The Future of Power BI’s Visual Experience, w/ Miguel Myers

Rob Collie (00:00:00): Hello friends. Today's guest is Microsoft's Miguel Myers, who is directly responsible for both the palette of visualization controls that we use in Power BI, as well as what they call the reporting experience, aka, everything downstream of building your semantic model. When you're crafting the experiences that your report/dashboard users are going to use, that's Miguel's purview.

(00:00:25): Now, coming into this interview, I already knew that Miguel was very, very popular in the community, and now, I absolutely understand why. First, it's super, super obvious that Miguel really and truly cares about the things he's working on. This isn't just some job that he got assigned in the round-robin job assignment that is Microsoft. He's been on a collision course with this job for a long time. This is his calling. And I think you're going to hear that.

(00:00:54): It's so trendy to say that you're passionate about something. And most people who say that they're passionate about something are just talking in their own book. This is an example of someone who truly is passionate about the thing they're working on, about the thing that they're doing for us, about the experiences that they're crafting for us, and that in turn the experiences that we're able to craft for our users.

(00:01:17): Second, Miguel is transparent with the community in a way that is very, very rare. He's very open. He just tells you what he's up to, tells you what the roadmap looks like, and we actually talked during this conversation a little bit about one of his upcoming plans for, let's call it, hyper transparency. Now it's obvious why the community loves transparency. Knowing where the product is headed and roughly on what timescale like that breeds a lot of confidence, but it would be very, very easy for that transparency to not be a good thing for someone in Miguel's position.

(00:01:52): In some sense, it would be much easier if he held his cards much closer to his vest. Transparency with the community can really open you up to criticism, and also you can feel trapped by your own prior announcements and proclamations if you subsequently discover that you want to change path. So you have to be very confident and competent to be this transparent without it coming back to bite you.

(00:02:16): Which brings me to the third thing. This is an enormously difficult challenge that Miguel faces, and he 100% strikes me as the rare person who can navigate it. You might think that this visual layer stuff is like frilly compared to things like VertiPaq and DAX and Fabric, but I used to have a job very similar to his, not as complex as his, but similar enough that I can see all of the difficult interrelated trade-offs and dilemmas he faces every day.

(00:02:46): This isn't one of those problems spaces where you can zoom in on a single detail at a time, get it right, and then move on to the next one. You have to solve all the problems almost at once. You have to come up with solutions that thread the needle of all of these seemingly contradictory trade-offs. For example, enable fine-grain control. That's good, right? But don't make things super tedious. Another example, empower advanced users to create finely tuned and uniquely sophisticated dashboards while also making the environment approachable to first-time dashboard builders. Here's a good one, continue to improve things while also not alienating people who have become accustomed to using things that work the old way. Oh, and along the way bring consistency to a suite of different visualizations which exist for the very reason that each of them brings something uniquely different to the palette.

(00:03:41): Honestly, there were times in this conversation where all that complexity and interrelated trade-off started to really dawn on me and I started to imagine myself in his job and I got just a touch sweaty. See if you can hear it in my voice. It was a great conversation. We talked at length about many different dimensions of the future plans in this space. We debuted a new bit called Unfair Feature Requests. We channeled some questions from our team at P3. Oh, and along the way, he helped me take an oh-so-gracious victory lap on some of my internet nemeses of a decade ago. Mwah. Delicious. All this and more when we get into it.

Speaker 2 (00:04:22): Ladies and gentlemen, may I have your attention, please.

Speaker 3 (00:04:26): This is the Raw Data By P3 Adaptive Podcast with your host, Rob Collie, and your co-host Justin Mannhardt. Find out what the experts at P3 Adaptive can do for your business. Just go to p3adaptive.com. Raw Data by P3 Adaptive is data with the human element.

Rob Collie (00:04:53): Welcome to the show, Miguel Meyers, how are you today?

Miguel Myers (00:04:57): I'm really good. Thanks for asking. How are you?

