new format now includes all Json elements for all the report side of things so it’s basically making it easier for the report to be consumed and used or built on top of so that was interesting as well so those are those are all the surface here so what can you do with all these things a lot of this is like prep work in front of so some of this is just sharing data like the pbids is just sharing a common connection string but a lot of the other ones like visual templates pbip format formatted

32:34 visual templates pbip format formatted files the Bim definition of different files all these other templates are kind files all these other templates are intended to store a pre-built of intended to store a pre-built solution and then help you get started from there so you you basically can cut out a lot of initial busy work or create a uniform experience when you’re building stuff yeah and we we really don’t talk about this enough maybe enough right like template templates are in many ways an opportunity for the hardcore Dev teams

33:04 opportunity for the hardcore Dev teams to pre build easyuse Solutions or shortcuts for the maybe less than technical user so like that maybe somebody who understands data or modeling but messing around with all the connections or obviously themes and we could talk about that at nauseum but we’ll hold off on that we like that one for sure yeah we we we do from the the visual layer reporting side but why don’t we talk about it

33:34 but why don’t we talk about it more I think is not really easy there to support it yet agreed but at the same time like even in steen’s article it it is developer focused like these are the the the back end pieces that are all there for us to build these artifacts or easy buttons for business to engage and It just strikes me as we’re ripping through each one of these that we’re we’re not promoting them as part of a

34:04 we’re not promoting them as part of a program or adoption or why aren’t they in the adoption right like hey a great strategy for as you’re introducing this thing to the wide audience of wherever is templates and or the pbds where you’re you know pre-pro work or whatever the case correct yes my theory is why we don’t promote or really talk about is I think there needs to be a lot more careful planning we have six different options

34:35 planning we have six different options with six different file types that are going to be stored more or less in six different locations and it’s not necessarily windows explor it’s not they’re all over the place so if us just start going pretty liberal in terms of hey we’re going to create some service ones oh we got some developer ones we got the report ones well how can you actually centralize and manage that I think that makes it ve overly complicated I think there needs to be more careful planning on what are we

35:05 more careful planning on what are we going to deploy and what are we going to allow I think from the developer point of view there’s some documentation out there but this is one of the few articles that I’ve actually seen where it actually goes through not just what a template is from the developer point of view but here’s some approaches that you can take we know what to do with a pbit it’s pretty simple use it for your theme file or use it for The Connect straight data stored in SharePoint so everyone has a centralized

35:35 SharePoint so everyone has a centralized location that makes it simple well I location that makes it simple well there there’s also a a the the one mean there there’s also a a the the one use case that I’ve used extensively which is the allowing allowing a setup of of a powerbi file to include parameters that somebody’s got to log in right away which you can con like make the easy buttons as far as connections are concerned or at least ease that that pain a little bit but the the end user

36:05 pain a little bit but the the end user who’s using it is automatically prompted to to input certain Fields without having to know that they have to go into Power query update data connection etc etc etc so that in it of itself is is the one use case that I I think is probably the most relevant that that you’re blowing past that I just want to point out but that it that is like you prompted right off the bat fill in this information and then it just goes and does its thing and that’s my my point I guess so I I feel like you’re trying to

36:36 guess so I I feel like you’re trying to argue with me for no reason but continue let me let me let me give one more perspective before we move on to keep arguing about that that point because I think you think you get some good points there on both sides I also think Microsoft has not done it justice on how they incorporate the use of templates into the rest of the product right so fair fair all these things exist and Microsoft them themselves uses these templates for things and I think the really reason why Microsoft has developed any templating on anything is self-serving they

37:08 anything is self-serving they want a pbids file so that way when they had purview show up perview could connect to any data source that’s being generated from powerbi so they could read a source and say hey look we have a shortcut that we can use in powerbi desktop to automatically connect you to this link server or this SQL database right so they want wanted that easy connection string so they built it for themselves because they needed it in other tooling that’s why it was built they just happened to get it out they document it and say here’s our standard

