Episode 184: Where do people get stuck with Power BI?

Episode 184: Where do people get stuck with Power BI?

Episode 184: Where do people get stuck with Power BI? 560 420 Carlos L Chacon

​You’ve seen the visualizations and want to dive in–perhaps you were assigned the task. However you started with Power BI, chances are you might get stuck at some point. In this episode, we chat about strategies you can employ to help ensure success. Have a take on this? Hit us up on social media. We’d love to hear from you.

Episode Quotes

“The two groups I’ve run into a lot [are] people that either just, “okay, we think we want to do this, but there’s so many nooks and crannies to this tool,” and the people of, “oh no, we’ve already dove in and we realize that it isn’t a smooth gradient.””

“I think the big thing that I’ve learned from trying to learn and understand Power BI is that if you try to take an incrementalist approach, which works for T-SQL, you will fail.”

“I think it’s going to be very hard for them to change the path that they’re on, and so I expect that five years from now, this podcast is still going to be relevant.”

Listen to Learn

00:38     Intro to the team and topic
01:10     Compañero Shout-Outs
01:56     SQL Server in the News
04:50     Should SQL Data Partners do a Spanish podcast?
06:23     Two groups: the Lost and Confused and the Out-of-Control
09:33     For 80% of everything, “there’s a button for that”
11:48     You can’t learn Power BI incrementally like you can with T-SQL
14:48     Ideas for the boss, from a staffing perspective
15:53     SQL Data Partners is available to help when you get stuck
17:06     These issues are not likely to be fixed or improved on
20:17     Closing Thoughts


Music for SQL Server in the News by Mansardian

*Untranscribed Introduction*

 Carlos:             Compañeros! Welcome to another edition of the SQL Data Partners Podcast. I am Carlos L Chacon, your host, and I am joined today by Eugene.

Eugene:           Howdy.

Carlos:             Who’s traveling, getting ready for the holiday. We’re recording this, we’re interrupting our holiday schedule, compañeros, just for you. And Kevin Feasel.

Kevin:              Hello.

Carlos:             Welcome, gentlemen, thanks for being here again, today. Our topic, it says on my sheet that I’m looking at, it says, where do people get stuck with Power BI?  And so, I imagine Eugene will have some insight here for us.

Eugene:           Yeah, I have some thoughts on that matter, yeah.

Carlos:             Okay, so before we do that, of course, compañeros, we have a couple of shout-outs we want to give, and as always, please forgive me as I attempt to say your last names. Damon Rygiewicz, Katie Keene, Amanda Cinfio, Fernando Sanchez, Mark Misa, Tim Fox, Lukasz Szozda, he was giving a little love for the podcast, Jorge Perez, Gilda Alvarez, Austin Thomas, Mark Creery, and Praveen Madupu. So thanks, everybody, for connecting or giving me a little shout-out on social media, I do appreciate it. And as always, compañeros, you can connect with us. And if you connect with me on LinkedIn, I should say, I really enjoy when you actually leave a comment, it gives me something to go off of. But if not, I’ll just ping you and then we will go from there. Now it’s been a little while, but I wonder if we do a little SQL Server in the News. Now this isn’t hot off the presses, anymore, as we’re at the end of the month, but SQL Server 2019 was released.

Eugene:           Yeah, it’s exciting.

Carlos:             And so, yeah, that is quite interesting. In other news, SQL Data Partners is in the process of being able to sell licenses, and it’s interesting, at least the organization I’m going through, you have to sell the 2019 and then you downgrade. So, I thought that was kind of interesting.

Eugene:           Yeah. Which like from my limited understanding was a challenge a while back, because I used to tell people, “oh hey, with SQL Server, I think it was like 2012, you get a free cold standby or whatever you would call it”. And then the licensing changed for ’14 and ’16 where you had to have Software Assurance, and it was very confusing where it’s like, “okay, but I just want to buy another 2012 license”. It’s like, “no, sorry, you have to follow the new rules.” And it’s like, “wha– what?” Yeah.

Carlos:             Yeah, yeah.

Eugene:           It’s confusing.

Carlos:             And so this is interesting. I wasn’t quite sure how much demand there would be for this, and I’m not saying that my customers are pounding down my door saying, “we need to upgrade now.” However, the whole idea of data virtualization is a poignant one, and we tend to enjoy our little bubble with SQL Server, but I’ve had a couple of customers, shout-out to Rob Wagoner, who reached out and was like, “hey, I have data in PostgreSQL, and I’d like to be able to connect to it. I’m having trouble with drivers, and upgrades are a pain. Hey, can I use the PolyBase tools in 2019 to connect?”