Rob Collie (00:05:00): Fantastic and very, very excited to be talking to you today. The community, I will go so far as to say that I detect a buzz in the community specifically about you and specifically about the things you're working on, and that is not something that usually happens with software or product managers. There's an optimism in the community about the things you're working on. Are you aware of this optimism? Are you aware of the positive vibes that the Power BI community has about you and your work?

Miguel Myers (00:05:31): Absolutely. Yeah. I'm excited as well to see how people are engaged and I'm happy that they are contributing to the whole product development.

Rob Collie (00:05:40): Again, this is not normal, a good not normal thing. I said the name Miguel Meyers, our internal Slack, and people were like, "Oh." There was a palpable excitement.

Justin Mannhardt (00:05:51): The comments and questions are still streaming in real time as we're recording, I know you're probably without Slack at the moment, but I've got us for the latest interest from the team.

Rob Collie (00:06:01): But let's start with the basics. What is your current job at Microsoft, Miguel?

Miguel Myers (00:06:05): Well, I am the product manager for Core Visuals in Power BI, and now I am as well the PM for report experience. So what is report experience? Is all about the format pane, data pane as well, the on-object experience that has been launched, the future ribbon format section and everything that is connected to what the report creators use to build reports.

Rob Collie (00:06:32): Music to my ears. Having all of that in one unified organization sounds like a really good idea.

Miguel Myers (00:06:40): It was a really strategic move because before when I was just the PM for Core Visuals, it was really difficult for me to handle changes that we wanted to do because I was always going to Rosie who was the PM for report experience and she had a vision of helping new users while I am more into the advanced users, so many of the things that I want to do were blocked by this team. So now that I own both, I should balance myself, always protect the new users as well, but I can move faster with the Core Visuals without having so many blockers.

Rob Collie (00:07:15): Yeah, I mean if you own the controls, the visuals, you want to create a new chart control or an upgraded chart control, but its settings, the knobs, the levers that the report designer needs to manipulate on that thing. If that whole experience is owned elsewhere, you can end up with a total experience that's less optimal than when everything's on the same page.

Miguel Myers (00:07:38): That's correct.

Rob Collie (00:07:39): Miguel, do you know my background?

Miguel Myers (00:07:41): No, please share it.

Rob Collie (00:07:43): I used to have your job. The closest parallel to your job that existed at Microsoft when I was there is the one that you got today. I mean, yours is a lot richer of a job than I had because the canvas was much smaller, but on version one of Power Pivot, I was product manager, program manager for the majority of the client UX. Now, of course, that was all Excel, but in Excel add-in as well, and so

Justin Mannhardt (00:08:09): Generational.

Rob Collie (00:08:10): There's a lineage here. Rob Colley evolves into Miguel Myers. This is an improvement.

Miguel Myers (00:08:16): Yeah, well, I'm not sure if it's an improvement, but I want to think so.

Rob Collie (00:08:23): I do too, actually. I would prefer it to be an improvement. I need you to do a really good job. We're both invested in the same idea. There's a temptation for me, I'm not even sure what we're going to find out how much I'll be able to resist this, but also whether it's a good thing or a bad thing to just really put my software product manager hat back on and talk to you about all of that. When you said balancing the needs of the new user versus the advanced user, I'm like, yeah, that is the stuff. That is one of the most enjoyable things and difficult challenges of the job that I used to do. How do you keep the easy things easy? The simple things need to be easy.

Miguel Myers (00:09:07): And more importantly, without changing the current experience because people are reluctant to change. You need to always find that spot where, hey, I'm improving your productivity, but the experience that you are seeing is not as different as the one that you learned from, the one that you first saw.

Rob Collie (00:09:24): The legacy of past decisions, which in fairness are always the best thinking that was possible with the information that was available at the time, even if it's our past selves, it doesn't have to even be someone else that made these decisions. Your past self can be made decisions that are now current self legacy. For example, I tried to change the name pivot table in Excel when I worked on Excel because the name was a liability. It scared people off. It prevented less sophisticated users from even trying it. Like what is a pivot table? Doesn't tell you anything about what it does. Even the legacy of the name was too much to overcome.

