Mailbag: Analytics vs. Data Visualization – Ep. 296

Chris follows up on Ep. 294 with a fair critique: KPI dashboards are built on what the business expects to see, but analysis is about uncovering what you didn’t anticipate. The conversation drills into audience, self-service reality, and practical patterns for enabling exploration—from Desktop and Excel to Fabric notebooks.

News & Announcements

• FabCon Community Conference — Event info for the Microsoft Fabric community conference; the episode mentions discount code Carlo100.
• Power BI Theme Generator — Quickly generate theme JSON files to standardize report styling.
• PowerBI.tips Podcast — Subscribe and browse recent Explicit Measures episodes.

Main Discussion

The key shift is separating operational reporting (repeatable metrics) from exploratory analysis (question-driven discovery)—then designing both the semantic model and the user experience to serve each job well.

• Match the experience to the audience: executives usually need stable scorecards; analysts need flexible ways to test hypotheses and follow the data.
• Treat the semantic model as the product: invest in a clean grain, business-friendly dimensions, and measures that make slicing and comparison frictionless.
• Expect exploration to be iterative (and temporary): many “aha” investigations are one-offs; the durable output is what you operationalize afterward.
• Use Desktop intentionally for discovery: Power BI Desktop is a legitimately strong (and free) sandbox for fast aggregation and pattern-finding before formalizing.
• Give report users real exploration tools (with guardrails): features like Explain the increase/decrease, Key Influencers, clustering, and Analyze can help—but only if users are taught when to use them.
• Don’t confuse “self-service” with “no training”: perspectives, visual personalization, templates, and office hours often matter more than adding more pages.
• Promote discoveries into repeatable metrics: when a pattern keeps showing up, capture it as a governed measure/KPI and a clear view for the next cycle.
• Use Fabric when the question needs more than visuals: notebooks, SQL, and Spark are a better fit for deeper diagnostics and statistical exploration—then feed results back into the model.

Looking Forward

Run a two-week pilot where analysts use one certified semantic model plus a “sandbox” report page (Analyze/Key Influencers/Explain) to capture 3–5 insights—then turn the best ones into governed measures and a repeatable KPI view.

Episode Transcript

0:30 hello and welcome back to the explicit measures podcast with Tommy Seth and Mike good morning everyone good morning gentlemen happy day happy days it’s almost the end of the week the week excited it feels like it’s going fast I can’t believe February’s almost done at this point been going quick what do they say the older the older you get the faster life seems to go yeah like there’s there’s a some science behind that apparently whereas as a as a

1:02 that apparently whereas as a as a kid you every day is long summer is the longest in the world vacations are awesome because they take forever and then all of the sudden it seems like things start to accelerate and there’s an There’s an actual I don’t know if you call it a theory or a proven scientific thing that there there’s a reason for that yeah I can tell you what that is there’s no science it’s you just like the sip that’s the difference that’s the the only yes that’s the only thing that SE Ates you and a child you go to the bank that was the

1:34 child you go to the bank that was the worst thing in the world all of a sudden I like I’m just gonna not do anything I feel like as a kid I feel like as a kid I was in learning mode all the time and so time perceived to be slower because I was constantly experiencing new things you go on a vacation it’s a new thing you go on a flight that’s a new thing everything’s new to you you get older stuff becomes less new and it starts speeding up and like I’ve already done that I’ve already done a flight before can I get there already like there’s maybe maybe there’s more expectation on the other side of things that maybe it’s

2:05 the other side of things that maybe it’s because things aren’t new anymore it’s just not it’s not fresh anymore I don’t know maybe I don’t know anyway our main topic for today is we’re going to jump into an episode that we’ve done in the past we’re revisit our prior mailbag Chris has responded back to our mailbag from an earlier episode that we responded to his question around and he gives us some flak for not going deep enough so we’re going to go in we’re going to go a little bit deeper on our question that we addressed on episode 294

2:36 that we addressed on episode 294 and go a little bit more in depth I think around insights and data analys and Analysis and what that means and what that looks like maybe focus in a little bit more of our comments around either how we get this done or what does this look like for organizations but before we get into our main topic let’s do a a little bit of news Tommy any things that you’re working on these days what what openers should we talk about here I got two and you could do Choose Your Own Adventure but let’s do both all right so the first

3:07 let’s do both all right so the first one is I think the biggest you if you’ve been using chat TBT like we’ve told you to and use open AI yeah yeah I saw this yeah and it definitely made it Dent to my my day and we’re all loving it so cool look at all the proms thing well it went off the AI went off the rails chat GPT is like freaking out right now and because apparently there’s a bug in one of their packages and oh that could be a problem like do you understand if I had a bug in my software

3:37 understand if I had a bug in my software before like it’s down my software’s down okay now the an is coming out when we have this Reliance he’s using Spanglish it’s using Spanish and English together as it responds to your questions I’ve seen some hilarious as that where just like repeating new lines and just like it was like oh that’s not good or to me it was just like it would say two things and then move on and go SL like what in the world is going on and then it starts I I was reading a couple things on so

4:07 was reading a couple things on so apparently Megan Longoria picked up on it and sent a little Twitter thread that’s that’s how I heard about it she’s like I think chat GPT has been drinking too much or something something along those lines and it just started randomly saying the same word over and over and over again and it was starting to throw in Spanglish and it would have like a paragraph of what you wanted and then it would just like have gibberish and you’re like yes these are words yes you’re writing them as if they are a sentence but none of them make sense together there’s not a coherent thought in the in the context of the

4:37 thought in the in the context of the words so words so it it it was very funny some of the the comments and the things that people were asking it and it gave you some very weird answers now what’s interesting to me is we we tend to even with animals right we we humanize them we put human true if if you’re a supercomputer and all people people are doing is asking you to do mundane tasks which is what we are like we want to save all this time like we’re going to ask you the same dumb question

5:08 going to ask you the same dumb question over and over and over maybe it’s just losing its mind right faster like really like can you imagine having the same conversation over and over and over and over over millions of times right and and the prompts that were so proud of are just different ways of asking for the same same thing let me show you where your system information is and how do you get your Wi-Fi back working exactly and I especially with this chat GPT so so I’m I’m going to go

5:39 this chat GPT so so I’m I’m going to go into chat GT chat GPT a little bit more here and potentially even co- I want I think this is a very interesting conversation Tommy I know you have purchased co-pilot Seth do you have co-pilot yet like the the Microsoft 365 version I do have you gone into Excel and used co-pilot inside Excel I’ve not tried that yet okay the other day I was doing some Explorations around co-pilot inside power query so I’m I’m using co-pilot and power query I’m trying a couple things I’m getting a little frustrated

