App Usage in Fabric – Ep.543

A mailbag question about tracking app and report usage turns into the more uncomfortable version of itself: can you connect any report you own to a decision that made or saved the company money? Both hosts have been on the wrong side of that question, and this episode is about the metrics and the process that keep you off it.

News & Announcements

• Fabric runtime release channels — A new toggle letting customers control when Spark runtime updates land, so you can test library upgrades, dependency changes, and security patches before they hit production. Mike’s take is that this matters most for heavy Spark users: a new efficiency feature can save real money on large jobs, and some organizations will happily accept early-release bugs to get it sooner.

• Deep dive into tooltip options in Power BI visuals — Zoe Douglas walks through the tooltip options now generally available, from default visual tooltips to field-well tooltips to report page tooltips. It prompts a good exchange about tooltips getting in the way of the analysis flow — you can’t copy from them, you can’t export them, and there’s no way to temporarily switch them off while you read the data underneath.

• Chicagoland Power BI user group — July 16 — In person at the Microsoft office downtown, with Rayfin demos. They’ll try to capture it, but don’t count on a live stream.

Main Discussion

Topic: Measuring app and report usage in a way that actually means something

Mike opens with the memory that shaped his thinking: realizing that if his boss asked what value he’d delivered that year, “I created 35 reports” was the whole answer — and he had no idea whether anyone opened them. Tommy’s contribution is a definition of reliable that puts trust and confidence at the center.

• Usage numbers only mean something next to the business metric. You cannot judge a sales app by opens alone. You need the sales team’s calls and demos alongside it — if their numbers move and usage moves with them, you’re supporting them. One side of the coin without the other proves nothing.

• Active users over a rolling window, not raw opens. Mike’s default is 30 days; tighten to two weeks if you want to be critical. The question is whether people come back, not whether they visited once.

• Tommy’s benchmark for an org-wide app: roughly 80% of the intended audience should touch it within a two-week window. Anything less and you’re carrying reports that are quietly dead.

• Reconcile your reports against the operational system. Whatever you’re reporting on should match what production says. It’s basic, and it’s the foundation of the trust Tommy is describing.

• “If a report isn’t making or saving money, kill it.” A callback to a position they took years before Fabric existed, and it still holds. The hard part is quantifying which decision a report supported and what that decision was worth.

• An org app needs a stated goal before you publish it. An organizational app is different from a pile of reports in a workspace — it affects the whole business, so it needs a target number attached to it from the start.

• Write down why the report exists. A single sentence on the report itself, stating its purpose. It aligns the page count, the measures, and the eventual usage conversation, and it gives you the baseline for measuring anything.

• Plan for sunsetting. Junk accumulates in every organization — that’s not a failure, it’s inevitable. What matters is having a process for identifying what isn’t used and turning it off.

Centralized, business-led, or federated

The right usage metrics depend on which structure you’re running. A full center of excellence builds everything centrally and distributes it. A business-led model pushes creation outward. The middle path both hosts favor is federation — find the sharp person on the ops team who understands modeling, give them more exposure, and let them own content for their department.

Scorecards close the loop

Tommy’s point on this is that a number has to make someone happy or sad — it has to invoke something. A scorecard puts that number in one place a person actually looks. Mike ties it back to his own GitHub spend: agents mean far more commits and far more build minutes, and the cost only becomes manageable when someone sees their own figure on a scorecard rather than an aggregate nobody owns.

Looking Forward

Pick your most-published app, work out its active users over the last 30 days, and see whether you can name the business number it’s supposed to move.

Episode Transcript

0:02 Open up the sky. Dance to the day to laugh in the mix. Fabric and A. I get your fix. Explicit measures. Drop the beat now. H feel the crowd. Explicit measures. Hello and welcome back to the explicit measures podcast with Tommy and Mike. Tommy, how you doing? Mike, I’m doing great. Happy 4th of July for those who are American who are listening. There we go. It’s a big holiday for us.

0:33 There we go. It’s a big holiday for us. We have some other countries celebrating with us as well. So, thank you though other countries who are celebrating our 250th anniversary this year. So, that was fun. But, moving away from that, the holidays and breaks and it’s good. Did took some time off this weekend. But, more so than that, soccer or as the rest of the world calls it, football. Football. We in America are really want to play well with the big guys, but Tommy, we cannot keep pace., it’s the equivalent of my

1:04 , it’s the equivalent of my five-year-old trying to play soccer with my 9-year-old. It ain’t going to happen. You can try hard. So, hard. So, yeah, it’s like the the the genetics of other countries are just like more built for soccer than even us in the US. Like, we don’t have the capability to even compete. So, this is not me being unpatriotic here. If you’re listening, you’re thinking, “Oh, I can’t believe Tommy never believed in the US team.” Let’s let’s understand how it works here. All the guys who were playing for the US soccer team, maybe some of them were in the US academy, most of them did travel ball,

1:34 academy, most of them did travel ball, high school,, in college. All these people were playing against especially in Europe. They’re in soccermies like AC Milan, Liverpool. They sign people when they’re seven years old. and then you do your elementary, your junior high, and your high school if you don’t get cut from the team all throughout in this soccer globe. That’s like the president of the United States buying the New York Giants and then signing Pop Warner guys throughout their high school career for football and just grooming them to eventually play for the New York Giants.

2:05 eventually play for the New York Giants. Like that’s what soccer’s been for ever in the Europe and South America. Yeah, it it feels like we’re trying. So I what I will say though is I’m very excited that the US is doing a good job hosting. There seems to be a lot of like fun news there. I’m really enjoying football at this next level. And so I’m 100% [snorts] enjoying watching. I’m still going to watch the rest of the the World Cup. I liked how many games there were. I think this has been a lot of fun. So very very excited about the rest of the series. and now I got to find my next team that

2:37 and now I got to find my next team that I’m going to root for. so if you have your favorite team from playing in this year’s World Cup, we want to make some predictions. Let us know in the chat. Send us in who who you’re rooting for. who is your favorite team that you think will go all the way. We’ll take some predictions here. Not that we’re going to do like a real prediction market. We’re there’s no rewards for this. It’s just No, no, no, no, no, no. Even though there’ve been enough commercials for that. that. Oh my goodness. All over the place. What does this remind you of? Hm. Yeah. I just remember like the Super Bowl a

3:07 I just remember like the Super Bowl a few years ago with the Bitcoin or whatever the BTX. Not that’s going to happen. I’m not going to go that way. But do you have a team right now that you find yourself leaning towards now the US is out? Well, I thought the the England game against Mexico was quite phenomenal. Mexico clearly outplayed them and that was a really good game to watch. There’s been so many games even just now just watching like a lot of like shootouts at the end, sudden death kind shootouts at the end, sudden death places that are just really fun to of places that are just really fun to watch. So, I’m probably we’ll see.