Miguel Myers (00:10:01): Yeah, yeah, yeah, I totally agree.

Rob Collie (00:10:02): In the podcast that went live this week, that's already live, I teased that you're going to be a guest on the show and I said in my teaser for your episode that the visuals canvas is the total experience. Everything that you're in charge of is a layer at which innovation can basically continue forever. And it can be valuable innovation. People often think of innovation as like, oh, making the DAX engine twice as fast or the Power BI engine being able to work with the [inaudible 00:10:35] Lake Parquet format like technical things. But capabilities and usability, usability improvements at the visuals there, that's a forever thing. Let me just throw you the absolute softball question of do you agree that there's sort of an endless capacity for innovation?

Miguel Myers (00:10:55): For sure. I actually took seven, eight months to map up a plan for core visuals and when I present that to my engineers, they say, "Hey, do you know that this map that you have here is seven years of work?" I told them, "Well, but this is just for us to get into a nice spot after that, we still need to add more stuff." And it's not just silly stuff, it's things that are critical for analytical stories that users have to find patterns or see differences or et cetera. So yeah, I totally agree. This job that I have, I can be here forever.

Rob Collie (00:11:32): There you go. Seven years of work, you understand that your immediate response is probably something like, we best get started then.

Miguel Myers (00:11:37): Or give you more people to make it two years.

Rob Collie (00:11:42): Let me put my vote in that column. More people two years sounds good.

Justin Mannhardt (00:11:46): And I think it still pops up here and there, but I think one of the major criticisms of Power BI for a long time was the visual layer compared to the first movers like Tableau and things like that. And it's sort of unmistakable the amount of progress that's being made. The pace is really amazing to me, Miguel, because in my role here at P3, I don't make reports and data models all day every day anymore, and so I open up Power BI Desktop to work on something and you move my cheese man, all the things that moved around and features were different, but you get used to it and you're like, oh yeah, I remember how this used to have to work. So what would you say would be sort of the top things you're excited about on that roadmap that you can share?

Miguel Myers (00:12:29): The things that I'm more excited is improving edit interactions. Because if we really make sure that the visuals and slicers and filters can communicate with each other, there are a lot of possibilities on how people can work with data. Right now, it's a little bit rigid and there are so many blocks when you're trying to create some basic stuff. For example, if you have a slicer and you want that slicer to always select the last seven days, just because we do not have that infrastructure of how they can communicate with each other, it's just impossible.

(00:13:02): If you have three slicers, one for region, one for country, another one for state, and then you select something in the state that doesn't apply for the region, then everything is blank because we don't have the same hierarchy or infrastructure to understand that they are connected and they should filter with each other. And there are multiple things, we have powerful components already, but they are not exploring their full potential because the foundation has never built, but they're missing seeing that huge opportunity that relies on a strong foundation.

Rob Collie (00:13:34): I think that's great. I mean the edit interaction button, it's so confusing to me even today. I click that button, I'm like, ah, what am I going to get? I always forget which one I'm supposed to select, the receiver of the filters or the sender of the filters. I flip a coin.

Miguel Myers (00:13:56): Me too. That happens to me all the time. Yeah.

Rob Collie (00:13:57): I did this yesterday. I did. I selected the destination visual and did edit interactions when I really wanted the source visual, right?

Justin Mannhardt (00:14:04): The downstream impact is always the end user though because you get people using reports like you were describing Miguel scenario like, "Hey, I clicked this state and everything went blank and the report's broken." And the report developer is like, "Oh, well Rob, you need to go back and do this." Why doesn't it just know? Making that more intelligent, Miguel, if that's the foundational piece that you're excited about, that makes the end user's life so much easier, and even the developers, like you mentioned the last seven days, the hoops we jumped through in creating data models with calendar tables that have a row in there that is last week and it's dynamically changing. It's like we could make this easier and I think wow, massive opportunity just to empower more people.