6:09 things I’m getting a little frustrated because I don’t really know what it can do and then a recommendation came to me it says go see what the Excel team has done with co-pilot and I go into co-pilot it I get into Excel and I click the co-pilot button it says hey we only operate on tables because we need to know a defined range of data I’m like okay that’s fair it says would you like me to add a sample of data okay boom table appears like six Columns of stuff I’m like okay this is interesting it made a whole table for me and then it prompts me with questions it says would

6:39 prompts me with questions it says would you like me to add a formula to enhance the data or the analytics in your table yeah sure do it here’s something we found interesting and it puts another metric on the table okay all right now I’m getting there’s this point in time when you start using co-pilot or chat GPT of things you’re like okay I see what you did there I’m going to challenge you now okay let’s go like you did something that I thought was maybe a bit impressive okay let’s see how far this game will go it’s like you’re going to push poke the bear at this someone made a shot like do it again do it again yeah do it again yeah yeah do it again

7:11 yeah do it again yeah yeah do it again let me see come at me bro come at me so now I’m gonna bring in my own table instead of your your table you’ve already you’ve already like I kept going further I kept saying okay can you add data bars to the column called whatever the name of the column was sales number boom let me let me let me do that for you here I formatted the column for you with data bars I was like okay all right then I said then I started saying okay what what if I could stack commands

7:41 what what if I could stack commands together like what if I could do multiple things and then I said can you sort this column in descending order and highlight any value that’s above 50, 000 boom it did it two commands at once and I was like oh snap we’re playing a game here like you’re showing up the to play so anyways I was very impressed with what I could find and what I could do there and I really enjoyed the concept of simple commands that I could just ask it to

8:13 that I could just ask it to it’s it’s the the mundane it’s to your point Seth you just said earlier it’s the mundane things that I don’t want to figure out where those but I know it can do it do I remember where all the buttons are and all the buried menus that are in in the ribbon maybe maybe not if I’m spending a lot of time in there yes fine maybe I’m using a search thing but chat GPT should just be like here it is Bing and maybe the next time I actually go find the button but I honestly I think people are going to rely on chat chat GPT like if I can just say make a bar chart on this column and

8:44 say make a bar chart on this column and also I’m in Excel this is in the web so Excel for the web I’m able to like have my microphone I could just talk to it I’m like make a bar on the column named blah blah blah and it just it would do it I was like oh this is amazing so I can start talking to my computer and it actually starts figuring out what I actually want and does things H this this is interesting this this is valuable because as you’re as you’re talk so was this just in Excel co-pilot or you were in in power you in Excel co-pilot no so this was this was purely

9:15 co-pilot no so this was this was purely a co-pilot in Excel experience I wasn’t in power query there because like your your description of things like I just wrote down two ideas I gota go try with like that I know have been challenging for for like some analysis that people have been trying to do right now yep and if things are that easy where I can just say yeah do this thing and that thing and like let me know what the output is and to be to be H to be perfectly honestly clear here like we’re at the

9:46 honestly clear here like we’re at the beginning stages of this this is how long has this been out for I just bought it like a month ago I haven’t even had time to play with it I don’t even know what prompts I can use right and so one of my main challenges here is here is my understanding is as this as co-pilot is being used and interacted with these different products it’s learning what is working well for these situations so for example it doesn’t hallucinate but It probably does but I didn’t get

10:17 mean it probably does but I didn’t get it I didn’t get it to do it but I was able to like look at like I was able to write statements against it and it can do things that the programmers of co-pilot didn’t intend it to do I was talking with the Microsoft person and they’re like yeah we’ve seen some surprising results out of these things I was like really so it it has learned like sometimes it learns how to yeah sometimes it learns how to like do things that they didn’t explicitly program it to do and it’s figuring out

10:48 program it to do and it’s figuring out okay this is what users want when they ask this I’m going to do this to the the whole action-based thing the whole simplified like the experience of being able to let people just communicate to

10:59 able to let people just communicate to the program and say I want you to do this and it it is able to interpret that find the menu option click the column and do the things those simple based actions are huge I think that would be really relevant and any yeah you had that with copilot Studio yes well so there’s there’s now so this is they need a library and so in Excel there’s a little like notebook at the bottom of the icon I didn’t have time to play with it but apparently there’s the ability for you to like go see all the

11:29 ability for you to like go see all the different prompts that are supposed to be used and things you can go test and try but I feel like I’m back in the early days of Google where I have to figure out how to Google everything again I don’t I don’t know how to write a a question to Google to get the answers that I want I’m still trying to figure out what prompts can I use here to get this to work and it just will take some time it’ll take some exposure all this to say is I was very impressed with Excel little underwhelmed about power query it feels like there’s needs to be some more work done there for what I’ve seen in power query it doesn’t quite do as many

11:59 doesn’t quite do as many so I was amazed at the Excel integration I was not Amazed by the power cor integration yet I’m still actually Dove tail into the part of the conversation today all right well then that was the that was one of our openers yeah what was the other thing you got Tommy let’s go to the your other opener do you have anything else I could like literally do this for the next hour speaking of also relying on all the different AI things I’ve been working like since you’re a personal one at a company you have to worry about

12:30 company you have to worry about your own upgrades to your computers and oh totally yeah so IC now you’re gonna be a true designer but like I have this custom PC with a bunch of slots I’m like I wonder what I could do with this so I basically and now Google has their new AI which is really cool that just came out and I was basically like feeding into some things like here’s what I’m trying to do what set up do I need the infrastructure I’m

13:01 need the infrastructure I’m running AI servers here but I’m also doing the database stuff but I’m also doing a lot of training so and it’s like hey what’s your motherboard it’s like okay well if you have those three hard drives purchasing these together here’s like certain like SD drives that will make sense da da da D oh here’s all the ports that you need to do and they gave me a diagram of my motherboard after like so I was literally going like slot seven right and I was like okay this is cool and then I asked him to compare the

13:31 cool and then I asked him to compare the top seven like hard drives for like what I needed because I’m like I am I gonna do partitions or am I G to do a skip drive whatever the heck that is and I basically said like basically telling me like you figure it out what I need but you asked me those questions that you need to ask and it was this ability now just to rely on it so like my computer’s about to be so close except I’m missing a screw because I dropped it into the computer but apparently