3:39 watch. So, I’m probably we’ll see. I’m probably going to root for England and we’ll see what if I, we they seem like they’re kind , we they seem like they’re not an underdog, but of not an underdog, but like they were they’re definitely outplayed in that game and they came through. so, very interested to see how the rest of the tournament turns out. A lot of people are on the Norway bandwagon. I don’t know if you see what they do after they win, but it’s one of the coolest things. Do you see that? Yeah., is it is it Brat is his is his name? That’s one on the Holland. Holland. Holland. Yeah. Holland on the team

4:09 Yeah. Holland on the team that. So there’s And I I will also argue like of the countries I’ve been to, I’ve really enjoyed some of the Nordic countries, Denmark, Sweden,,, you countries, Denmark, Sweden,,,, Norway, very very awesome know, Norway, very very awesome countries. Really enjoyed those. And they have they have they seem really fun. Like if you went to a game with them, it seemed like you’d have a lot of fun when you went to the Norway game., literally, they have the drum and people are going like the they’re doing the freaking Vikings. They’re Vikings. Vikings. Well, and Tommy, in America, we’re very

4:39 Well, and Tommy, in America, we’re very much like we’re proud of like our innovation and the industrial revolution. There’s a lot of like, you revolution. There’s a lot of like,, innovative things in like museums know, innovative things in like museums and stuff around the US like very science driven and like, hey, we’re we’re building net new things. you go to Norway, they’re extremely proud of like their their heritage around traveling the world and a lot of like seafaring adventuring and going over Antarctica and and exploring over the the North and South Poles. Like it seems to be like that was like a really big moment in their in their heritage. And

5:10 moment in their in their heritage. And so there’s a lot of like rowing and being on the water vibe that goes on there, which is super fun and very awesome to watch. This has been fun. All right, Tommy, what what do you have? So, our main topic today, let’s talk about that main topic real quick and then we’ll come into some news here. So, our main topic today is tracking app usage. So, we have org apps now and we have old workspace apps and now we have Rayen apps. That’s another one that’s in there. Tracking app usage inside

5:40 there. Tracking app usage inside Microsoft fabric. Why we should do this? how do we do it? What’s what does this look like for organizations?, and you how do we know if these apps are being adopted? Well, are they working the way we expect them to work? So, we’ll unpack a little bit of that topic today. We’ll probably talk about or throw in some agentic something somewhere in there,, which will be,,, that’s every episode nowadays. But, Tommy, before we get into the main topic, what do you have for us for news today? Yeah, we got three items. One is an announcement. If you are near the Chicago area, somewhere in the Midwest,

6:10 Chicago area, somewhere in the Midwest, well, guess what? Next Thursday, July 16th, we are having the next Chicago Microsoft Fabric and PowerBI user group. We’re doing it in person at the Microsoft office downtown. Mike, are we going to be doing live stream, too, or is that just too much to do? So, we are going Well, I’ll bring some of the gear. I have the gear for it. So, we’re going to attempt to do as much as we can. We So, I would argue don’t bank on us live stream it. We may record it and publish it after the fact. So, we’ll

6:41 and publish it after the fact. So, we’ll see how it goes. I’m actually probably not going to put a lot of emphasis on doing the stream of the actual event,, just because we want you to show up. We want you to be there. We want you to be present.,, enjoy come up to Microsoft Office. Tommy has waved. So, in in t typical,, Microsoft events or Tommy do them at the Microsoft office. Traditionally, they have had a $3 kind Traditionally, they have had a $3,, just get the door fee. of,, just get the door fee. Tommy and I decided, look, we’re going to wave that. We want everyone to sign up that you can possibly sign up. Make sure you sign up and come to the event.

7:11 sure you sign up and come to the event. So, we have waved that $3 fee. Make sure you just go to the page that the link is in the description and also in the chat window as well. Go there, sign up. Sign up if you’re coming because you need to be let into the building securely. Bring an ID., but we’ll be there., you want to hang out with us in Chicago, sign up and join. Should be a lot of fun. So again, the meetup link is on the podcast and it’s also in the YouTubes and it’s anywhere you can find if you go to meetup for the Chicago user group. All right, so that’s the first announcement. Mike, we got

7:42 the first announcement. Mike, we got some two fabric blogs. The first one is about Microsoft fabrics is introducing fabric runtime release channels and these are new a new feature or a new toggle so to speak that allow spark so customers can control their updates with certain workloads and test changes early before they hit production. so these are certain things and again we’re talking about spark runtimes here in Microsoft fabric library upgrades dependency changes gear security patches

8:13 dependency changes gear security patches and even their operating system updates on the back end and Microsoft knows that there’s no really test matrix to cover all customer scenarios. there’s you all customer scenarios. there’s there had really no preview before know there had really no preview before what came the default. So right as of right now there are now two types of release channels. This is very common if you’re an office admin. We have the default channel, the production grade runtime everyone uses automatically. And now there’s also the early access channel, a production grade preview for upcoming updates for validation.

8:44 upcoming updates for validation. So Mike, this pretty cool and I like this because I think Spark has become so integral to really what Microsoft Fabric is all about and a lot of people want to test the beta things. A lot of people are also coming from other environments where some of those features are only in preview. So Mike, what’s your take? Yeah., Tommy, looking at this one, right, and this, I think, makes a ton of sense for us,, to have these release rings. As you were saying this, Tommy, you’re talking about it. I’m always guessing, and this is probably just something that I I see or observe, right? Up until this

9:14 I I see or observe, right? Up until this point, it’s been like a gated release, like there’s nothing out there. Maybe something gets into preview and then finally gets released. Some customers may want to start working on,, staging those changes, right? it. One of the advantages of having like a pre-release particularly around Spark is Spark keeps getting better. It keeps getting more efficient. There’s always like these new features that are coming out and people always want to run and use the latest features. I think Data Bricks does a very similar thing where they release a lot of features very early. They give out a

9:45 features very early. They give out a early release ring, right? So bring in that audience earlier. I think this does two things for the companies that are more interested in trying out the latest greatest features or hear about Spark changes, they’re able to get to those Spark changes faster than they usually would. So, one, that’s really exciting because now if you’re on the cutting edge or you’re trying to build something new or you want to use all these greatest, latest features of Spark, they’re there. They’re going to be there sooner than than before. But that’s a decision in the business. they can decide how much

10:15 the business. they can decide how much risk they want to take on for those new features because stuff likely, well, not likely, I’ll say that back. Yeah. Yeah. Code you had that exists previously could start breaking because of these

10:27 could start breaking because of these early changes. It increases the risk. It increases the risk, but some companies are okay with that risk and they want the latest features. And you they want the latest features. And, Tommy, if you’re running a really know, Tommy, if you’re running a really big job and a feature comes out in Spark that saves you a good amount of compute, you could be in a position where you could be saving thousands, tens of thousands of dollars potentially on jobs that are running. If you’re running a really big system where you’re running lots of Spark, that that you’d want that as soon as it came out. As soon as it’s released, you’d like to start trying and who cares if there’s a couple of bugs here. You can solve that

10:58 couple of bugs here. You can solve that problem, right? So, I think I think there’s also some advantages there on that on that front. The other side of this coin is there is like the stable releases still. So you don’t have to opt into this. This is there for the companies that are financial and need that regular rigor of testing. They can use these preview features in a test bed and then only approve them and only move them through after they’re ready to go and really officially released. And I think Microsoft at the end of the day they get a lot of extra testers on this stuff.