Rob Collie (00:14:48): Right now it's sort of like the visuals are all just really good at their own island, but the opportunity to treat each visual like it's its own piece of software in a way, the last seven feature on a slicer, most recent seven, would be amazing, and your job I think is really difficult because at the same time it's really a million individual improvements. It's just this incredibly long list.

(00:15:16): You don't have one bottleneck place to go typically and say, we're going to improve this one thing. The engine team has a different problem. Oftentimes there's a central bottleneck that goes solve one thing and they just gain end-to-end benefits. So you have all these million different places where you need to make improvements, but at the same time you need to maintain consistency, vulnerability. This is an engineering challenge,

Miguel Myers (00:15:40): It's design engineering, and even something that I always say to people that they ask me about what is the hardest part of NPM and that you have to keep your reputation intact, so you have to be careful on the features that you launch, that you have data that proves that it was successful. If that decision was not successful, then engineers and designers are now reluctant to take the next step because they won't believe that you're driving the whole team in the right direction.

(00:16:09): It's a lot of things that you need to keep in mind. That's one of the major reasons as well why I really engaged the community because they can not only see quantitative data, they can see qualitative data based on sentiment analysis on how well people are perceiving the new features that we are launching. Trust me, even design struggles in certain points that I need to come up with ideas and help them to be unblocked.

Rob Collie (00:16:36): So let me clarify what I meant by engineering. I understand what you mean. You're aware of the internal organization of Microsoft design, engineering and product management, and I wasn't saying engineering in that sense. What I was saying about engineering was including the design, including all of it, I was actually meaning everything but what you think of as engineering. The product management and the design that's required to be successful in this space is itself, I think could be taught on par as an engineering discipline. It is software engineering. It is that difficult. It is that challenging. As unforgiving as building a bridge. It has the same sorts of high standards and requirements and it's got to handle all possible environmental conditions. I was going out of my way to emphasize that as an engineering discipline is from my experience, Miguel, and I think this is different for you today. Kids these days, you all have it so much easier than us old timers had.

Justin Mannhardt (00:17:34): Get out of here.

Rob Collie (00:17:36): I know I'm going to be 50 in a month, so I'm practicing. Okay. So when I worked on the PowerPivot team, that was the first time that that team, analysis services, had ever owned anything resembling a user experience like a non-super high-end developer user experience anyway. It's the first time that they had a consumer experience, first time that they had a power user experience of any sort. That was a hardcore engineering team. They're the ones that built VertiPaq, the ones that built MDX back in the day. What I was working on, the user experience was thought of as less than it was the second class citizen of that world.

Miguel Myers (00:18:15): Oh no.

Rob Collie (00:18:15): And I had come from Office where it was a first class citizen, so I kept trying to tell everyone or get the message across that no, the things I'm working on the product can succeed or fail based on these things, and they're very difficult and it's just as much of an engineering discipline as y'all over there optimizing the storage engine, which has a different engineering discipline. In your world engineering means that the people who are actually programming it. Anyway, it was a compliment, but lowercase e engineering, not uppercase E.

Justin Mannhardt (00:18:48): Got it, got it.

Rob Collie (00:18:48): So what I was going to tell our listeners, I think a lot of them don't know this is that when you say you're in charge of experience this, right? I think what everyone hears is everyone has to listen to Miguel and then you talk about your reputation. And so the product management position at Microsoft is deliberately designed so that basically almost no one has to listen to you. You have to win buy-in. If the developers, the programmers, the engineering team, if they don't believe in what you're doing, you're going to have a hard time getting anything done.

Miguel Myers (00:19:21): It will be almost impossible.

Rob Collie (00:19:22): Absolutely. And Microsoft learned this the hard way because in the early days of Microsoft, the best programmers were the ones getting promoted. And they turned them into the ones making product decisions and they weren't well suited for it, but all the developers' development manager was in their reporting chain, so they had to listen to ideas whether they were good or not. So they explicitly separated this function and created what they call the program management function and said, okay, we're going to have people who are our kind of people who can think about these sorts of problems, but also think about them through the user's lens and then they have to convince the developers. So you have all of the responsibility for this, but at the same time, not nearly the authority that most people listening would think. And is a very, very, very challenging job.