37:38 document it and say here’s our standard you’re more than welcome to use it if you want but no one else has and this is a good example is no one else has like a full perview type service that they can build that actually start creating source connection files now you’ll also notice too datab bricks has an instant connector that talks directly to the so you go to their SQL endpoint the SQL endpoint has a connection string inside that connection string is all the information you need for for datab bricks to connect to powerbi and you can

38:09 bricks to connect to powerbi and you can go ask for data from the the SQL endpoint datab Bricks now has a single button press that says go visualize your data in the parbi tool that’s making a PBS file that is then passed over to parbi that you then click immediately click on it you log in and you’re in so again big systems big companies or cooling pieces are now able to leverage that thing that’s that’s why Microsoft produced these things I think what’s missing here is more of a catered story for like the development team the leader

38:39 for like the development team the leader of that development team the governance and and making a unified experience that doesn’t seem like it’s there yet there’s no concept of like a global theme that sits across your entire tenant what I’m pushing that’s what I’m that’s what I’m s suggesting like why it’s it’s interesting because the the the the the because we’re talking about this and all just laying out all of these like shortcut these easy things that are buckets of work that we could

39:09 are buckets of work that we could implement it I’m calling out it’s interesting that that it’s not part of the conversation more yes well okay let’s look at Microsoft hey we just turned on this I’m not I’m not suggesting that Microsoft should have laid all of this out for us because we know they don’t do that they say here our technology hey guys go have fun and we have fun I guess maybe maybe let me repoint my question no I I I asked it correctly why haven’t we as a community

39:41 correctly why haven’t we as a community talked about templates more or made them more part of our adoption or our our implementation again is it because they’re they’re still pieces and they’re disjointed to some degree Tommy I get it like yes they’re all over the place but at the same time where can you go build a report or how many different ways can you slice the same thing in the powerbi ecosystem that’s not confusing for end users no but that’s different because there’s a dependency where I need to find my template files for my team so

40:11 find my template files for my team so there’s a the origin that you want to keep centralized let let me ask you a question when would you let’s take the simple two different ways to create a powerbi template when would you would you encourage the create a new report in the service you

41:28 template template template template template template any data set any semantic model every like literally make it a centralized thing that is here are workspace level assets that are going to be used across the entire powerbi ecosystem and again Microsoft if you’re listening hopefully someone at the team over there is listening to the podcast here’s here’s your next three4 of a million dollar idea that’s that’s going here’s the next here’s the next quarter million idea that’s going to take a lot of time to develop but I’m going to give it any anyways but you

41:59 going to give it any anyways but you going to give it any anyways but thinking through that and going know thinking through that and going okay what if I had not only building workspace assets for your

42:06 building workspace assets for your templates power pbip files or pbit files upload them get them started that way everywhere in that workes workpace experience I can and then let the admin of the workspace control everyone must start from a template or you have the ability to create something without a template a blank sheet and then also add the ability for you to add a theme file that can be forced to be on all reports or not and have that happen at the workspace workspace level but you can add that in the template but what happens when you have someone create a brand new blank file

42:37 someone create a brand new blank file with no template attach so but think yeah yes and no but like because we’ve talked about we’ve I I love all of your ideas and I agree 100% with them but it’s because it’s the way you create standardization even if you have a powerbi template that is like here’s the theme here’s what this looks like here’s and then raw like here’s the connection to the model everyone starts here if you want to share this out widely within an organization because we’re going to put some things in here that everybody knows that this is a finance report everybody

43:08 that this is a finance report everybody knows this is a commercial report you want to go build your own stuff fine but you’re not sharing it with everybody in the organization because we have standards we have a process like we don’t want the wild west of everybody having to figure out how you built a report and I think that’s where exactly where you would fit this in is it everywhere in the ecosystem of RBI no but it is in the realm of like where I’m going to start sharing things throughout and you could really streamline and make I think much