Kevin:              If only we knew somebody who knew anything about PolyBase.

Eugene:           Right.

Carlos:             If only.

Kevin:              I could bring that person on the show someday.

Eugene:           Yeah, well, I think a book came out and that’s why the demand’s up, is they’ve been reading about like PolyBase Revealed or Explored or Exploded or whatever. I forgot the title.

Carlos:             Is is the book out now?

Kevin:              Not yet. So, as of Thanksgiving Eve, gonna have the proofs will be sent to me so that I can check them over and find all of my awful typos, or at least half of them. the other half I’ll leave in as Easter eggs.

Eugene:           That’s great.

Kevin:              And then, it’s supposed to be released in January.

Eugene:           Well, I can confirm you can pre-order it on Spanish Amazon right now.

Kevin:              Yeah, also, this will be way too late for anybody, probably, but right now Apress has a deal where all of the Springer and Apress computer science books are $7 for the eBook.

Eugene:           Seven bucks.

Carlos:             That’s for eBooks, right?

Eugene:           So, all of you listeners, either have a time machine, or set a calendar reminder for next year to take advantage of that.

Carlos:             Now it’s until what time? Cause I think we’re actually going to release this pretty quickly.

Kevin:              Cyber Monday.

Carlos:             Oh, it’s for s– no, we’re not going to release it that quickly.

Eugene:           We’re live-streaming right now. Everyone, go.

Carlos:             Yeah, that’s right. Okay, so yeah, that’s exciting as far as the new version being out, and so I’m interested to see kind of where adoption goes. And the other, so I’m not sure it’s necessarily a bit of news, however, I’m looking for some thoughts, compañeros. I know Mariano pinged you about this on Twitter, Eugene. So Mariano, Lucas and Gonzalo, the guys down at Precision IT, we’ve been working together for some time, they’re in Argentina. We are considering doing a Spanish version of the podcast.

Eugene:           Oh, you guys are actually considering that?

Carlos:             Yes, and that’s–

Eugene:           I thought he was teasing me.

Carlos:             You kind of walked into a back conversation, yeah.

Eugene:           Yeah, yo comprendo un poco. I can be the comic relief.

Carlos:             There you go. Hey, maybe we’ll have you.

Eugene:           Well, I mean, it is a goal, because, geez, hopefully, I can say this on the podcast, but, at SQL Data Partners, we had an opportunity to do that project for, was it for the Argentine gov– it was for some thing, and I’m like, “oh, it would be good if I knew Spanish, because all of the Power BI files have to be in Spanish, so I’d better brush up.

Carlos:             Yeah, that’s right, that’s right. Yeah, we’re looking at that, so compañeros, I know lots of you, and all of a sudden, I’m not going to remember everybody, but we’ve got folks down in the Dominican Republic, in Argentina, in Mexico, we have quite the following. And Spain, as well. So, if you’re interested in a Spanish version, let us know, and we may accelerate our plans a little bit, there. It’s going to be different, because I’ve learned things on this podcast, the format will be slightly different, but we’ll see.

Kevin:              I think he’s saying it will be better over there.

Eugene:           No, no, no, no, there’ll just be like more bells and sound effects. Like I’ve listened to some Spanish language podcasts and their intros are a lot fancier than what I’m used to.

Carlos:             Yeah, there you go. Some cumbia music, something like that. Okay, we digress. Okay, as always, the show notes for today’s episode will be at sqldatapartners.com/powerbi.

Eugene:           Nice.

Carlos:             And so with that, without further ado, because I know that we’re almost getting in the way of Thanksgiving preparation for our guests, here. Where do people get stuck with Power BI?