Miguel Myers (00:20:09): You're right, people think that because you're the PM, you have the authority to just take the decisions, sign the paper and get it developed. That's not the case. I always say I am just like the glue that connects the bridge between one side and the other one. But if one side and the other one they don't want to connect, then the glue is useless. So I'm just that piece.

Rob Collie (00:20:31): I used to tell my new hire program managers that reported to me that their goal was to never win an argument with the developers. Your goal is to get what you think is best for the users, get that thing to happen, but never win. You win an argument that's not going to help you long term. You got to get everybody sort of on the same side of the table. But people would tend to come in saying, so how do I get everybody to listen to me? I'm like, oh, no, no, no. That's not actually what you want. You want the right thing to happen.

Miguel Myers (00:21:03): Yeah, that's correct.

Rob Collie (00:21:05): Justin, you said that visuals were at one time considered a weakness of Power BI relative to Tableau. I don't use Tableau, so do we think that's still the case?

Miguel Myers (00:21:19): Where we are in comparison with our competitors, in terms of visualizations, we are moving ahead, gaining, and trying to be ahead of the game, but I think we are moving in the right direction. Before I was PM, I was just a report designer consultant for the customer advisory team. So all customers, anyone that were struggling or having issues with designing reports and using data visualizations, they were coming to me and I was just helping them.

(00:21:53): I was actually a pretty active report designer. I was designing reports for companies even traveling to their headquarters in different countries to work with their team and roll up my sleeves and get into the job, make sure that everything works fine. That's where I realized, hey, this is a problem. We have a big problem here. It's not fair for anyone to spend three hours trying to get a visualization behave normally or as expected and even praying to any God that nobody else will break it after you are leaving and then starting all over again.

(00:22:32): So I myself try my best to reach out to the product team and say, please, have a glance of what we need to do with the data visualizations. Just please give me a chance if not being the product manager to be there just to advise someone in that side of the planet to help them. And it was after four years that I had a chance to finally join the team and now I'm coming from experience as well. That's something else that I said to users, "Hey, I am not just a product manager. I am one like you that once were in your side, I know your pain." Even as a PM I'm still forcing myself to design reports to make sure that everything works as expected, but because of this change, not because of me, because the awareness of the importance of improving data visualizations is what is taking us into the competition. I think we are leading in the right path.

Justin Mannhardt (00:23:31): It's fun to reflect on the frequency by which you would encounter those types of objections from a customer. We want a visual to work a certain way. We want a report page to work a certain way and they'll show you another BI product working this way and like you said, okay, now we go and we're doing all these crazy workarounds with bookmarks and funny things in the model and measures to sort of make it behaviorally work. And I love how you said that and hope it doesn't break. That's certainly trending in the right direction. You encounter that a lot less often. It still comes up.

(00:24:09): The other day someone was like, I'd love if I could synchronize access across two visuals or the examples you brought up about the slicers, like that stuff where it's like, I can do this over here, why can't I do it in Power BI? But where I always put my chips on the table, when Microsoft decides something is a problem, they can muster so much energy to that problem. It's just exciting, Miguel, there's a lot of work to do and you got a lot of fans here at P3 that's been made clear today.

Miguel Myers (00:24:39): When I joined the team to be a product manager a year and a half ago, I said to the team, hey, I'm going to start adding a bunch of settings, things that people always wanted, but this format pane that you have here and whatever other settings that allows us to modify them, I'm just going to let you know that if 20 options were a problem and I'm going to add another 50 or 60, I have no idea how everybody's going to handle this. Everyone is going to be overwhelmed.