43:39 I think much better Solutions in this ecosystem with templates versus just what it is right now which is everybody just building their own thing unless or are running your own templates behind the scenes off powerbi desktop files you scenes off powerbi desktop files because most people don’t even work know because most people don’t even work with the pbip file format or pbit format and they don’t they’re like oh scary it’s not the PBX they’re like well in reality you save that into that it’s there’s an easy path to that ecosystem

44:09 there’s an easy path to that ecosystem and it just it’s done for you yes and and I think this is restricted because of that extra little bit of friction it’s restricted companies from actually developing we to I’m going toot our own horn here for a b we are the only company in the world who’s building pbps like we’re the only company in the world who’s building wireframes pbps template files literally talking about let’s build a standard that needs to happen for your company or your team start here and we’re and we’re provided with AI based tools inside the template building experience like we’re generating visuals

44:41 experience like we’re generating visuals on the page for you I can create an entire report look feel style within five or 10 minutes as opposed to having to go into the service or starting to build one from scratch that takes hours to build shaving off for those listeners for those listeners who don’t no what like power bi. tips theme generators what you’re referring to that’s what I’m referring to yes correct sorry I assumed everyone know knew already hey hey assume anything but yeah so but there’s power there there’s obviously right we

45:13 power there there’s obviously right we take a file we let you do a ton of stuff and then you have a starter that instantly turns into a report for you correct and and yeah we’re huge timesavers go check it out for sure but that’s what you can do like actually I’m glad you brought that up because yes we don’t like toot our own horn a lot on this podcast which is probably good we probably a little bit no we wouldn’t we wouldn’t have an audience I don’t think like hey go check out our our stuff all the time but what I’m arguing for is

45:45 the time but what I’m arguing for is exactly what we’ve done we have we’ve leveraged a lot of these little bits and pieces granted we’ve created an experience for users that shaves off at least half if you’re going to if you’re going to spend a lot of time building a report we’re shaving off at least half of that time because it’s all look and feel set your properties once like reuse it’s the reusable thing you can use everywhere and then we create the pbip file for you right like that’s what I think I’m saying is like hey you

46:16 what I think I’m saying is like hey you what I think I’m saying is like hey what why aren’t we talking about know what why aren’t we talking about the formats of these things or what else could be extended by individuals in their own deployment ments of powerbi leveraging templates more so start with us get that in there and then like have a standard process right like I’m we’re going to create a workspace this is this is a workspace that’s you this is this is a workspace that’s going to be used and shared out know going to be used and shared out throughout the company our standards are you’re using these template files this

46:46 you’re using these template files this is where you start then we’ve we’ve driven a lot of all of the other stuff we’ve talked about into the fabric of how people build their new processes and it’s all driven from standards and I think that’s what’s cool about templates and where my mind is going is these are ways we can build standards when we’re making deployments and I’m I’m gonna I’m going to reevaluate like huh where where can I talk about this more or leverage

47:17 can I talk about this more or leverage templates better than ultimately the great starting point that you get coming out of theme generator well I I think that’s a big part especially with the developer side that hopefully there’s a solution built on change management right so the theme generator or having the pbit or those standard basic Solutions well I’ve created great we have our template everyone’s using well let’s say we need to make we’re making updates to that things are out of sync people are out of sync and there’s

47:48 sync people are out of sync and there’s no easy way to update your theme unless you’re doing it individually to me that’s another Roadblock especially for team where I may be adding new pages to the theme file may not happen all the time but you will find where you have a wide range of variance between different reports over time this is so you’re speaking to things here Tommy that are very interesting so let’s let’s just take like there’s the there’s a macro and

48:18 like there’s the there’s a macro and there’s a micro version of what this is talking about right so the the the very focused I don’t know which one’s right the micro version right focusing on the very detail pieces is I have a report a report gets a theme file done like okay so now when you scale up and look at the organization I have hundreds thousands of reports across hundreds if not thousands of workspaces that stuff is amazing and then you step back and say how do I