11:29 a lot of extra testers on this stuff. Yeah. Yeah. Right. That’s the So, in some way, I feel like these early release rings are like a privilege to be in, but it’s also like you just saved Microsoft a ton of money because now you’re doing the testing, you’re finding the bugs, and you’re submitting them back to a feedback channel so Microsoft can go fix and resolve them. So, it also feels like to some degree these early preview sites are actually Microsoft lifting the testing and putting it on you as the as the product owner. It’s an enterprise CIA. That’s what that is. This is what that is. So

11:59 is. This is what that is. So that’s probably true. Yep. Enterprise. No. And I’m glad regardless, I like it. I I I do too. So the next one, this is from Zoe. Nothing new, so to speak, here, but it’s from our friend Zoe Douglas at Microsoft, and it’s really a deep dive into tool tip options in PowerBI. We’ve talked about tool tips a little, but let’s talk about the really the main ones here. We have the default visual tool tip tool tip field well the tool tip fields only in a sentence format something that’s

12:30 sentence format something that’s generally available now newly generally available and then the report page tool tip tip and there’s a help icon tool tip as well so Mike we’ve talked about tool tips I don’t know if we’ve talked about the essential part of it here but there’s a few things here the new sentence format one with a global hierarchy which is neat Sure. But let’s actually dive in, Mike, with these enhancements, but also just tool tips in general. Where would you prioritize tool tips in terms of making

13:02 prioritize tool tips in terms of making or breaking a report or visualizations in PowerBI? How essential is this for your customers or for consumers to know? Let’s start there., Tommy, this is interesting that you say this because I’ve actually seen, tool tips are interesting. They work well when I believe you’re on like a desktop application, right? If you’re on PowerBI desktop, it seems to work fairly well. I’ve had reasonable success using tool tips like on my iPads and things like a smaller a little bit smaller form factor of a screen. When I

13:33 smaller form factor of a screen. When I get to mobile, they don’t make sense at all honestly. And so I’m even trying to like So I I’ll just say this. So, there’s been a couple reports that have been built where I think sometimes the tool tip is a bit aggressive and I it actually hinders me from actually clicking on what I want to click on or see what I want to see on the page and maybe I’m move I’m just having to move my cursor around the page to look at a table or some other information in there and all of a sudden it’s got a tool tip slapped on top of it and I’m like what what’s going on here? Like this seems a bit egregious.

14:04 Like this seems a bit egregious. So, egregious work for this. Well, it it just seems excessive. Like I Well, it it just seems excessive. Like there’s no there’s no ability for mean there’s no there’s no ability for me to like in the visual or on the page to like say look I I’m happy that the report builder builder built the bookmark inside or the the tool tip into the report. But sometimes I feel like I just don’t want that right now. I just want to look at the data and see what’s there and not always be annoyed with an aggressive tool tip that’s always popping up in my table. So I kind always popping up in my table. So I have a lovehate relationship with it.

14:34 of have a lovehate relationship with it. I like it in some some ways and other ways I don’t like it. It seems very purpose-built, right? A lot of these bookmarks are very specific. Some of the examples that Zoe gives me that I like to use., not to be any, I’m I’m not going to u deny this, right? One of the tool tips she uses is another page that has a line chart on it for a bar chart. So you’re on a report, you’re hovering the bar chart and you just cursor over and you get the historical values of data for the last,,

15:05 values of data for the last,, couple days or the price changes over time, right? That’s useful. I think I think that’s helpful. I think it’s part of the effect of how the report is used, but there’s no global way to like, okay, I’m I don’t really want to see that tool tip. Can you just disable them for a hot second? Let me just look at the data that’s there. And so that’s the part where I get a little bit more like it feels like it’s in the way more than it’s helping to me. And it’s interesting that you bring up the report page tool tips because you mentioned the egregious part. They the report visual tool tips can be very very

15:36 report visual tool tips can be very very aggressive. I think with tool tips I I I see the value and I always try to be very intentional with tool tips, but they’re tough. I think if you do it the wrong way or I think without the intention where you really lose a lot of context. So, I’ll give a good example here. On any bar chart, you can add as many tool tips as you want, but I always try to make sure that everything’s related. For example, if I’m looking at sales year to date, well, I’m only going to I don’t want to add everything. Just

16:06 to I don’t want to add everything. Just because you can’t do something doesn’t mean you need to do all of it. So, I’m only going to add, hey, sales year to date. That’s the obviously the default last year’s numbers and year-over-year and add those as the main tool tips there. When it comes to the report page tool tips, yeah, the same thing. I’m very selective when I choose to do them. When they first came out, I don’t know if you were like me, Mike, but it was something I was trying to do to every report visual. It’s like, this is so cool. Everything’s going to be helpful. And then you realize that it could take people really out of the

16:38 could take people really out of the what they’re trying to do and more more the analysis flow or like yeah, the analysis flow that’s their own logical flow. you’re looking at the table and you’re just trying to turn in,, looking at the bar chart, you’re just trying to turn into a table so you can see what data is there or and or you’re just trying to,, copy a number off of a table somewhere and the tool tip keeps showing up and you’re like, what? Wait, what’s going on? Like, I just want to read the data that’s behind the tool tip. I don’t want the tool tip right there right now. It’s almost like I need to like temporarily disable them while I’m looking at stuff like, no, this is not what I want. Thank you, tool tip, but that’s not the data that I need at this moment. So, again,

17:08 that I need at this moment. So, again, it’s a lot of preference on the user creating the report. And, it’s funny because I took the approach of creating reports for myself again, like my own cycling data, my own like kind my own cycling data, my own like local testing data. And when I was of local testing data. And when I was creating reports for me, I realized when I wanted certain tool tips. For example, my I I was looking at my cycling data. I’m like, “Oh, cool. I could see a line of my spikes, but that’s not what I want to see right here, right?” Even though that’s cool, I just really want to see the numbers here because that’s really what I’m trying to

17:38 because that’s really what I’m trying to dive into. So it’s always tough building tool tips without again being the eye of the beholder or actually being the one under the microscope. This is always part of my sandbox phase. I don’t create report page tool tips without going through with the user saying this is an option and unless they’re really excited about that or they actually suggest something, it’s probably not something I am pushing or implementing without that because again to your I think you said it perfectly, it can take people out of that analysis flow. So tool tips though,

18:10 that analysis flow. So tool tips though, they’re important. I still think they’re I still think they are they’re essential that you do it right. I’m going to agree with you there, Tommy. I think that’s that’s definitely it’s a judgment call, right? It’s a judgment call on what the users want. I’d almost it’s almost like a feature like I would want to show people that it exists. exists. Yeah. But then really make sure I have a defined requirements around like what what information actually needs to be

18:40 what information actually needs to be shown in the tool mark. I’d almost argue, Tommy, in favor of tool tips, I would rather use like a drill through feature instead, right? Hey, here’s the bar. I can see the data. It’s highlighting the way I want. I want to understand more information about that bar. Right click, drill through. It’s a little bit more clicking to get there, but I think you can actually do a much better job catering like the messaging around what exactly you selected and and like that again the story of like I’m starting here on the homepage or this landing page. What is the next step of

19:10 landing page. What is the next step of the story? Where what am I walking into to get me to the next area? Psychology and technology is amazing. I don’t tell me if you ever felt this way because for me Mike this is frustrating because I feel on paper that tool tips should be the most essential thing probably like the second most utilizing thing PowerBI seems like a great feature but more often than not I’m finding myself out of the flow with it or users are and it feels almost illogical like it it defying the reason for the tool tip. I

19:40 defying the reason for the tool tip. I don’t know if you’ve ever felt that way but for me I love tool tips. I try to implement them and then I realize that they’re not using them even though it feels like it should be such a supplemental part of the report. You you can’t copy the text from them. You can’t export them. You can’t like It’s it’s it has to be there. Like mean it’s it’s it has to be there. Like every visual program that you have has some tool tip capability with it. Like it just has to be part of it, right? But I just I just think the use cases of it or the places where it makes the most sense or where it’s like best utilized is is a bit smaller than I would like.