Eugene:           Yeah, so I wrote a short blog post about this, and in retrospect, I wonder if I have like insulted any potential customers, I don’t know. But I described it as like two categories: the Lost and Confused and the Out-of-Control. So, Power BI is unique in a lot of ways in the fact that it’s self-service and it’s easily scalable. And so, that means that people’s on-ramp into it is a lot different than other products, in the sense that, you can get a free license, you can get Power BI Desktop, you can start playing around with it for cheap, you can scale up for cheap and just add people with the $10/month license. And so it’s different than, say, SQL Server, where, sure, you can get Developer Edition or Express, but a lot of times you’re going to be buying a $60,000 service to support SQL Server, and so, there’s a lot more thought that tends to go into it. So, at least in my experience with customers, there seems to be two big groups. The one group, the kind of Lost and Confused, are people that are dipping their toes. They think they want to maybe switch to Power BI, but there’s all these weird nooks and crannies, in part because of the way that Power BI is developed. It is developed in these two-week/one-month sprints and it makes for a really messy narrative about all of the advanced features. So, the basics is simple, they show you the Dashboard in an Hour, Dashboard in a Day, Clicky-clicky-draggy-drop. But then, if you’re in Enterprise, it’s, “okay, well, how do I license this?” And it turns out, there’s like seven different ways to license it and nine different ways to deploy it, and so, these are people that just want an hour, honestly, to cover the unknown unknowns. They just want to know, “am I missing something?” So those, I’ve described as the kind of Lost and Confused. They haven’t made an investment; they’re worried that there’s something they’re missing that may impede a successful deployment and really, they’re just dipping their toes in the water. The other group, and you know, we had a customer kind of like this, is kind of the Out-of-Control where, they’ve started using Power BI, because it’s easy to get started, but then the data model gets bigger and the logic gets more complicated, or they’re adding more reports. And so, they run into these issues where, “okay, I’m able to do the basics really easily. You know, I can make some simple DAX logic, I can add visuals. But how do I do year-over-year?” Or,  “I want to be able to select month-to-date versus quarter-to-date.” Or, “how do I make it faster, because my data model just got ten times bigger?” And so, everything’s going fine until it isn’t, and that seems to be the two groups I’ve run into a lot. Is people that either just, “okay, we think we want to do this, but there’s so many nooks and crannies to this tool,” and the people of, “oh no, we’ve already dove in and we realize that it isn’t a smooth gradient.” Those seem to be the two big groups, which is very different than what I run into, say, with like SQL Server.

Carlos:             Right. Well, there you go, and the answer is going to be very simple. You just call us up and we’ll help to get you unstuck.

Eugene:           Sure, yeah. Absolutely.

Carlos:             Shameless plug, right. Kevin laughs, so I’m curious there, Kevin, do you actually interface with business users? I feel like your role requires you to do that a little bit more.

Kevin:              No, I stopped talking to people in about 2014 and it’s worked really well for me. I just haven’t talked.

Eugene:           Good for you. Yeah, no, if you can pull it off. The problem is there’s people everywhere, that’s the challenge.

Kevin:              Right, so I just stay inside my house all day. Life is good. Nobody has figured it out yet. As long as nobody’s listening to the podcast.

Eugene:           There’s just a cardboard cutout of Kevin in the office looking very stern.

Kevin:              Right.

Eugene:           Saying something about like Gauss.

Kevin:              Yep. No, I actually I don’t really spend much time talking to business users. They do keep me securely locked in an undisclosed location. But as far as pain points go, I think Eugene’s got it nailed. There’s a point where you get from, “I was able to do this fairly simple thing pretty well.” Up to, “wait, I want to do something that I think is pretty simple. Like, I want to show maybe the top 5 and then everybody else.”

Eugene:           Right, exactly.

Kevin:              It seems like it’d be really easy, but then you start getting into it and it’s not that easy.

Eugene:           Yeah, and I think that’s a natural outcropping of how it was designed, because everything’s an optimization problem when you’re designing a tool or product. Just how you don’t see in real life, you maybe see in movies, submarine helicopters that can transition between the two, but its two very different sets of constraints. And in my opinion, Microsoft has optimized for what I’ve described as maybe the first 10 or 20 hours of development and for low-code business users. And so, you see this a lot, especially with Power Query and M, where for 80% of what you want to do, like Pareto’s Law, 80% there’s a button for everything. You want sentence case, there’s a button for that. You want a column based on the examples you type in, there’s a button for that. You want to parse JSON, there’s a button for that. But then you want to do something that’s in the 20%, like okay, you want to read from a WebAPI and you want to send authorization information and all of this stuff, and there’s not a button and you’re just out in the wastelands. And so, I think it, in some ways, it’s a very intentional design, whereby optimizing for the on-ramp of new low-code business users, they create this chasm when you need to get to the more advanced stuff.

Carlos:             Right. Yeah, I mean, so like most things, is that just, you know, you have to plod through a bit of that? I guess we’ve talked a little bit about training, I suppose in the past, but because it’s so new and things are changing, I’m not sure there’s a lot of great options there.