48:48 then you step back and say how do I understand all the themed elements across all of my company and does any of my quote unquote standards and again this is where I would also push for you this is where I would also push for when you’re talking about things know when you’re talking about things that are certified versus not certified yeah I think you care more about the certified things having that similar look and feel you’re willing to bend more on the self-service side of things when it’s not certified so even so let’s just take that window right you have a huge organization I have a lot of things going on and then I’m looking at only the certified artifacts that I care about so just focusing on the certified

49:19 about so just focusing on the certified portion how do I configure or how do I update or how do I modify settings across all those reports and I think there is capability now with pbip format there is capability now with Git that you can use for those things however again to to your point Tommy and Seth earlier was there’s no tooling to support this there’s no tooling to support the the quick ingestion of all those different theme files comparing them and making a common

49:49 comparing them and making a common change across all of them yeah you have the you have the capability with Git to make the changes you have the capability to like modify very small pieces of a file but if you think about the bigger challenge for me in my mind right now is look we have these great things called deployment pipelines if I change a single measure definition in the dev environment I don’t get to deploy the measure definition across the environments I’m deploying the entire semantic model across all pieces so I think what the the community doesn’t

50:19 think what the the community doesn’t even know they need yet is they need finer grain controls around identifying what is the change and pushing only the change piece and then not and then ensuring that things don’t break long term and I think this is where the challenge comes from because Microsoft has to support small changes on lots of things and today they don’t It’s On All or Nothing they’re not good point it makes it makes sense from like a pre-build standpoint that you you can start these places but the change the

50:50 start these places but the change the change management parts yeah would need to get much cleaner if you’re going to do wider ad do correct but that’s but that is a Dev way of thinking and and you have to imagine like 80 90% of the powerbi community is not Dev thinkers they’re thinking like the business side and it’s self-service and it’s I make my change on my one file that’s easy enough for me to understand I’m not trying to go into pick and choose and it’s it’s value added but is it easy for users to to comprehend what changes are being

51:20 to comprehend what changes are being made and how to push those incremental changes across with things that’s not built yet yeah so really really cool thoughts there one thing I think set that you’re pretty interested around here is there’s this new thing that came out in data flows Gen 2 another template the power query template files I’ve I’ve always been well I’ve always been a fan of when doing trainings are teaching people how to how to do things like hey hey here I go quickly from here’s the user interface look at all the steps and by the way

51:52 look at all the steps and by the way check out this Advanced button right here that like bam they go oh my gosh that’s code yes right and and then it’s like copy this paste it into your other report right and then like change the data sort bam and it’s like bam all there and they’re like so the Reus the reusability right of M code or you reusability right of M code or the the sometimes know the the sometimes extremely challenging transformations of

52:23 extremely challenging transformations of data is the value there right and I think what what Steen calls out in here is that there’s a power query template that allows users to easily leverage

52:37 that allows users to easily leverage that M code and then points to a Chris web article where it’s easier more than now than ever to reuse that that transformation code in between data flows or from Excel even to take it out of there and use it in a data flow Etc and I think the the call out here is just that’s it’s a it’s a time well spent whoever developed the the template for that because those those are once again huge Time Savers of saying I did all this

53:08 Time Savers of saying I did all this over here but like can I just move this this basically this giant pipeline of transformation data into a different flow and you absolutely can do that now but again and this is the last thing I’ll say there’s still that same roadblock piece right now the only way that you can utilize templates is upload from your local computer there’s we’re still lacking I think in all these Solutions and to do that circle around or wrap around everything that’s negative well I’m giving you I’m

53:40 that’s negative well I’m giving you I’m saying that because you guys point that so P go ahead sorry no I’m you asked the original question why don’t do we think that we’ve really promoted this I’m G you an answer at least my answer well okay so yes I agree with your point Tommy there is no there again something that’s missing here is there is no Central store integrated with what this is right to your point Seth and to your point Tommy it’d be really nice if the power query template files had the