(00:25:06): So I was sending that message in February, April, June, July, another in September, and when they told me in November, now you're going to own that space. I said, I'm going to make sure that we have a plan to improve these horrifying, overwhelming controlled settings and keeping them light. So then when a new user arrives into Power BI Desktop, they won't see that you can now change the pattern of lines by specific dash gap pattern which is coming next week. That's super advanced. Why should I see it with a bunch of other advanced items? So we are planning to really concise this experience and make sure that the way that users create reports is more pleasant and not that overwhelming.

Rob Collie (00:25:55): That tension between the simple things need to be simple and the advanced things possible. I mentioned it earlier and something that we struggled with a lot on Excel because the temptation, especially in a software development organization is to treat everything as equivalent. Like every single setting has the same amount of importance as every other setting and therefore you end up designing for the advanced user, only for the advanced user. The way I feel every time I open up any sort of Adobe product, it's like, okay,

Justin Mannhardt (00:26:27): I am advanced.

Rob Collie (00:26:30): There is no training wheels version of this. This is straight into the deep end. This is why I don't really use their products very much. I just bounce off every time. There are three categories and maybe you even think that there's more than this. In my view, there's three categories of places where the types of improvements that you can make, in no particular order.

(00:26:52): One type of improvement is making things more learnable, making things more approachable, more discoverable. The second category I have here is reducing the amount of time and labor and tedium that's required for a report designer to achieve a given result, even if they know how to do it. The first one is knowing how to do it. The first category is knowing how to do the thing, making it discoverable, how they do the thing. The second thing is now that they know how long of an effort is it and how frustrating of an effort is it to get the result that they want?

(00:27:22): And then the third category of improvements are in terms of end results that can be delivered to the report user, the types of experiences that can be built, expanding the universe of what's possible there. So a lot of the things you talked about in terms of the edit interaction stuff, I'm just going to cherry-pick an example out of my personal experience.

(00:27:44): With the table visual I found myself recently wanting a find feature on the table visual that didn't filter the table down to that one row. I want find like find on a webpage find where it just scrolls me to that row. I want to find where I am or where a particular product is in the list, but I still need to see it in the context of the rest of the table so that the conditional formatting still tells me what I want, right? If Microsoft doesn't give me that feature, I can't deliver that experience. I don't know of a way to deliver that experience to my end users no matter how hard I work. There's no find that just scrolls the table. So that's an example of that third category.

(00:28:24): Does that match your understanding of categories or would you say that there's even a fourth category?

Miguel Myers (00:28:29): No, no, no. The same order. I have them as well in the same order. You are a hundred percent right. The first one is we need to make sure that everything is organized and people can discover them intuitively. We need to maybe provide some [inaudible 00:28:44] marks or giving opportunities for somebody to type something and we will let them know where they can find it, but we will make sure that the path is way clearer than what it's right now.

(00:28:55): Then the other one you said, you are a hundred percent right, is all about how fast that is. An object is a good example. An object was designed for new users to make their life easier, but then in reality when I did an experiment with even the new designer that is going to now become the owner of that section, we are actually making more clicks than ever. This thing that's supposed to be for new users is making my job longer than expected.

(00:29:28): So yeah, of course all the settings even in Power BI, hey, the new core visual that you can add multiple cards to one container to change one thing from card to card, even just in the form of pane, moving away on object, just the format pane itself is a painful experience going back and forth between one setting that is called apply settings to, because you have to go to that dropdown, select the card, then go down there and change the color, go up there, open up the apply settings to drop down and select the other card and back and forth, back and forth.

(00:30:00): They have to be easier to manage, easy to see, easy to understand the states that you're applying. If it's hover, press, everything needs to be way faster so then you're more productive. And you will feel even more motivated to use it because you will know that you're not wasting your time to do a silly change. And the last one, as you said, okay, you are giving me this amount of new line styles. You're giving me a lot of... Now I can even make my columns a new color in the field and now an outline. Why in the hell should I use this? It's useless to me. Well, the tool needs to give you, hey, this is good for these reasons. If you want to make a report fully accessible, we can automatically make these tweaks. So then you can see the utilities of the slicers and the metrics, et cetera.