Eugene:           Yeah, I mean, I think it’s a mix of things I think, one, I suspect that the documentation and community blogging and that kind of stuff over time is going to get better around that. Like back in 2015, 2016, there wasn’t a lot, but, for example, Melissa Coates just released an updated diagram of all of the different components of Power BI and the whole ecosystem and stuff, and so I think you’ll see some of that coming from the community. But I think the big thing that I’ve learned from trying to learn and understand Power BI is that if you try to take an incrementalist approach, which works for T-SQL, you will fail. So, what I mean is, if you try to say, “okay, I’m just going to do just-in-time learning. I’m going to just try and Google it whenever I run into a problem.” You’re going to run into a wall that’s a lot higher than you expect and you’re very likely going to fail. What you need to do is intend on learning some fundamentals and theory from the get-go so that whenever you come to that wall, you’re more easily able to scale it. So, get a book on DAX, get a book on M, maybe even read some of the Microsoft whitepapers on some of this stuff, but just assume from the beginning that you’re going to need to create some of that scaffolding for yourself ahead of time so you’re able to scale that wall instead of running headlong into it, is my experience. And it’s very different than what I’m used to with T-SQL. T-SQL is so prevalent and so well supported that I could Google any problem and find a blog post and find a DBA Exchange post and solve it.

Carlos:             Some examples, or yeah.

Eugene:           Yeah, I need to split a string; there’s 20 blog posts. I need to be able to do a switch statement to create some bands; there’s a blog post. All this stuff. It’s not the case with Power BI, and in part because it requires a lot more mental models to understand what you’re manipulating, where again, you can clutch your way through SQL, in my experience, anyway.

Carlos:             Yeah, so it is a bit of a different mindset  in how you go on, there. Which also then makes it a bit more challenging for those that do just want to poke a stick at it, because if you’re going to say, “hey, how much time do I want to devote,”  so tha–?

Eugene:           Yeah, my boss would call that Poke and Hope. But I think this is a natural kind of consequence of their business model. You see it in a bunch of places, Microsoft is trying, with the whole Power Platform to approach those low-code/no-code users, the people who, they do really good with a VLOOKUP, but maybe don’t expect them to write C# in their entire lifetime. And I feel like I see a similar pattern with, say, PowerApps where yeah, you want to make a CRUD app that can read and write from a single table? Awesome, super simple. You want to make something approaching more normal app functionality? Now you’re in the deep end or over the wall or, I don’t know which analogy or metaphor I’m supposed to be using, but you get the idea. I think by trying to target these users, it’s just an unavoidable consequence, in a lot of ways.

Carlos:             So then let me flip the switch, so I’m a CIO, I’m a manager, I’m I’m the person responsible for the team right or for people.

Eugene:           Sure, you’re the economic buyer, so to speak.

Carlos:             Yeah, exactly. So then, does that mean from a staffing perspective or from a resource perspective that I need to be thinking about, “okay, well, who am I going to,” I don’t know if siphon off is the right word. “Who am I then going to allocate this investment to, so that they can be our go-to person?”

Eugene:           I think to some degree. I wouldn’t say that you need to expect them to be able to do all of the work, but I would say plan that there’s going to be one or two people that maybe you send to an all-day training. Like a pre-con or something like that, that you’re going to provide some resources for so that they can learn some of the higher levels of this stuff so that they can help other people out. So, it’s not like SSRS where, okay yeah, no, maybe like two business users and then a bunch of devs are going to be able to use it. I think a lot of business users are going to be successful with it, it’s just that, okay, it would be good if we had a couple expert– not even experts, but like highly proficient users to help other people out, I would say so, yeah.

Carlos:             Sure, yeah. Well, and again, so shameless plug here, but this is an excellent resource or an excellent scenario where having a partner, or I like to call them escalation resources. Somebody that can be part of the team, that can help, that’s maybe not necessarily delivering every single report, but that when your team gets stuck, you can reach out and say, “hey, guys, we need some–  or guys or gals.

Eugene:           Yeah, and get you unstuck.

Carlos:             That’s right, yeah. “Hey, can you come help us and engage with us, here?” And I think there are– you mentioned Melissa Coates already. I mean, obviously we wanted to be in that bucket. There are others that are in that vein that, as long as you create a model or get a model that would work for you, then it can make a lot of sense, I think.

Eugene:           Yeah, if you’re like one customer, just laying out track as you’re going and you want us to help you, it’s a bit difficult. Also, it’s worth noting, Melissa Coates just went independent, so she’s got, I think it’s Coates Data Solutions, which has nothing to do with the outerwear or anything like that. It’s spelled with an extra E.

Carlos:             Yes, but good luck getting her.

Eugene:           I’m sure she’s busy. She’s said as much, yeah.

Carlos:             Yeah. Anyway.

Eugene:           But, if you like my dulcet tones, I have some availability with SQL Data Partners, so give us a call.

Carlos:             There we go. Yes, there is that.

Eugene:           Very cool.

Carlos:             Okay, well, I think those are some good points. Is there anything else we should consider?