54:11 power query template files had the ability of saying I’m just going to have an artifacts folder inside my workspace it’s literally a folder where I could say upload my theme file upload my templates upload my whatever whatever whatever I want it goes into the workspace and then hey let’s just say for example I want to I do want to muck around with power query templates what if I could just export them out of the data flow and just put it in my workspace folder and then come back and go get it I don’t have to nothing comes down to my desktop it all stays in the servers one that would be genius so

54:41 servers one that would be genius so there’s another good Microsoft idea for you we’re fixing your tool as we speak right now right now Microsoft so enhancing enhan enhancing enhancing enhancing the tool but to your point though this is also and I I’ll point out my my comment earlier this is also a very self serving feature of Microsoft to build power query template files because what what is this for the reason this is there is to help people migrate from other things into Power query template pieces so it’s a migration effort that Microsoft is producing that that you can then lift

55:12 producing that that you can then lift your old Power queries out of the old service and again I don’t think it’s there yet I don’t think power query gen one has this and data flows yet but I again I would imagine this is not too far off in the future where they could just say oh look here’s a data flow gen one export this P pqt file go over to Gen 2 and upload it so I honestly I think it’s a very self-serving feature here again to get people to migrate to fabric to make it easy for Microsoft to

55:42 fabric to make it easy for Microsoft to store this information and maybe they internally will say upgrade this data flow gen one to data flow gen two and what they’re doing internally is creating the pqt file creating the new data flow and then uploading it there so that could be all experience pieces that Microsoft may be developing behind the scenes I don’t know but it would seem to me that this is what they want you to do is just have an easy upgrade path from a a gen one this is this is why this exists so anyways I will say when doing training classes now

56:13 will say when doing training classes now having the ability to load things using a power query template file is much easier so that to me feels like a really good experience and so here we’re going to connect to this North Wind data Services I’m going to give you all the power query steps load this file boom we’re done we’re all on the same page we have the same information that’s super helpful so anyways really neat things there as well I think we’ve actually beat this topic to death we’ll probably have to have another one of these in the future because we’ve we’ve had so much good

56:43 because we’ve we’ve had so much good conversation already about it but I really want to say this has been a great topic I really like thinking this way about things and feel this is this is wonderful I like it so with that us say thank you very much we should probably do final thoughts here before we really wrap let me not skip our our format here so final thoughts real quick format Tommy let’s do final thoughts what’s your final thought I want to keep it really brief I think templates are great I’m excited to see where they go so I’m not gonna say anything the hot

57:13 so I’m not gonna say anything the hot takes anymore or at least tonight today but no I I think the the new Solutions out there especially from the developer side I really don’t think we’ve touched the potential use cases for that tooling that can be created from this and that’s what I’m most excited to see Seth final thoughts of all the Fantastic uses of templates power bi. tips theme generator does it the best Shameless plug

57:44 best Shameless plug sham with sham with you we’ve worked really hard to make that a Primo experience we really have we’ve tried hard to make it awesome it I I agree it’s a great place to start but I I I love I love all the bits and pieces it’ll be it’s still worth exploring how far you can prean templates to help business to create standards to just enhance the usability of powerbi for users that might not know all the technical pieces

58:15 might not know all the technical pieces I think as Mike pointed out there are certain things that have to be in place for us to like widely push them but at the same time I think it it is now at the Forefront of my mind I’ll be thinking more about how to leverage them and look for the enhancements to do so yeah I think my final thought here is look this stuff exists if you are a technical user you should definitely invest some time learning how to work smarter with these tools and features that exist I think this is and I would also Echo when you’re if you’re talking

58:45 also Echo when you’re if you’re talking about a team that needs standardization and you’re talking about a central team that’s doing this 100% this should be part of your planning road map thoughts and thinking about that this should be part of your you should be having knowledge about this so if you are a leader of a team this is important figure out how to use this and use it in what works inside your team all the new features that are coming out pbip format enhanced report formatting the the the timle format these are all designed to help developers be better at building that certified content and having