14:01 apparently very we but like all the stuff I was like yeah I could have Googled all this which I tried Googling too because again when it said here’s all your comparisons that you had based on a ranking none of these numbers are correct I they’re all made up and I was like okay thanks anyways so it’s getting pretty impressive they can go find and again it’s when it can Source the internet for information and scan different web pages and then say okay here’s 30 web pages that tell you how to build PCS build PCS here you go boom I’m going to look at all of them and give you a summary of

14:31 all of them and give you a summary of what’s going on there that’s the value like it feels like these AIS are able to like when I did Google Google would say you search a question I will return the one result that has the answer that you need what I feel like is happening now is you are asking co-pilot questions or you’re asking the AI questions and it’s summarizing multiple pages of stuff and returning back okay looking at these 50 pages here’s what we think is the best options for what we’re seeing consistently across these

15:01 seeing consistently across these different areas which is very interesting to me we’re just touching the surface because everything we’re doing right now is that simple prompt based chat base no memory thing no guided part where like kind no guided part where like like the agent side hey I it set up a of like the agent side hey I it set up a jupitor hub for me on another computer and I basically let it run my computer and did all the configuration so like the longer memory side of things rather than just that simple question answer which that is right now co-pilot for Microsoft we’re that’s where we’re

15:31 Microsoft we’re that’s where we’re touching the surface Mike and it’s very scary to what else we can do so help me understand what you’re doing with this you are using your computer to host your own large language model and you’re asking it questions directly yeah and then there’s a a package called open interpreter and I had a Raspberry Pi I’m like set up Jupiter Hub on the Raspberry Pi all the configuration whenever I log in and I connect to that IP you should launch this a spawner thing so I don’t have to put anything on my my computer Tommy I knew you were a

16:01 computer Tommy I knew you were a nerd but this goes did it perfectly we’re we’re at another you to me like you just leveled up in my mind like I would not have anticipated like Tommy going out and saying okay I’m going to get a large language model I’m going to go figure how to render it on something and then I’m going to use it and and get to it via I knew you I knew you’re technical but Tommy this is another level your IP add You’re Building IP addresses and then submitting questions to a random IP address in your local network this is crazy I give it admin I said run Jupiter help here’s I want I got cof and I did

16:32 help here’s I want I got cof and I did on my computer on the Raspberry Pi and it was done it was done it was done it’s insane yeah so Tommy Tommy at this point I would just like just just a reminder you have a family upstairs sometimes you need to pull yourself out of the dungeon go say hi I I heard the same thing this morning Tommy has just recently installed Tommy has recently installed a shower and a bathroom and the kitchen is currently being installed downstairs in the basement there will be no leaving for

17:02 basement there will be no leaving for Tommy the wife will Contin the door on the lock the lock on the door upstairs has been switched it’s now the other direction so now Tommy is locked in the basement and is required to continue to make money while he please create me a shower buy you a 3D printer while it’s while I’m working pal please please build me a bathroom yeah why hasn’t why hasn’t anyone made why hasn’t anyone made an AI that’s called how that that that’s a Miss I think we know yeah we

17:35 that’s a Miss I think we know yeah we probably have probably already there you just don’t know it is how yeah imagine you’re yeah I don’t I just I just because I just said it now it’s in it’s in Internet to life yeah right interesting all right so one more announcement here real quick I’ll just throw a last one and we’ll get into our main topic today so this March March 25th through the 28th the Microsoft fabric conference is coming out you can go visit Azure datac comp. com and if you’d like a $100 off your ticket you can get $100 off by using the coupon

18:05 can get $100 off by using the coupon Arlo 100 to go get $100 off your ticket if you’re going to the Las Vegas conference for Microsoft fabric should be a great conference looking forward to it we’ll see how we it it doesn’t feel like this is a dynamic conference taking over fabric it feels like this is an exclusive fabric-based conference so very looking very much looking forward to that so hope to see you there if you want to join us out in Las Vegas from March 25th 26th through 28th there’s also some preons and some postc cons as well if you want to learn more things about Fabric and how to use it all right with that any other RBI

18:38 it all right with that any other RBI blog is out oh yeah okay let’s maybe we talking about the desktop we’ve been talking about desktop updates and now it actually came out so give us give us a couple things Seth maybe give us a couple high level idls here of what you’re find in the blog I don’t want to take too many too many things away like We we talk about a little bit visual calculations the biggest thing for sure that for sure dive into that all that all right I’m I’m not saying that the timle

19:10 right I’m I’m not saying that the timle thing isn’t important okay I’m just saying from a wider audience perspective that’s probably true visual calculations and timle is much more in the development side so visual calculations I was very pleased to see there’s a lot more detail around on object interaction updates M I’ll admit it’s still a bear to work with but I definitely use the on object updates as well well now they have changed it so that you can actually

19:41 have changed it so that you can actually select two visuals and resize them so properties that are common across visuals such as width and height you can now select two different kinds of visuals and have the the size and style still show up amazing that’s just like why why wasn’t this done like turn the like why not grab all visuals and say turn title off like that should be a thing like every visual has a title sure let me turn it off another one here also is when you don’t add data to a visual you could never adjust the

20:11 visual you could never adjust the position of it so there are properties of the visual that appear only well it traditionally was the visual would be put on the page this is one I’ve been whining about since day one like I’ve been complaining about this you put the visual on the page I need to be able to adjust the position the length the width the height the direction like all that that should be adjustable for me even when there’s no data attached to it don’t restrict me so that they fixed that one and there’s a whole bunch of bug fixes and the main one that I’ve really been having an issue with is

20:41 really been having an issue with is when you work with certain visuals the build pane would accidentally close or automatically close so you would click on different objects on the report page and some of those windows on the right hand side would then accidentally close or shut down or minimize themselves when you were clicking on across different visuals you maybe you were doing like a cross filtering or cross highlighting and all of a sudden boom all the panels on the right hand side start closing and it was very jarring to have the UI changing like that so a lot of bug fixes is there so I felt that that was a good another good solid

21:11 solid update anything else that stood out to You Tommy I was gonna say the PowerPoint thing but I figured I knew you were gonna say PowerPoint I saw that I thought that’s Tommy’s feature right there I know no I honestly I I think yeah the visual calcul calcul ations are probably going to be the most significant obviously timle and they talked about the like how they’re actually migrating that anytime let’s go back to PowerPoint like you said PowerPoint but you didn’t describe it like no because it’s I was just doing

21:41 like no because it’s I was just doing that to you because oh you you you don’t really like the feature you’re just trying to get Meed up I’ll just get you riled up because oh okay on it no the PowerPoint is rather than adding the link you you can actually get a shareable link from PowerPoint they’re still working hard on the integration so no