20:10 than I would like. And again, that’s just probably my preference. If I had more control at the report page level or settings at the report level like, hey, turn tool tips on, turn to tool tips off, then I might actually put more tool tips on the report page and then let users decide, do they want them on or off? I’ll just have them there by default, right? So that’s that’s something else that would be interesting to unpack as well. Okay, great. Good topics. for both of those two topics, I will put them in the chat window and they’re also in the description as well. So, if you want to go check out more about these tool tips and the announcements from Zoe

20:41 tool tips and the announcements from Zoe that she was presenting here, want to see how those work, feel free to go check out the chat window. You’ll be able to see them there. Tommy,, jump us into our main topic today. What’s going on? going on? All right, so we have a mailbag here, and I’m going to ask people as you

20:54 and I’m going to ask people as you submit mailbags to one, don’t use AI, and then two, try to keep it to a single question. We’ve gotten a few mailbags lately where there’s like bullets of questions. Mike and I are gonna talk about probably three sent three words in a sentence for an hour. So, you give us six sent six questions. We’re not going to cover it all. So, what we’re going to do is we’re going to break down I’m not going to read the whole mailbag here, but we’re going to talk about I think something that’s so important and it’s always going to be important. It’s about Microsoft app usage report usage. And

21:26 Microsoft app usage report usage. And I’m going to take the main question here. It’s really what’s the most reliable way to measure PowerBI app usage over time, not just individual reports when you’re distributing content via the app experience. And Mike, I’m going to just set us up here because we’ve talked about usage before and the importance. I don’t think we’ve ever argued that, but I’m going to take a few keywords from the question here so we can reframe our conversation today. Sure.

21:56 Sure. what is the most reliable way, not just should you do it and then how do we actually do it over time. So, and I think there’s those are two things I want to break down with you today because there is obviously the the foundation of should you do app usage, should you do report usage? Yes. Should you look at it? Yes. Now, how do I actually trust it and what’s the best way to look at it is one se area that I’m looking at and then over time I want to take you on a journey here and not

22:26 to take you on a journey here and not just say we’re just going to track historical data but how should I be looking at the usage over time in terms of should I be tracking the growth is only that matters because the reason we’re tracking this is not just to say our numbers have grown hopefully you have tied this to what adoption is for your organization. What are the goals of adoption? They should reflect in your app usage. So there’s a ton of things here. I’m going to just start with again

22:56 here. I’m going to just start with again those two ways. What is the most reliable way and then we can go into over time. So Mike, in your own experience that let me hit the ball over the net here and ask you when you look at app usage, you look at report usage, have you found a tried andrue way to make sure that you first just got the right numbers. This is always it’s not tricky. Let’s be honest, there is real answers to this, right? right? Yes.

23:26 Yes. But as we all know and as I’ve done lots of these engineering projects around how do we build things what goes where what where are we loading this data to it’s quite complex to understand how we’re able to weave all of the data from your source system through some data engineering provide that business logic and get it out to the final answers. So there always needs to be a process on how you quality check your data regardless. Doesn’t matter what you’re

23:56 regardless. Doesn’t matter what you’re doing. There’s some process. Go to the operational system, grab a customer, see total sales. Your reporting should match what’s inside the actual operational system. How you’re tracking billing and and items there, right? I have a system where I track my usage of my team and what they’re working on and how many hours they’re they’re doing. I have a whole reporting system that reports on that information. Well, I need to make sure that my reporting system matches what’s inside the summary. It’s inside production. So, that that’s something that we do

24:27 that that’s something that we do internally and that way you have vetting of those numbers. So, I’m going to say as long as you have a process in place where you’re able to vet the numbers, that’s that’s the most critical thing for making sure that you actually have real data coming in. I’m not sure if I even answered your question, Tommy, but maybe went too far. It’s a it’s it’s a good start in terms of obviously there’s I think if I’m hearing you correctly here, it’s no matter what work you do at the end of

24:57 matter what work you do at the end of it, you need to have a way to in a sense measure if you’ve done it right, if you actually have met some expectations or goals. If you’re going to build an app, you want to make sure that it’s going to provide some value. Well, and let’s talk maybe more about the app thing specifically, like how do how do we know what’s working there? Just because you have more people touching the app, does that really mean you’re actually answering the question? Is that really meeting the need? Right? Are what are what’s the story of how people are using the data that

25:27 how people are using the data that you’re producing through the app? Now, regardless, I would argue that there are some other metrics here about how many people are logging into powerbi. com., are are we expecting people to create their own workspaces? How many workspaces are being created? how many reports and or semantic models are being generated across the different their platform or different teams that you’re you’re leading this with. So there’s a lot of questions I think you can ask here that you would want to write down and say what are the goals of your team? How do what what is the expectation that

25:57 How do what what is the expectation that you’re trying to deliver here? I would also argue for teams that are just starting with PowerBI, you want to think about the persona of the people that are consuming things from powerbi. com, right? are you I think there’s definitely Microsoft has this very clear data engineer and then or analyst and data engineering like that’s like two key roles I see right right but I think the data analyst actually breaks down into like many more roles or or more nuanced roles that are a little

26:28 or more nuanced roles that are a little bit more specific an analyst could be someone who just shows up to a pre-built report and uses said report to do some work on data that’s existing in a model like they’re not doing the modeling of Other people define analysts as the data modelers and report builders and they’re that’s in the analyst role. And so I would argue sometimes I feel like some of the data engineering experience that I’ve h I’ve gained over these years and I have been doing that actually lends like even though I’m I would consider myself more data engineering than I am

27:00 myself more data engineering than I am analyst. analyst. I still do a lot of the tasks across both of those roles, right? So I look at this going what does your consu define your consumer of stuff? What do they do? Because even just saying what is the consumer of PowerBI content that can mean a lot of different things to many different organizations. One one I’ll walk in they’ll say oh we’re only allow we’re only letting them use and look at reports. Another organization will say well I want them to look at reports but I also want them to build their own reports off of models I provide them. A

27:30 reports off of models I provide them. A third may say, I want to share shortcuts to lakehouse tables and let them build their own models however they want or I’ll give them a sample model to start with and they can customize it from there. So there’s different levels of like what permissions you give to that analyst. And so that’s why I want to define like more specific roles in there and I think that’s where you want to sit and think about what does success look like for when you deploy the apps. I I love where you’re going with this because I think this is I why I think our podcast is well because I think I’m