59:16 that certified content and having Central teams manage those pieces of content so I really do think that that piece of this story needs more love I think Microsoft needs to provide more support for the community to be able to do more with these templates and build more integration points across the powerbi platform that makes this in general easier for us to deal with these complex issues so I think you’ll get there we’ll see what happens we’ll see what Microsoft opens up and again I

59:46 see what Microsoft opens up and again I know they’re busy right they’re thinking about we want to we want to work on the co-pilot features we don’t want to be working on the Michael’s stupid template feature because he just wants it and it would make his life easier they want they want to do big movement pieces right now and I think they’re focused on some really big closing big gaps in the fabric ecosystem which I get that’s fine there’s not enough people to do everything so I’m sure it’s just very hard to find funding for all these new features that are getting produced there’s not an infinite

60:16 getting produced there’s not an infinite number of people in the world to do all the infinite number of development activities well shocker if everyone would just start buying f64 and stop would be fine all right that’s exactly what we need to do everybody go buy an xf64 even if you don’t need it you can convince your organization to do so go buy it it’s okay you can scale it up and down when you need to but just go buy one actually buy

60:48 two that redundancy yeah that’s a thing right and then we’ll have enough money to fund these other features that we need so that’s that’s what we need so go get more consumption people stop being bashful about what’s going on here anyways thank you very much for listening we appreciate your ears for this this last hour we hope we found some of this informative for you hopefully you’re able to leverage some of these technology and hopefully there’s some some pieces here that stepen pointed out that you didn’t even know existed so hopefully there’s some other pieces there that you can leverage inside your organization as well with that I’m going to say thank you very much we appreciate your time please tell someone else about the podcast we want

61:18 someone else about the podcast we want them to hear about this cool fun stuff that we’re doing we’re passionate about data we love powerbi and fabric it’s changing how we do business it’s changed all of our careers and we think it will change other people’s careers too so find other people who are passionate about this stuff and engage them as well Tommy where else can you find the podcast you can find us on Apple Spotify or wherever you get your podcast make sure to subscribe and leave a rating it helps us out a ton do you have a question or idea or a topic that you want us to talk about in a future episode head over to powerbi tips podcast leave your name and a great

61:49 podcast leave your name and a great question join us live every Tuesday and Thursday a. m. Central and join the conversation on all powerbi tips social media social media channels is that what I said you did said powerbi tips leave it you tips too it’s actually live there too right it is I’m not funny Tommy you’ve been saying the same thing for like weeks now and all of a sudden you like the last two you’ve like cut it off at the end and like Seth and I have had to

62:19 the end and like Seth and I have had to finish it I think we should do that from now you should just Tommy what you should do is you should just randomly stop somewhere in that and we’ll see how well set then I can finish the statement what do I need to change in the template oh it’s funny I’ll talk to you offline Tommy we’ll fix it we’ll fix it no big deal anyways thank you all very much we’ll catch you next time see you

62:54 [Music] later [Music] n

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.

#Explicit Measures #Podcast

Previous

Lakehouse or Data Warehouse – Ep. 305

Next

Fabric Conference 2024 – Ep. 307

More Posts

Mar 4, 2026

AI-Assisted TMDL Workflow & Hot Reload – Ep. 507

Mike and Tommy explore AI-assisted TMDL workflows and the hot reload experience for faster Power BI development. They also cover the new programmatic Power Query API and the GA release of the input slicer.

Feb 27, 2026

Filter Overload – Ep. 506

Mike and Tommy dive into the February 2026 feature updates for Power BI and Fabric, with a deep focus on the new input slicer going GA and what it means for report filtering. The conversation gets into filter overload — when too many slicers and options hurt more than they help.

Feb 25, 2026

Excel vs. Field Parameters – Ep. 505

Mike and Tommy debate the implications of AI on app development and data platforms, then tackle a mailbag question on whether field parameters hinder Excel compatibility in semantic models. They explore building AI-ready models and the future of report design beyond Power BI-specific features.