Eugene:           I don’t think so. I think just that I don’t see this changing any time soon. There was a different discussion on Twitter about Source Control and Power BI. I think Steve Jones was talking about how we live in this devops world now, and there really should have been perhaps some more thought given to how do we make Power BI support devops and that sort of thing. And I think the reason I bring that up is it highlights the fact that it’s one of the downsides of taking the agile approach to something, is that there’s certain architectural pieces that are very hard to retrofit back in. Like if they had designed the .pbix format in a way that wasn’t a zip file, that was focused on being a bunch of text files or what have you, then it would be so much easier now to put it in Git and put in Source Control. But it’s going to be a nightmare trying to retrofit that in, in reverse, and so, I think the same thing applies to all of those pain points we talked about. Maybe Microsoft’s Learn is going to start providing some additional learning materials for Power BI, but I think it’s going to be very hard for them to change the path that they’re on, and so I expect that five years from now, this podcast is still going to be relevant. Just a little bit different, there’ll be 13 different ways to deploy instead of nine, and so on.

Kevin:              Yeah, and that has been a constant pain point that I heard throughout PASS Summit this year, which was, “what is the CI/CD model for Power BI?” The answer is, there isn’t a good one.

Eugene:           OneDrive. The answer is OneDrive.

Kevin:              Yeah, OneDrive.

Eugene:           Well, you know what? Yeah, I’m actually poking around with a demo for Power BI Sentinel, so I’m excited to see if that community product is going to fill some of those gaping holes with that. But yeah, it’s a perennial problem, and again, it’s just whenever you decide that, “okay, we’re going to start with a minimum viable product and make these two-week sprints,” it’s great for being flexible, but it’s really, really bad for making certain architectural decisions. It’s a challenge.

Carlos:             Yeah, which we have talked about, and I don’t remember the episode number all of a sudden. I should always just pull up the–

Eugene:           Just edit it back in.

Carlos:             There we go, that’s right.

Eugene:           You’re like, “it’s in Episode 102.”

Carlos:             Amanda is laughing at us now.

Eugene:           She’s not laughing with us?

Carlos:             Yeah, I don’t know, she may be crying.

Eugene:           Just slice it back in. No one need be the wiser.

Kevin:              Dear readers, it was in some episode, just go back and listen to all of them and you’ll be fine.

Eugene:           That’s why the /powerbi, just watch them all. Listen to them all.

Kevin:              Catch ‘em all.

Eugene:           Yeah. Dude, I’m playing Pokémon and you pick a language and it says you can’t change it, so playing Pokémon Shield in Spanish, I’m that dedicated.

Carlos:             There we go. Here we go. So, while compañeros, we do invite you, particularly those who are joining us as they say, have been– like Lucas who reached out to say they’ve been listening for about a year, some of those original episodes. The first 20 or so, please forgive me, I was learning, I was very new, then. But we did talk a little bit more about that in Episode 170, Devops with Power BI. There we go, but as Eugene said it’s kind of a little cliffhanger for you, there. We are looking at a community tool and he’ll have his official review here on the SQL Data Partners Podcast–

Eugene:           Oh, I’ve got a deadline.

Carlos:             Whenever he’s ready to do that.

Eugene:           Oh good, okay.

Kevin:              It’s a lot of pressure.

Carlos:             In 2020. That’s right.

Eugene:           Yeah, yes.

Carlos:             So, I don’t know if you’re old enough to remember. Do you remember Barbara Walters?

Kevin:              Of course.

Eugene:           Not particularly, on my end.

Kevin:              “If you were a twee, what kind of twee would you be?

Eugene:           I’ve heard the name.

Carlos:             Yeah, so I’m excited. I wish Barbara Walters would come back around, and say, “this is Barbara Walters, and this is 2020.” Funny antidote.

Eugene:           There’s going to be a lot of 2020 jokes.

Carlos:             That’s right. Like hind sight’s 2020?

Eugene:           Yep, yep, yep. I’m ready for it.

Carlos:             Okay, compañeros, I think that’s going to do it for today’s episode. Thanks to Kevin and Eugene for being with us today. And as always, if you want to reach out to us, you can do so on social media. Eugene?

Eugene:           Yeah, so you can follow me @sqlgene on Twitter. My blog is sqlgene.com. And I made an Instagram, but I’m not telling any of you what it’s called.

Carlos:             Hm, inquiring minds. And Kevin?

Kevin:              You can find my social media information in the 14th copy of PolyBase Revealed that you purchase. The first 13 won’t have it.

Carlos:             Very nice, very nice. And compañeros, I am at LinkedIn at Carlos L Chacon. Thanks again for tuning in, wherever you might be, and we’ll see you on the SQL Trail.

1 Comment

Leave a Reply

Back to top