21:58 no there’s but this is a big they came out with a bang for February well they had like a month like they didn’t release anything for January so it’s like it’s a lot of bug fixes and things in there and so one thing I will note about timle and I think there’s a Twitter thread that’s also talking about this as well so my impression was when you so timle is this new format if you’re in desktop and you save your project down as timle and you publish your project your timle project you publish your timle project through Des toop back to the service it

22:29 through Des toop back to the service it translates it back into a tmsl the model. Bim file so the desktop recompiles all the timle files and pushes it out to the service and the service will retain that tmsl the dayon object for model depm however ruy I think made some comments that says if you grab and create your pbip file and you publish that into the repo before you you synchronize that back to the

22:59 you you synchronize that back to the workspace then the timle format will be obeyed and and retained when you work with the file so it’s it sounds like it’s if you publish through desktop with the pbip format it converts it back for now but if you publish your report or if you save it and then save it into a repo and then link your repo directly to the workspace then you have the timle format that you can use so I don’t know really know the implication that means yet or how that will work but it sounds

23:30 or how that will work but it sounds like it it feels to me like you’re not going to be able to publish over an old data set you’d have to delete it save it as timle push it to a git repo and then synchronize that git repo back with the workspace in order to get the timle format to format to work but I understand why they’re doing that because ultimately if you’re using timle format anyways desktop should be accessing your powerb reports from the git repo and the service should be using the git repo so ultimately you shouldn’t be publishing from desktop

24:00 you shouldn’t be publishing from desktop anymore basically if you’re using Git You should just go directly to the repo and use that to edit or pull down your files and change things and check things in when you’re done so it it it does feel more like a proper development process process anyways interesting notes around that as I was discovering more about the new timle format and how it works sweet all right with that let’s stop talking about fun new things let’s let’s go back to our our actual topic of the day Seth you want to there’s a lot let’s revisit all of our deficiencies yeah let’s this is going

24:32 deficiencies yeah let’s this is going to be he’s poking a little bit at us here I think this is good I think it’s good to push back the community is pushing back on what we said earlier so let’s good let’s go ate it how do you want to do this Seth do you want to read out the the the statement again okay oh yeah all right otherwise otherwise it’s not a true mailbag all right let’s read it it out hi fellas it’s Chris and we’ll leave the last name off sure again this question is a direct followup to episode 294 294 being data analytics versus data Vis Vis visualization I felt you guys were

25:02 visualization I felt you guys were dancing around something really important but not really getting there snap related to finding the AHA Insight in data I think it’s really hard to design a powerbi report or app that does that which is why so many report consumers want to dump the data to Excel and then start working with it when you gather requirements and build a report it’s all premised on what the recipients expect to find as insights that’s not data analysis so someone in it or Central bi as the data modeler report

25:34 Central bi as the data modeler report Builder are not being are not doing data analysis You’re Building reporting around kpis okrs that were designed based on assumptions that the business has about their area of the business expected results don’t make you go aha it’s finding unexpected things that is really the goal of data analysis what’s some correlation that we didn’t think about before what’s some outcome that we didn’t see coming when we’re in powerbi desktop we can play with the data sift through it and find anomalies where we

26:05 through it and find anomalies where we might say that’s funny what would that look like if I plotted it in a scatter chart or what would I see if I showed it by this Dimension by this other dimension so that I can see what’s really going on here that’s data analysis so my question is how do we enable that experience for report users who do not who not going to design their own ad hoc report in desktop I suspect this might involve perspectives and visual personalization

26:36 perspectives and visual personalization Plus for Boss training the goal would be to allow consumer analyst to find the insight and then be able to visualize it in a way that they can then communicate out in an understandable way so this isn’t something to design for executives of course but for analysts in different areas of the business I think this is the Holy Grail that’s so hard to achieve thought thought question mark say yes okay let’s move thank you for not having a character limit I I I so so one I love the fact

27:07 limit I I I so so one I love the fact that he ends everything with thoughts question mark because I do that all the time old actually this is a question directly from Seth Seth just yeah I decided to label it as just lab new name some random names you went to a random name generator and just pulled this in no no this is actually a real person this is not Seth so by the way there’s a lot to unpack here I think there’s multiple levels that we can we can chat through and talk about give our perspectives on things many different levels in this one there’s lots of unpacking going on here all right let’s

27:37 unpacking going on here all right let’s break it down to main category Seth what do you think are like the main things you want to hit up because I think there’s a ton of direction we can take take this well I I think ultimately we’re going to talk about the you know right right back at me well here let me let me give you a moment to think Seth I got one here that I think I think part of the topic here is the audience matters right so I do

28:08 is the audience matters right so I do think there’s different kinds of reporting for different levels of the organization and I think that was very clearly pointed out here like insights insights or analytics how you how you want to describe that if you’re an executive member the insights you’re looking for are consistent numbers and I think Executive leadership is also looking for consistent insights and there is a let’s call the Insight something around a

28:38 call the Insight something around a budget and then measuring myself to the budget right I have an expectation of where I’m going to get to and then I’m going to have how close did we get to that expectation so if you’re at the higher end of the organization you’re trying to move resources around either into sales into product development into supply chain like you’re making high higher level decisions where you’re moving groups of people around to make sure that your end objective which is likely making money or being profitable right those end objectives are being delivered so at that higher end level I

29:11 delivered so at that higher end level I think your insights are very consistent and they’re repeatable day-to-day when you talk about the role of the analyst the analyst and I agree a lot more with this with with Chris’s comment here the analyst is looking for those aha moments right so we know sales are down that’s that high level okr that main goal okay sales are down analyst go figure out what do you need to go find in the data that is describing what happened oh the analyst found out we

29:41 happened oh the analyst found out we lost one of our largest accounts or second largest account they’re no longer buying from us that therefore we have to go make up sales somewhere else to to increase our revenue for the year or for the quarter whatever that may be right so I think to me there’s like different levels here that analyst level is doing more of a self-discovery I feel like there’s when I was going through like creativity some workshops to help be creative and things there’s this concept of like you start with a very narrow scope and you widen out to a very large area so you

30:13 out to a very large area so you start with a okay the initial assumption is sales are down okay then we do exploratory things in the data we we widen our scope and say let’s explore our data and figure out where that’s coming from and then you narrow your focus back down again so it’s just it’s just like narrow funnel goes to a wide funnel and then come it Narrows back down to an answer or something like that so that’s one pattern but you could also think of your Discovery things as like a funnel that has a wide end so if