28:00 our podcast is well because I think I’m on the other side. I’m more of the analyst but also doing both tasks but I think that’s my other side of the brain. Mike, I I’ll take you back a little here because I want to talk about the value here and I think you’re really touch Okay. Yeah. Can we unpack what you mean by the other side and the analyst? I’m not sure if I really understood your point there. You said you’re more of a data engineer but you also do both tasks. I would I would think more like my strength is in like data engineering, pulling data together, joining thing like modeling and and like the big

28:30 like modeling and and like the big picture of like the enterprise data system, right? That’s how your brain tinkers first. That’s that’s where my brain goes towards. And so I can also build down to like semantic models and reports, but like there’s also this like large architectural piece of this. Some organizations don’t push people to directly move into that area extremely fast or or they’re not,, [snorts] they’re you talk to them like, well, our users are not going to be able to make reports, right? right? They’re just not going to understand it. They’re not going to get how the model

29:00 They’re not going to get how the model works together. They’re not going to understand how to build visuals and things. They just want to get to a report, select some filters, and consume the data that’s there. Right? That may be where your company’s at or a large majority of your consumers are. Great. How do you want to distribute that content? Do you want to give them access to workspaces? Do you want to simplify it? Do you want to give them only apps? Do you want to have shared apps and land them on a SharePoint page and just say, “Go here for all your reporting. Click the app you need and you get into the app.” So, there’s this whole distribution method and like how you want to distribute this content. If

29:31 want to distribute this content. If you’re at that level where you want most of your company to not build inside powerb. com apps are the way you should measure adoption and the way you should distribute. And I would even argue that’s where you should start. You should start there. And you should have a very small subset of people that are doing the actual build of things. And then as the company becomes more knowledgeable, you open up to more experts in departments, experts in domains.,, we’re going to talk to the ops team. Okay, Tommy’s in

30:02 talk to the ops team. Okay, Tommy’s in the ops team. He’s really sharp. He understands how to build some models. We’re going to expose a bit more to Tommy and federate part of our center of excellence into that department. Tommy will take some more ownership around building and creating content for their team. team. I wanted to make sure we touch on that. But before we before we go down that road, I want to touch on too on the the side of the reliable thing. And I you’re I know you’re touching on this but Mike when we think about defining reliable and what what I was trying to

30:33 reliable and what what I was trying to mention was I think I have more of the analytic brain just from my background. It’s data is driving behavior. Data desri describes behavior that’s the sociology background that I have and that’s when I anytime that I look at numbers I’m always starting with that. For example, I think to your point, just because app usage is high is not just a critical number in itself. Well, what is that actually explaining? What is that actually telling me and the value? And I was actually while you’re

31:03 value? And I was actually while you’re talking before I actually looked up the definition of reliable and let’s not make the mistake of defining reliable as just accurate, right? Because I think this is a critical juncture here. And I think you really are touching on this in terms of how are we actually looking at this? So what does it mean reliable

31:22 this? So what does it mean reliable person information that can be trusted and that reliance or confidence may be placed is trustworthy, safe and sure. I want to double down on the word trust and confidence. Yeah. Yeah. Yeah. Yeah. Because I because I believe I believe that’s really if you give out a report or even an app for that matter and you’re not able to justify the numbers numbers Yep. and the teammates are are looking at it and going, I don’t I don’t know if this looks right. And and you’re going to have some questions like that.

31:52 to have some questions like that. There’s going to be questions that come in like this number for this customer doesn’t seem to look right. And so, we need the ability for like consumers of our apps to push back on us a little bit and say, I don’t know if that number makes sense. Let’s let’s I I think it should be this. You say it’s that. Let’s go in and dive in and make sure we’re actually talking about the same thing. And I think there’s there’s a little bit of this is where the center of excellence I think comes into play. This is where, hey, we’ve done rigor, we’ve done testing, this is what we’re going to do. As you’re releasing data through apps, I I think the

32:23 data through apps, I I think the expectation here is by the time you get to app publishing, we’re talking like certification of stuff. There’s there’s an internally documented process of like how do we vet these numbers or or and how do we make sure that they’re right so that way when we give them to our users we can trust the numbers that are there are there right and I think the trust though too is I think about this in really from a BI team or the people responsible what are you getting measured against or what are you getting measured for Mike I I

32:54 are you getting measured for Mike I I remember a critical moment or seinal moment in my career when I was building reports and we weren’t really looking at app usage. And I remember thinking if my boss came to me at the end of the year and said, “Okay, what value did you do this year?” I’m like, “Well, I created 35 reports.” Okay, did people look at them? I don’t know,. And I thought about that. I’m like, that puts me in a bad spot,? So, I remember even from an individual point of view, if I’m going to create all these reports because I got requests and no one’s looking at that, first off,

33:24 no one’s looking at that, first off, not great for me. What value am I actually providing for the company? And I think when you look at the org apps too again that’s make the assumption that’s a team of people doing that well what are they getting valued against are we just looking at well we want people to use the or app just views right just the usage is it that the whole team is looking at this because I think you and I have matured on how we look at usage where some reports do not need to be looked at at every day some are a once a week thing some are just meant for

33:54 week thing some are just meant for meetings so just because their usage is lower than the average does not mean it’s a bad report. And I think you were talking really talking about Yeah. And I think so when we talk about trust, those numbers may be accurate, but if I’m just comparing that against all the reports or all the apps, well, I you’re missing the point. And again, that’s I think losing a little trust here. So you have to start with a place of what are you trying to actually measure? What do you from an app? If I’m going to publish

34:25 you from an app? If I’m going to publish an app and I’m going to have all these reports in it and I’m going to have notebooks in it because I can do that with an org app. What am I actually trying to provide here? What what’s our goal with it? And to me, org apps are different than just having a collection of reports like in the workspace where there should be a goal for there should be some number that we’re trying to reach. I I’ll stop. Yeah. Yeah. I’ll pause there because I have an assumption here with what an app should be. But let me

34:55 with what an app should be. But let me pause there. I I think I hear your assumption, but also I’ll look at this, Tommy, and go it also sounds to me like you’re talking a lot of this is a single team that’s producing all reports for the whole organization, right? It it also feels like the way the way you’re using the language here, it sounds like you’re funneling all reporting through the center of excellence. They’re building all the reports. They’re building the apps and they’re distributing out to the organization. I don’t think the second part, but yeah. Okay. But I I would that’s one approach I think of many approaches on how organizations want to leverage and

35:26 how organizations want to leverage and use PowerBI. I do think there’s this idea of the center of excellence. I do think that makes sense. there’s also this need to have leadership involved here so that the data that we need to get into fabric right from our source systems whatever wherever they may be onrem in cloud SQL databases Oracle who cares where it’s coming from we have to have someone unblock barriers so we can get access to that data and pull it into into fabric and powerbi because then then the reporting solution can take off

35:56 can take off right right but where I I think there’s many company structures here. And again, you could go two two ways. There’s there’s two main directions. You can go enterprise deployment only. Everything goes to a central team. All reports are built there. Everyone gets the reporting from that central team center,, full center of excellence, BI centralized team. And then the other side of this is businessled. There

36:26 side of this is businessled. There really isn’t a center team that’s defining standards and things. You basically say, “Okay, team, here’s PowerBI. Go use it. Go find value from using PowerBI across the organization.” So, you give it to everyone and you let them explore and build their own workspaces. I think both sides have their advantages and disadvantages. I think giving the business full control to build whatever they want gives them the ability to fail fast, the ability to go find what works for their team. It may not and it may not be the same as every other team.