30:44 like a funnel that has a wide end so if you think about a funnel at the top that has a very large hole there could be a million different questions you want to ask of your data and you’re going to try and use the report to narrow down to a specific area you’re going to you’re going to tell a story then there’s the opposite effect you’re going to come in with a very specific answer and you just want to explore the data you want to widen your scope of what you’re looking at so I think I think there’s many different ways to portray this and I think audience matters here and that’s my initial thought yeah I just recapping

31:15 thought yeah I just recapping where where we were in the previous conversation I’m trying to like pull it up in my head right we were we were describing the difference in in the two what is the difference between an data analytics ICS and Analysis versus data visualization and the analytics part finding the meaning a lot of this is ad hoc it’s it’s very exploratory in nature and then it gets to a point where you’re ready to refine the message a bit so I I don’t I don’t want to like pick apart

31:45 don’t I don’t want to like pick apart the parts but like in the statement that stood out to me is when you got like Chris saying when you gather requirements and build a report it’s all premised on what recipients expect to find as insights and I don’t I don’t think that’s the case all the time right like I think I think there are many different types of reporting and when we start to talk about visualization and powerbi and building requirements around things the requirements for how we display information is important and in

32:15 display information is important and in many cases those reports after having that solidity of or that requirement sets around what we’re going to build the value is in the data changing underneath all the time that direct or influence the different insights that are provided to the end User it’s it’s not about and and some of that is exploration and some of it is understanding like what is happening in in just a couple areas like planning

32:47 in just a couple areas like planning forecasting Finance even right like these are the same visual it could be the same visuals it could be the same thing things in the report that are indeed driving the ins sites without you

32:58 indeed driving the ins sites without you having to do raw analysis and I think that’s where the distinction is in like I I get it I do raw data analysis all the time and there’s aha moments and very rarely am I in powerbi right I’ll start with queries I’ll start in the source systems it’ll potentially go to Excel and I absolutely get the okay I I want to see this on an aggregated level I’m going to play I’m going to use powerbi desktop as a play around tool

33:30 powerbi desktop as a play around tool totally I do love the point though there that that is a huge use case of powerbi desktop and the fact that that’s free is crazy amazing that’s very true because just from an analysis perspective right power that tool alone to be able to be like what I have I have these Excel files it’s I’d have to go either pay for pay for co-pilot or Y or generate all these

34:01 co-pilot or Y or generate all these things on my own or I can just bring them into powerbi I can drop these fields in and then they automatically get aggregated for me and I can I can group I can slice and dice I can do all this Dynamic stuff in the tool to give me the Insight I’m looking for and that that is really strong well let’s go I want to make a point here you made Seth I want to just emphasize one thing Tommy before we jump in there Seth you said I could go to Excel or I could use a free tool go you have to pay for Excel Excel just doesn’t show up for free and so th

34:33 just doesn’t show up for free and so th this is like one of the things I think that’s very Genius of Microsoft they’re giving away analysis services on your computer for free they gave it away they they gave it to you in desktop you get this really rich power query engine you can have the vertec engine you can build what you want in the report now the genius I think here is or the Strategic part is well okay great you have the data it’s not valuable unless you can share it with somebody else so like there they’re making the money on the service side of thing so we give away the software but when you want to collaborate on it that’s the that’s the

35:04 collaborate on it that’s the that’s the the hook that gets you to buy the rest of the software pieces so I just want to be very clear here you can do so much with just powerbi desktop and that could be a heavy data analytics tool just in and of itself if you’re a team a small business unit if you can get your hands on that that’s incredibly powerful anyways I just I just was very shook by your comment that it resonated very well with me as far as the cost aspect ratio here guys I’m this is all very

35:34 here guys I’m this is all very much voting well to my argument on that conversation where not every visual is analytics and that’s what I was really my point of view was just because you have a report doesn’t mean you’re doing analytics because we don’t to get in that but I think that’s a this is a big Point here I think what Chris is at least alluding to where some someone asked me to build a report they are not asking me to spend time telling them what to do with the data or can you dive into why the

36:06 data or can you dive into why the numbers the way they are they are asking me to build a report with not necessarily the expected number but the expected metrics if I add it in there oh by the way here’s your analyze by in the breakdown of the distribution they’re go how many times you that’s fine but that’s not necessary and there’s that big distinction be between a report that is like on the observational side but that’s what people are really paying

36:37 but that’s what people are really paying us to do and unless it’s explicitly part of the of the conversation with with business was like hey go grab the data we want to see X Y and Z also can you help us understand why our sales are down over the last six months can you put something together in there I love doing stuff like that but that’s what’s changed to me with my rool as a data and analyst where when we just had Excel

37:08 analyst where when we just had Excel because I didn’t have a lot of other you because I didn’t have a lot of other there was no modeling so it was know there was no modeling so it was like where you’re going to spend your time it’s like hey we just did this campaign can you pull together what worked what didn’t now we’re just saying we’re going to show you the trending no one’s asking me for my opinion on the data right I’m not going to go back up and just so you guys know you’re spending a lot in this they’re GNA go great we we deal with us we you deal with you and I think that’s that’s a big part of this and go ahead

37:38 that’s a big part of this and go ahead try it and break down your next report to them and more often than not that’s just not the expectation of the visual side of powerbi and yeah it does have all those features but we’re also never tasked with that either or more often than not when we’re doing modeling I I don’t I don’t I understand what you’re coming from and I think I agree with on the most part but to

38:08 to say to say standard reports with common known okrs or kpis on them to say that that’s not analytics I would disagree I think it’s a part of the analytic story right sure you’re just describing more of a self-service method right so where I think I I delineate my lines here a little bit more in this is there’s going to always have to be standard kpis and known things that’s been developed and it’s agreed upon by the organization that this is the way we do business

38:38 that this is the way we do business those don’t change what I’m I guess what what I’m hearing in this in this comment here where I think this where Chris is going here is I do think so I think there is a lot of Discovery self-exploration figuring out what things are going to be working inside your data set that’s totally okay and that’s why powerbi exists but this is why you develop Cube that you can then or semantic models that you can go share right so that’s why we that’s why we design things the way we do so the idea is the the design of the data should support both types of metrics it should support standard kpis

39:10 metrics it should support standard kpis and standard analysis against designed goals and it also should allow exploratory things and where I want to push on some topic here is while Chris says getting the analyst to dig through the data and go aha I found something those aha type reports typically don’t last very long I feel like they they are you dig through the data I’m looking for a question I

39:41 the data I’m looking for a question I get down to oh I think this is the answer to that question you think about art articulating that story and then the data changes and then that story is no longer true so the same group of visuals to do that analysis around here’s a thing maybe don’t pertain long term now what may come out of that analysis is you may decide hey we need a couple more actual kpis or things that help us decide what is good data or not good data or what is good health of our business or not good health of our business but I don’t know if I