36:56 may not be the same as every other team. It does create problems where you have potentially many different stories of the same data, conflicting metrics and repeating values and data that’s being stored multiple times when you give it out just widely to the business. So I really subscribe to like a middle ground approach, a federated approach. And I back to your point Tommy I was reacting to like your comments around well what if I build a bunch of executive reporting does that carry more weight like the Say that one more time.

37:26 Say that one more time. How do we how do we quantify this report was used to support this decision and then this decision allocated more resources, more money, figured out more materials, was able to find problems., how do we allocate that report to to dollars that the business makes? And I want to go back to our comments we made a long time ago back before we had fabric around, which was if a report isn’t making you money or saving you

37:57 isn’t making you money or saving you money, money, Yeah. Yeah. it probably shouldn’t exist. There’s probably any correlation you would try to do would be poor. I I think right off the bat if you said, “Oh, our numbers went up with report usage and then the next quarter our sales went up.” You can make that argument, but I don’t think that’s a really accurate or or reliable way to actually track usage unless you actually had something more direct. That could easily be one of those perist correlations. It’s it’s I’m going to tease myself here a bit because it’s easy to say that phrase of is this report making me money

38:27 phrase of is this report making me money or saving me money? Right? It’s easy to say it, but how do you physically quantify it? How can I tie a report directly to dollar amounts that is you like driving the business one way or the other? other? Can you not how I think the question to me is can you actually I do think there is ability to do that because in my business I run a consulting business and every hour that we spend our time on throughout the week is either billing a client or working on internal projects. It’s one or the

38:58 internal projects. It’s one or the other. we’re only doing one or two things. I’m allocating my resources towards a client work or I’m allocating my resources to app or appdev work that we do internally or internal meetings or whatever. It’s very clear to know if we’re spending our time more on producing or consuming. So, I’m not saying that every report can be done like this, right? right? And again, this is a very simplistic look at,, where do I spend my time? Do I measure all my time? Is it applying to the right things?

39:28 applying to the right things? That’s the level of I think detail that we need to think about things is does this report and and physically I have a report that tells me if I made money this month or I didn’t make money this month and then I have to repivot you month and then I have to repivot my team members what we talk about know my team members what we talk about how we work on to make sure that we’re doing the right thing to change those numbers and and this report’s designed around I know what drives revenue for my business business and so the report is around what that looks like. I was init I think you’re on to something here because my initially

39:59 to something here because my initially was going to go you would have to do an AB test take the report away and actually see numbers go down but no one’s going to do that. However, I think right to your point and I think a lot of reports can do it. People don’t realize it. If you’re just trying to equate report usage with revenue, you’re probably not going to have a really accurate or reliable way to do that. However, However, what has revenue in it?, , well, yeah, but even then too, what if there’s a big product that occurred? You have to set up what’s supplemental to it. So let’s take a sales team, right? A sales team who has a pipeline. They have

40:31 sales team who has a pipeline. They have to call, they have to do,, demos, whatever the case may be. Well, my report for them should show again what keeps them up at night. I always go back to that. How many calls each what are the amount of calls that people should be making? What are the amount of demos that they should be doing on a given quarter? those we have to make the assumption and any company makes the assumption that supports revenue. If if my sales team makes more calls and therefore they’ll do more demos therefore we should make revenue.

41:01 demos therefore we should make revenue. If that model’s in place already that business model’s in place that’s what a report track would you track. So now let’s combine that where the report’s just not looking at the sales team revenue. It’s what are your calls this month? How many demos are you on track for? I would argue, Tommy, I would argue like those those are sensors, right? Those exactly what you’re talking about. Those are indicators to it’s a volume game in a lot of these situations, right? How many

41:32 these situations, right? How many marketing things was made? How many contacts with the customer was made? If you have a really large customer, how many points of contact was made? When did you talk to them? How often are you talking to them? There’s likely proxies for good behavior driving certain outcomes that you want. And Tommy, to your point, I love the demo. I like the

41:48 your point, I love the demo. I like the demo part. If you’re selling software and you’re not doing 30 demos a week, obviously people are not understanding what your product is and how to use it. So, the odds are the higher the amount of demos you get, the more likely you’re going to get someone in that pile to buy and find the value from your software. Case in point, if I have a report that shows my calls and demos for the sales team, I then need to see usage go up and a cyclical nature of I should see report

42:18 a cyclical nature of I should see report usage go up and if the report’s helpful, I should see demos and calls go up. Those are the things you should be tracking in piece by piece. If people if I’m actually showing those right numbers and people are relying on it, then they’re going to behave differently. if they’re behaving differently, those numbers go up. And so I think you have to you can’t just look at the usage of the app and say that’s success. You also have to I think to me tie that a little with what’s critical to that team changing their behavior.

42:48 team changing their behavior. I think there’s also a leadership moment here too and especially around what’s usage of the apps are doing. Leadership has to say we’re looking at this. We’re monitoring this and checking in to let people know like hey good job you increased the number of demos this month and funneling that data back. Okay, an increase of 10% in demos yielded an extra two more customers than normal. Great. Keep doing that. Which then the people who created it go and also the app usage went up. The reports that supported that went up. So we know people are using that. We’re

43:19 we know people are using that. We’re providing our value to the sales team which are then providing their value. So I think I’m having this model here where there you cannot have a reliable way to look at usage without having both sides of the coin. It has to you have to have the correlation or the numbers for the that team along with the usage. One without the other, it’s impossible to say whether or not you’re actually supporting them or whether or not you’re successful with I think your maturity. I’ve been I’ve been exploring a little bit more here, Tommy. And again, I don’t

43:49 bit more here, Tommy. And again, I don’t I don’t know how much you’re into this one, T. I feel I I’m going to project here a little bit. Okay. Okay. Right. I’m going to project here a little bit. Project away. I don’t use them as much as I probably should. Scorecards highlighting like over time. So there’s always like a metric, a KPI that I need to like monitor and get out get my hands around. around. This isn’t I maybe underrated, underutilized feature. Get a semantic

44:20 underutilized feature. Get a semantic model out there. Put something into scorecards. track over time the KPI for Tommy. Tommy. How many demos did he do? I think one one area, let me give you another let me give you another proxy example here, Tommy, that I think maybe is another a really good one. We are a heavy company that uses tons of GitHub. We we’re like all in on the GitHub experience. We have bought all the GitHub features. We’re using GitHub Copilot. There’s a couple APIs that tell you like if you’re doing things and building software and apps, GitHub charges you per minute used on a

44:52 GitHub charges you per minute used on a VM to build the application. There’s a bunch of things there that cost you money to like run and and build applications inside GitHub. Now, with the advent of agents and having agents help us build code, we can commit way more than we ever have before in the past. Make so many more code changes. One thing I found is a disconnect is I the leader of a company is looking at like the expenses, managing what’s going on and so I see all this usage going up. Great. That means we’re being productive. Things are happening. We’re getting stuff completed. Awesome. On the

45:23 getting stuff completed. Awesome. On the other hand, all the engineers that are building on the other side of things have no recollection of these numbers. They don’t understand how much co-pilot they’re using. They don’t understand how much an action is running to build a project or a software piece. So engineers in my team don’t see that information. So the biggest disconnect that I see right now is I need to build some reporting around all the APIs coming out of GitHub and produce a dashboard that is there then shared back to the engineers and say even it it’s got to be

45:54 engineers and say even it it’s got to be in a place where we can easily use it, Tommy. Like just having a report and saying it’s done, go look at it. People will never go look at it. having something that pushes information into your team’s inbox or your mailbox or something that people are going to regularly look at. And yes, there’s still a link there back to the original report, but having that input of data regularly to people’s workflow, that’s important. That’s that’s where people need to physically see like what’s going on. So, if you’re having this disconnect to your point, Tommy, like I’m going back to reports and apps now. the sales

46:24 back to reports and apps now. the sales team. team. If the sales team isn’t seeing the results of getting more demos being closed, those turning into customers, the revenue generated by based on those demos, demos, you’re not giving signal back to that team to do a better job. It’s funny you bring up the scorecard feature here because I feel like you’re you’re throwing a softball at me right now because this has been near and dear to my own heart in terms of the success. So I think a lot of people go they look

46:54 So I think a lot of people go they look at scorecards the wrong way or they start on the wrong side. Typically I I have this program with my consulting firm and we call it the measurement framework and it’s helping companies out not creating the reports but just starting way early on trying to index or calibrate to measure what matters. So we actually calibrate the teams on looking at their normal numbers and then where are those indexes to say where are those pressure points and again so and then we do a calibration workshop basically and

47:26 do a calibration workshop basically and we then start that after we do that we build semantic models just for scorecards because we have to start somewhere and again giving everyone the report to your point can be very helpful but you want you need accountability in order for any metric whether it’s app usage message or demos to actually be effective. It doesn’t have to just be accurate. It has to be impactful to the person. To your point, the eye of the beholder. If the number goes up, it either has to make

47:56 number goes up, it either has to make someone happy or sad. It has to invoke something. And a scorecard is a great way to do that because assigning the user that one quick place to look at it. There are ways to do this, but it can’t just be a given number. Let’s get your example with GitHub. You can actually create a scorec card that would actually have hey everyone’s cost basically or estimate a cost based on what they’re doing or create someone else’s their own individual scorecard. Hey you’re doing a lot of cost on code spaces let’s say you lot of cost on code spaces let’s say that’s where the bulk of your

48:26 know that’s where the bulk of your calls we want your cost for code spaces be under 20% of what is allocated to you. So that’s something that they would look at every day. That’s something I would look at every day. It’s like oh my gosh I did a lot in code spaces. Shoot it’s over 25%. I need to change my behavior. So, you can’t just create a scorecard and expect people to look at it. You have to start and again, let’s go back to app usage again. You can’t just do app usage and expect it to make a difference. I have to tie it to something. It has to tie

48:57 to tie it to something. It has to tie something that has some sentimental meaning to me to the BI team or to that team group of people who are creating it. it. So a lot said there but I think to your point point if you were creating and this is one thing I was going to mention earlier but bodge very well right now. If you are going to create an organizational app, you have to make the assumption those numbers, the reports, the metrics in that report are affecting the entire

49:27 that report are affecting the entire business, right? I I think that’s a pretty easy assumption. An orwide app is going to affect the business. So you cannot create an org app without having some goals for the app usage in itself. Yeah. Yes. I think so. I I think what I’m looking at is like where I would put my my money here a little bit is looking at how many times the report’s been opened. looking at actions from that workspace so that that app comes from an area where there’s a reports and it’s in

49:58 area where there’s a reports and it’s in the workspace. Getting the usage metrics back from that workspace helps you out there. How many times are people exporting the data? I think there’s some goals with the app, right? What’s what’s the expectation? Yeah, there’s maybe a larger concept here, Tommy, that you’re you’re touching on with the app side of things is, and this is maybe something that we miss some of the times or we don’t put as much emphasis on. A lot of times we really get excited about building the semantic model and immediately going into like building the the data set and then the report. That that’s that’s where we

50:29 the report. That that’s that’s where we like clicking on the keys, stuff starts showing up on the page, it looks good, comes natural. But how often do we step back and say, “Okay, let’s actually look at the report page or what’s going on here and say, why are we looking at these numbers? Why does this make sense? Why is this important?” Right? A much more measured approach as to exactly why we’re in here and what’s the goal, right? What are what are our stated objectives of this? And I think if we step back and and before even building the semenic model going back to your

51:01 the semenic model going back to your point earlier Tommy you said what keeps you up at night. What’s what is the thing that you’re doing or that is requiring you to sit back think about why you’re building the report to begin with? Just having that statement available to you. this report exists to stating the stating those facts because then it aligns a lot more of like why the report exists, how many pages are in there and then that also I think measures that’s part of the initial measurement of okay Tommy I’m building

51:31 measurement of okay Tommy I’m building this report to measure sales. Why? Tell me why. Well, we want sales to go up. We want to make sure that we’re maintaining the same level of sales we had last year. Right? Great. That’s the stated fact. Now, how do we tie usage, the report, the analytics back to driving more sales? And once we can really understand and unpack that question, then I think we’re able to able to evaluate whether or not the app is yet

52:01 evaluate whether or not the app is yet is is winning or losing. Should we keep it or kill it? Yeah. And I I think it you mentioned the overtime thing here, too. And I’ll give you a baseline metric. I I’ll give this one away for free. Okay. in terms of you want to look at app usage in a reliable

52:15 want to look at app usage in a reliable way. It’s not just number of uses, it’s active users and depending on your organization, you can,, increase or decrease that rolling window, but typically I like to do 30 days of app usage. So, what are the active number of users? How many people are going back to this report? And again, if you want to be a little more critical here, you can say even on a two-eek window. Okay, our app usage for the last few weeks or the last, you

52:45 for the last few weeks or the last, you for the last few weeks or the last,, few windows has been high. 80% of know, few windows has been high. 80% of people are looking at it, right? You people are looking at it, right?, but if that’s beginning to go know, but if that’s beginning to go down, there’s going to be a reason for that. And I think is it the perfect metric? No. But I think it is so much more impactful and valuable than just act like number of views or just number of users because with an app, with a report, with any metric, it’s not about someone looking at it. is what’s their reason to return to it? Why would they go back to look at

53:16 to it? Why would they go back to look at that report? And that to me where success is coming from. So I would highly recommend here that you look at active users if you were just starting off and you’re not sure like like where where do you even begin? But these are trickier metrics to come out like to build Tommy like how do you that’s that’s hard to figure out. Hey this user ID came up and looked at the report. I have to check that now from like the first time they looked at it to like are they using it regularly each month. Those are more complex metrics. It’s

53:47 Those are more complex metrics. It’s just harder to calculate a windowed period of time and then evaluate against that user and that ID., are we getting a lot of one-time users of reports or are we getting a repeated user, right? Is this report actively being what is that what what defines active to you, Tommy? What does that mean in lie of this report? report? Is that Is that Well, that’s why it’s to the app, especially when we’re looking at the app because reports are tougher, right? Because again, some reports you don’t need to look at every day.