40:13 business but I don’t know if I agree with the fact that these self-discovery things will live on for a long period of time I think these are I things that you do for a period of time they work and then they eventually need to be deprecated well it’s not it’s I I think you’re on a great point and I want to I want to agree with it and push it in a direction and then I want to Circle back to I think the core of of what he’s asking one one I totally agree with you I think I think there’s this premise or

40:43 I think I think there’s this premise or thing we’re picking up here where analyst analysis always like always has to be in that exploratory exploratory stage and all of our our our things need to be there and they don’t to your point what is what is the the outcome of a lot of that exploratory stuff in your case you just said one-off reports yep right happens y okay quot in quotes the query right you have these you have like somebody asks for

41:14 these you have like somebody asks for something and then they’ll come back at random times every four or six months and say hey remember that report you ran for me yeah can you do that again right yeah but then like a lot of that exploration requires aggregation right and then we need to aggregate it output it create an output where people need to see this aggregation and that’s where the value is because we’ve already done the exploration we want to now enhance and refine the outcome of that

41:44 enhance and refine the outcome of that exploration right yes and then we take it into advancing it and and that’s where I think powerbi shines from the standpoint that regardless of if you’re wanting to look at the past like like and this is where you pull out all the buzzwords of reporting descriptive diagnostic predictive prescriptive right go down more you go down this thing all you were’re saying is there’s a more refined approach to how the organization wants to like utilize the data that you’ve found these things in right

42:15 you’ve found these things in right because there’s a lot of value there and there can be a lot of value in these discussions to not just look at what happened but figure out what is making it happen and that’s what that’s what the fun I think of is a lot of in data analysis like what’s causing all this but then ultimately a business wants to get into can we predict what’s going to happen right yeah maybe and or eventually can we can we have the outputs of these data machines tell us what to what to do so I I think that’s where I like we

42:47 do so I I think that’s where I like we land in this landscape which we should which is not everything should be analysis like straight up exploration however if we go back to the key part of his question and I want to repeat it my question is how do we enable that kind and my that that kind this exploration I think aha first time moment experience for report users who not who are not going to design their own ad hoc report in desktop I suspect this might involve perspectives and visual personalization

43:17 perspectives and visual personalization so like the the idea is can can we have that mod like the model itself is creating value for the business his expectation seems to be that like these an analysts aren’t going to build their own desktop reports but can can we use powerbi as a tool to be that analysis tool to be ad hoc and dynamic in a way that isn’t the desktop is what I’m I’m reading here yes and and maybe not right like it

43:52 here yes and and maybe not right like it if if you’re in my mind there’s there’s always there’s been this theme over my

43:57 always there’s been this theme over my career right where people who don’t understand I shouldn’t say that there’s this perspective that if you if you build a a UI or a tool for business users to go build their own reports that they’re going to do that and know how to do that and they don’t and they won’t because it it requires them to learn something in a hardcore way a very few will always learn the tool and always find the value and

44:27 tool and always find the value and you’re not going to be able to spoon feed that I I guess in report Pages enough to cross a spectrum in my mind to do true data analysis of like exploration Insight on data sets because you’re going to already be utilizing tools like Excel or powerbi desktop as you’re doing that exploration because you want to shift gears radically Based on data sets that’s the the basis of doing just raw analysis yeah and I I don’t think you can produce

44:58 yeah and I I don’t think you can produce that output in powerbi do you guys or do like that’s the way I read his description like can we do that for the end user to go find their own insights but still give them the con like the structure of like the tool sets to operate within there right to me this sounds like throw everything in the model and and then have like all infinite choices of of visualizations to slap data together no

45:30 visualizations to slap data together no I think Mike you said this in the beginning I think there’s a bit of training we do on the exploration side is more than just like obviously the bookmarks and drill throughs like there are a ton of features I think that we just don’t utilize I love the analyze by and just being able to explain the increase explain the decrease is another good one too yeah and the key influence yeah so like just having that in being able to have the author do that

46:00 in being able to have the author do that or the per like the self-service side not just the consumer because then they can add that visual I think the obviously are the clustering showing people the automated clusters the key influencers just those alone if once you showcase how they work and I I think that’s a huge beginning starting point the problem is unlike with something where I think your nor like your citizen develop data analyst 10

46:31 your citizen develop data analyst 10 years ago right you had those Excel files you’re like oh I had I got that one result I’m going to reference that one cell and I go step by step the powerbi the visual side is not it’s not really meant to do that kind like it’s not doesn’t behave like a Jupiter notebook like oh let me get that direct output of this one analysis and now I want to dive in a little more I oh man Tommy you touched on a lot of points here that I was going to reach out to talk about as well

47:01 reach out to talk about as well okay ahead I I loved your comment there Tommy and I think I think we’re in a stage right now where powerbi is directly we have let me say this way prior to fabric you’re right Seth I think a pure powerbi desktop could only do so much and we had a little bit of extra tooling to help out and and what I extra tooling to help out and and what by that is we had pagein reports we mean by that is we had pagein reports we had Excel files going through

47:32 had Excel files going through this where I think some of the real analysis starts coming through here is building a bunch of pivot tables on top of a cube very helpful getting into other statistics of things right looking at different numbers now so if we if we extrapolate now and say okay powerbi is becoming the great analyst tool for for everything that we need to your point Seth like yeah we still have the cube the cube still makes sense but now I’ve got like a lake housee now I’ve

48:04 now I’ve got like a lake housee now I’ve got spark notebooks and with spark notebooks I have everything that comes out of out of python statistics that go along with this I can now use SQL wherever I want to go grab data and make new groups of tables so I think what we’re getting now in the service with fabric is we’re now expanding the ability we’re expanding the ability for people to not just focus only on the power port because to your point earlier Tommy I think you were saying this as well the desktop is limited it’s you can do bookmarks you can make a story with it but that’s kind

48:35 can make a story with it but that’s kind can make a story with it but that’s maybe after you’ve already thought of maybe after you’ve already thought about and chewed through the data right I really really like this experience of having like a cube connected to a a notebook and just digging through data like I’m going to run a bunch of statistics on things which which I have a whole big table which let’s do a correlation Matrix let’s figure out which s sensitivities of which variables align to other things so we’re getting to a point now where we actually have the Tooling in front of us that is going to allow us to do more of these true data analyst type things very