54:17 need to look at every day. Yes, understand. But I’m just saying I don’t even want to talk about that. I’m just talking about what do you define as active? Someone looked at it in the last seven days. Someone looked at it just today. Like what what does an active report look like to you? And even then I think so going back oh yeah maybe a bit more conceptually is we’re really good at creating we’re really good at reading and we’re really good at updating. Those are three areas of development that I think most companies are very well verssed in. We know we need to create more things. We’re getting a lot of

54:47 more things. We’re getting a lot of requests we can move towards that. But if you think about where information needs to live in the organization information is like a moving target. Sometimes there are certain parts of your business where the information moves., and what by that is, let me say this a different way. way. Information moving or why what your insights are. If you have a report, I see this multiple times. You build a report has a specific purpose. Here’s the purpose of that report. People go in

55:18 the purpose of that report. People go in the report, get familiar with the report, understand the insights that are developed from that report. sales up or down.,,, inventory management, product up or down, right? This is good, this is bad. I can Okay, great. But over time, there’ll be additional questions that that report can’t answer. And so the where I want to spend my effort, where I need to spend time. Okay, great. We have this inventory report, but why are these six products always low? Are we not ordering enough?

55:49 always low? Are we not ordering enough? Is this a supply chain issue? like there’s a second layer of questions that we may get into to ask to get into that next idea of information. So, and the reason I’m saying this is because this is it’s a bit fluid. The report we build today may not be the report we need tomorrow and we may need a one-off report for a couple weeks, but then get rid of it once we’re done using it and the problem or issue has been now resolved and we,, we fixed the issue or fixed like, hey, we’ve switched suppliers now. Now those

56:19 we’ve switched suppliers now. Now those six products are no longer an issue anymore. Like so that report highlighting those six products and having them low inventory level may not be relevant anymore. Right. Well, we don’t have a good method to delete it. it. And I’ll go back to when you’re looking at app use usage. That’s a better way to look at it because again to your point, yeah, I may have that report. Obviously, one single report is going to have these peaks and valleys. my quotota report or my sales quotota report I obviously you’re gonna see at the end of the quarter all the numbers go up and

56:50 the quarter all the numbers go up and the next week is going to be terribly low. I would look at if I am doing an orwide app Mike and I’m building this for the company or half the company I should expect that 80% of the company’s looking at this on a two-eek window and I think that’s one just another way to look at it because then you’re accounting for reports that are going away or those nonactive ones. So Mike, let’s wrap it up here and I’m going to ask you a question. I’ll try to answer it myself first, but if

57:21 try to answer it myself first, but if you cannot prove usage, does the effect can you actually say that you’re investing in fabric or is fabric just a cost? And I and I’m going to try to answer this here because Oh, a good question. Yeah. Okay. What do you think? I I’ve been doing the a lot of these roundts around how do you prove how do you make PowerBI an investment out of cost? And I think most organizations when they look at data running through the organization without seeing the value of it just because people are working on and spending time on it is just more of a

57:51 spending time on it is just more of a cost. It’s like I guess this is something we have to do. the app usage actually tying that again to that AB example I said of we are tracking demos and calls calls are going up usage is going up is actually showing this correlation or the best way that we can to actually this we’re investing because we’re making people better what they do just because you have 25 reports that show numbers doesn’t necessarily mean it’s worth the squeeze for a lot of leadership so for me I would hot take

58:24 leadership so for me I would hot take this and say yeah If you are not having any usage or can’t prove usage is a valuable asset, people are eventually going to look at this. Someone’s going to come in and say, “We’re just spending a lot on fabric right now.” They’re not saying it’s going to be an investment. Yeah, I’m I’m going to Yes, Tommy. I think so. I think that’s I agree with you in a lot of ways there. I just want to be critical about it. I think I just want to align company

58:54 think I just want to align company objectives to where we’re spending effort around reporting., again, I heir a lot more on the side of letting the businesses or business units or teams or federated systems within your company design a lot more with this. this. I think I think many organizations should start with apps in fabric. I I think many far often too many people are trying to go after sharing workspaces and giving viewer access to workspaces. And I think it honestly it overco complicates the

59:25 honestly it overco complicates the powerbi. com story. It’s too convoluted. There’s too much going on. There’s too many things for people to click on when you go into powerbi. com. I think apps are a nice way to simplify the information you’re trying to provide and give to users. Apps are also wonderful because I have a single URL that I can put anywhere I want and anyone who clicks on the URL, it asks you to log in and you get into the app. Period. Nice, easy, and simple. So, I I like that experience that we get there from that perspective. And that’s why I’d

59:57 perspective. And that’s why I’d want to like spend more time evaluating what does our business need to run and then aligning reporting and I’ll I’ll do the full crud thing, right? Have a process in place. When when when can reports be sunseted? When can we turn them off? When can we get rid of things? There’s going to be a bunch of junk that’s creating your organization. There just is. It’s just going to happen. just have a plan to figure out what’s being used, what’s not being used, and how to get rid of it, right?

60:27 right? And I think a lot of your most important stuff should just get funneled into an app. And that’s how that’s the main distribution method you should use for that. And again, to your point, Tommy, I think the metrics we talked about here are pretty useful. Count of users used in the report., active users, right, as how many users are actively using the report. I think those are things that you should be regularly looking at., and I would also argue scorecards with the little timed metrics things, you can attach them to semantic models. So, why not attach them to reports and key

60:59 not attach them to reports and key metrics that you care about and monitor them and have that sent to someone regularly in a team’s message, right? Or an email or the scorecards is a great way of saying, look, we care about number of users in this report. Let’s make sure we put that on our dashboard. We need a handful, not a lot, but just a handful of KPIs that going to tell us what’s valuable. You don’t want to get into paralysis by analysis. Too many metrics will will just muddy the water and you won’t actually know what to do because they’re going to conflict.

61:29 because they’re going to conflict. Thousand%. I love it. I love it, Mike. This is good. Awesome. With that being said, I think this was a good discussion around, you this was a good discussion around,, usage of apps and fabric. How do know, usage of apps and fabric. How do we monitor them? What does this look like for your organization? Are you centrally managed? Are you federated? What does that blend look like for your organization? So hopefully this gave you some thoughts to think about with the apps. If you are a new company starting out with PowerBI, I’d highly recommend for you to check out apps in general. Start learning how to use them. Organizational apps both at the fabric and PowerBI level. That being

61:59 fabric and PowerBI level. That being said, Tommy, where else can you find the podcast? podcast? You can find us on Apple, Spotify, wherever get your podcast. Make sure to subscribe and leave a rating. It helps us out a ton. If you have a question, idea, or topic that you want us to talk about a future episode, head over to PowerBI tipsodcast. Leave your name in a great question. Singular de finally join us live Tuesday and Thursday a. m. Central on all Power Tips social media channels. Thank you all so much and we’ll see you next time.

62:35 Dance to the day in the mix and I get your fix. Explicit measures. Drop the beat now. Pumpkins can’t feel the crowd. Explicit measures.

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.