49:07 these true data analyst type things very technical pieces and again I think this is why I was harping on earlier the audience to me matters because the audience audience defines what and how much of this information we’re going to go after does that does that make sense I feel like we’re yeah absolutely does right now in this conversation and particularly with fabric Microsoft has made an inflection point for me this is an inflection point where I’m looking at back and going wow things were great before but oh my goodness if I look forward now into the

49:37 goodness if I look forward now into the future the analyst or doing true data digging to find problems and and answer questions can now be really developed at a much higher pace and with a much more rigor than we had previously whether that’s whether that’s now applying some data science to what you’re doing whether that’s now applying just general statistical packages whether that’s now applying I might just go through Dan one of my employees that are our company just found this really great Learn Python

50:07 found this really great Learn Python it’s a book but the whole Learn Python thing is in Jupiter notebooks and it has here’s something we’re going to learn in Python and then here’s the code run it right here test this out here’s an example build your own thing right beneath it so this idea or this concept of having a single Report with just visual bolted into it why not do the exploratory data and tell the story in a notebook I could have I can have words describing what’s going on here’s where I’m thinking about my analysis here’s a graph or an image that

50:38 analysis here’s a graph or an image that powerbi are were producing inside the notebook here’s a table of data that supports it here’s some other like boom boom boom bo and now you can have a congruent analysis thought all inside the ecosystem does does the does and I I don’t know this does the fabric ecosystem in notebooks have have a visualization feature baked in there yes it does I know data bricks does but like I guess I guess what I’m hearing you say is for a lot of the data analytics or data analysis a lot of this lives in a

51:08 data analysis a lot of this lives in a different tool as far as like a better solution I think my maybe my point is there’s now a variety of tools that can now do this right so like so it was just desktop and reports now we have other things that potentially also could help us out as well well and I think I think what’s really cool about what you’re talking about is this the rallying around a Delta table right yeah or ultimately a model oh because we huge huge moment the the engineering to get a data set to the point where you have a

51:39 data set to the point where you have a model that previously yeah you create you try to Cobble together a report page that is a mix of table and aggregation but you might miss it and how do you how do you allow the flexibility of that within that framework and it is harder it’s it’s more difficult you’re not going to serve all the audiences you’re not going to get it all and you’re not going to have as wide of scope as you would have in tool sets that you’re talking about in a notebook I can get into the data into the model yeah I can

52:09 into the data into the model yeah I can look at the data and then event like as I’m thinking about how I I discover things or analyze it always starts with the data I see a pattern something doesn’t jive with business logic or my understanding of the business and then I say say oh th this is this thing what is the scope of that like it how rampant is this issue in my data set boom I go into aggregation right and a lot of that time is just summarizing writing a query in a

52:39 is just summarizing writing a query in a different way but that’s where these tools now having the built-in visualization aspects to aggregate the data or show it in that way in a visual way is very easy for me to digust and see what type of problem I have in a visual sense not just the table sense so like like to me that is a fantastic method to do that type of analysis as opposed to creating that in a tool that isn’t as versatile yes as as that fabric EC

53:11 versatile yes as as that fabric EC system right as you’re outlining or different tooling allows us to do guys can I can I can I one point here there was a moment in time when I was using powerbi desktop where I realized powerbi desktop is just SQL and I was like what and what happened for me was I was slicing and dicing data by dragging visual like Columns of data to visuals and or maybe I again I was making big wide tables inside powerbi on pages and

53:43 wide tables inside powerbi on pages and I could literally drag and drop data fields over to the filters and I had this Epiphany and I was trying to explain this to users I’m like why why do we want to do this I kept saying well it’s it’s like you’re you’re writing the SQL but you don’t have to do the SQL writing so so I remember some times where I was doing some SQL analysis and I was I just got tired of writing select these columns from the table do the join on I was like it’s faster for me to whip through and just drag stuff on the page make a really L like take your take your your power behavor report page double

54:13 your power behavor report page double the size so you have plenty of space to put lots of little tables on it throw all these tables on the page and just go to town okay what’s this column do okay put it here great here’s the sum of that boom drop it there done okay great now I’m going to have these other dimensions six dimensions on the page Great Click each one bing bing bing bing stuff starts changing in real time like that was like incredible for me I was like oh my gosh I’ve just written like 70 squl statements to dig into the data to figure out what’s going on but to your point Seth it’s not a pretty report it

54:43 point Seth it’s not a pretty report it didn’t have bookmarks it wasn’t all the the fancy bells and whistles it was literally just a shape of data where I joined these tables via relationships and then I wrote a whole bunch of you

54:55 and then I wrote a whole bunch of you and then I wrote a whole bunch of not wrote but I I use the UI know not wrote but I I use the UI to basically do a bunch of SQL conditions to figure out what the data is doing yeah I think I think ultimately where I where I land is I I don’t I personally wouldn’t try to create that experience in a shared method I would I would invest the time in training people on how to how to use rud rudimentary basics of powerbi desktop to get the most value out of it because there if we’re going to create that same experience as a notebook and visual

55:25 experience as a notebook and visual ations and tables and whatnot you can do that if you’re the author correct because you have the full control of what you want on the canvas and and ultimately the insights that you’re going to pull out of there if they know how to create a more formalized thing to share with other people you’re you’re magnifying the impact that they can make in the in the organization I I apologize Tommy I’ve been interrupting a lot Tommy say something I’m glad you guys are so excited about notebooks

55:56 excited about notebooks are we forgetting that like we’re literally trying to get the business users to try to start using powerbi for data analysis and then we’re throwing Jupiter in notebooks I’m sorry I think you have just completely split up the the audience here of the business user who has problems in Excel and we’re trying to give them got look have you seen have you seen the financial reports that are built in Excel from the finance team yeah and they’re very cool now to Learn Python it’s a leap I’m saying there not all data analysts will

56:28 saying there not all data analysts will be equals I think you’re right Tommy like you’re not you’re not gonna you’re not going to be able to walk into this and be like okay here’s a notebook everyone have at it and be fun like do fun things I agree with you 100% there but what I would agree with though is I would also say like yes but there’s a wide variety of experiences of people that there’s going to be a lot of people that do different things I you’re right but that doesn’t mean Jupiter notebooks which is a very subset of of a product language and tooling okay it’s just part

56:58 language and tooling okay it’s just part of the product just one aspect yeah I think you have a very granular view because someone has to know python where with powerbi if you can drag and drop you can why do you need to know python to use a jupyter jupyter notebook you can use SQL you can use SQL too great everyone so whatever language they’re going to use it’s a completely different way of doing it it’s meant for data engineers data scientist it’s not meant for the lay person it’s not I I

57:30 meant for the lay person it’s not I I disagree with that statement because I think we’re starting from a world of bi and and business individuals this is just yes so this is this is where I think Microsoft has not done Us justice in the fabrication of the different personas that are going to use this Microsoft powerbi thing like I think to your point Tommy a lot of people are going to come into the powerbi ecosystem they’re going to come in with the mindset of data flows data models or data sets and reports that’s what we’re going to start with all I’m

58:01 what we’re going to start with all I’m trying to say is if you’re doing analyst stuff if you have a true analyst going in and digging through the data eventually or the the variety of tooling that they’re going to be given is now way more capable and so we’re yes I agree with you you’re going to talk about the data engineer or the data scientist using notebooks but I love that even as a pretty technical I love the experience of it and I think I do too it’s democratization of these other tools there was a much larger barrier to

58:32 tools there was a much larger barrier to entry to get into those tooling because you had to go turn on synapse or turn on a spark cluster and configure things in Azure like all that’s now gone and so what we’ve done is we brought the barrier to entry to all these other great tools lower for anyone just because yeah but I agree with Tommy just because the they’re there doesn’t mean people are using them and this better than anybody else because power query has been part of Excel and part of our big pitch people don’t know that exists yeah right with powerbi is hey guys this has been there for like five years yeah but

59:02 like five years yeah but so it it is going to be a very a small subset of very technical or people who want to learn stuff analysts that are going to be pushing back into notebooks Tommy like do you agree with like H how would we produce that for the lay analyst then is it training them into powerbi desktop or do you think you can do this in a more formalized report that allows them the exploration they need I think there’s a lot of

59:32 need I think there’s a lot of opportunity with powerbi desktop because I don’t think I’ll half the people who are especially the audience of the self-service side they’re not modeling data who even know about some of those AI not co-pilot features with key influencers with the clustering with binning with the what if analysis like and some of those you do need to like build in the model but a lot of it you don’t it’s like just diving into like what did cause X Y and Z that again that

60:02 what did cause X Y and Z that again that predictive or not predictive but that more descriptive again there’s some limitations there that I mentioned but I would highly encourage like yeah look at all these features from the insights to the Q& A which is a lot better compared to telling someone hey you we going to use a notebook here the thing SQL oh now you’re going to have to connect to probably not we’re not going to make them connect to a database at that point because then that’s

60:32 at that point because then that’s getting out of the workflow of our semantic model well I think I think our workload is changing slightly I think we would then I think we focus on the Lakehouse or the The Lakehouse becomes the Landing part The Landing location for all of our data right so we no longer want to burden reporting systems Enterprise reporting systems we’d land all the data into our Lakehouse and from The Lakehouse then we handle and distribute the data from there oh my gosh Tommy look at the time we are like to go back to our early

61:02 we are like to go back to our early conversation we are old I was just getting started on this whole thing and this is a great question and oh my gosh we’re already at an hour already so I apologize I totally lost track of time I got way into this conversation and really think this is a great way to figure out how we can what we can push on the different again I think this is I think this is a a great Point Tommy you making here as well all right I apologize very random thoughts let’s do final thoughts here and then we’ll wrap it for today because this was

61:33 we’ll wrap it for today because this was a great question this was a great followup to what is an analyst how do we do it and hopefully we didn’t dance around it too much and hope we were able to bring out some other topical pieces around what does the data analyst look like and how does that impact what we do inside power. com so Tommy give me give me some final thoughts I I feel pretty good with what I said that’s like I just made the last the the shot so I’m good Tommy threw some shots Seth yeah I I think so one it is a great question two it throws out

62:03 is a great question two it throws out yes one of the most challenging things to produce in a reporting is a report that would allow the versatility that an analyst would need or want to do across data sets and that’s why I think I would I would land more in training them how to use powerbi desktop as itself like un unblock and unlock those features and functionality to them so that they can be a problem solver instead of creating

62:33 be a problem solver instead of creating more problems for you in the fact that you would have to like reshape and give them another page or another view of the data and and I think that’s the first step in getting them to the data access that they need for those aha insights or deep exploration and then to Mike’s Point fabric is a fantastic ecosystem that as I think organizations adopt it there are a lot more paths for graduating people up into earlier and earlier aspects of data and

63:04 earlier and earlier aspects of data and or deeper insights that they can they can generate on their own using coding languages and and things like that so the door being open is fantastic or at least there but I I would lean more into training them and getting them to use power VI desktop as the the analysis tool I like that as well Seth I think that’s a really good point I think I think my final thoughts here are people are going to be comfortable doing what they want to do right so if you’re an Excel user and you love Excel analyzing Excel is probably going to be your friend if you are comfortable in

63:37 your friend if you are comfortable in desktop and shaping data and moving data around you’re probably going to like playing around in desktop a lot longer it doesn’t mean there’s not a variety of tools you can use at your disposal U but to to go back to the question from Chris directly here is I think there is a place for analytics or analysts where you find those aha pieces of data it will need to happen you’re going to need people just to continue to wash through that data and figure out what makes sense I think what comes out of this aha moment potentially will turn into new kpis or new things we need to

64:08 into new kpis or new things we need to measure as a business that would then turn into more of a formalized goal and or okr type report but I think I don’t want to limit the scope of what an analyst can do and say we can only do okrs we can only do goals I think this is all part of the process I think it’s this is all analytics or analysis data analysis that has to happen people have to get in they have to see the data they have to push it around until they find what they’re looking for and then from there you decide what you do next so I like this this is great I think this is really good about storytelling with data

64:39 really good about storytelling with data and I love the fact that we’re in a very exciting space right now very thrilled to be continually working on powerbi and fabric super fun anyways thank you so much for listening to us for over an hour now we appreciate your ears and listening to us we thank you so much for your time we know your time is valuable we would really ask if you would help us out if you could help us out by sharing this episode with somebody else letting them know that they this would be an interesting episode to potentially talk through and think through what is an analysis and

65:09 think through what is an analysis and what should we be thinking about as a company maybe we spend more time in desktop helping people actually go through the exercise of finding the insights and the data with desktop it would be very very interesting Tommy where else can you find the podcast and you can continue to find us on Apple Spotify or wherever get your podcast make sure to to subscribe and leave a rating it helps us out a ton do you have a question idea or topic that you want us to talk about in a future episode which you guys have obviously done thank you guys for all the topics lately head

65:40 you guys for all the topics lately head over to powerbi tips podcast leave your name and a great question and finally join us live every Tuesday and Thursday a. m. Central and join the conversation on all powerb tips social media channels awesome thank thank you so much and we’ll see you next

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.