(00:30:49): So we can guide users on not best practices, but possibilities that depending the report specifications, they can take advantage of it and they will know exactly how and where to find it. So those are the three pillars that I have. They sound that easy, but they are not.

Rob Collie (00:31:06): No.

Miguel Myers (00:31:06): The advantage is, all right, now that I own the experience and the core visuals team, what I can ensure from the core visuals team is give me all the ingredients. So imagine the core visuals team is just that this team that is going to the farm, picking up a bunch of vegetables and spices, and now they are going to deliver this to some restaurant. The report experience team is the restaurant. Their job is all this bunch of ingredients they have to be stored, they have to be processed, and they need to be transformed to serve to their customers. So I'm continue moving ahead with the core visuals, continue adding more stuff because I know that these are important ingredients for the perfect recipe and I will have to deal with the experience team to make sure that this is transformed properly. It's quite a challenge, trust me.

Justin Mannhardt (00:31:57): When you're describing the friction, even what is seemingly simple, maybe in an end user's eye, I just want this behavior, this simple behavior idea, the friction between that end user and the report developers or an author supporting them, having to push back. It's like Miguel, I know it seems simple but it's not, and we've got all kinds of other priorities. That's a huge area of opportunity in this whole game.

Miguel Myers (00:32:21): Yeah, that's actually a good one. For example, I myself take some time to go over Fabric ideas and social media channels and I asked one of my previous coworkers from the customer advisory team, he developed this app where I can just mine all the data. It puts them into this app that I can read and then classify. So in this roadmap that I explained before, that is almost like seven years of work each milestone I put a code. For example, people are asking to apply conditional formatting to line charts, which right now it's impossible. So I put that particular request into a set of other requests that fit into a big milestone. And then big milestone, I introduce a code. When I go and classify and read and I'm saying this, this is real, I read comment by comment, and I need to not only read it, sometimes even I need to try to decode it.

(00:33:16): It is so difficult to understand what they're asking. I need to really understand the meaning of it. And then I put the label, okay, this one belongs to this milestone. What I'm looking at it is there's a lot of people asking for that one, for making lines accessible for conditional formatting. They want the line to automatically turn red if it's negative, et cetera. So what I did is I started exploring and investigating all the components that will make that happen. And what I found out is that before even doing that, we need to have a infrastructure that manages borders, transparency, fields in a range of these settings across visual shapes, which is columns, bars, ribbons, lines, the shade areas that lines sometimes have and then markers, because error bars is a combination of bars, lines and markers and obviously error bars will be taking that.

(00:34:15): So people were saying, "Hey, why are you doing these improvements for columns? This is useless. We were asking for line charts forever and you're not hearing our feedback." What people don't know is that in order to do that work, right, and they will be happy with the power capability that I will bring them I need to make sure that the columns are working fine because the same system that I need for columns, they have to apply for the lines. And right now it might sound silly or you might not understand what I'm saying and maybe only graphically I can share why or show your audience why is a connection but the community don't see those details. So sometimes they get angry, "Hey, you're doing things that we are not asking." And no, there is a lot of connections there that before were not highlighted that now I'm mapping and I know that if I do something before time, it will screw things up.

Rob Collie (00:35:09): Sometimes you have to build the framework and have the first, essentially it's not really a unit test, but it's for the first proof that it works.

Justin Mannhardt (00:35:18): Right.

Rob Collie (00:35:18): You got to build one flavor of it before you can go build the others. But yeah, I can see how from the outside people will be like, why is that more important than my line thing? What are they doing? What are they doing back there? And it's hard to know, right?

Miguel Myers (00:35:34): Well, not anymore because that's other thing that I'm changing. In the next months I'm going to share a dashboard in Power BI that contains all the roadmap, all the seven years of work and the connections, the dependencies. So if they see, for example, we're going to launch next week, table layouts. So the people table, as you know have three different layouts. Compact, what was the other one? Compact, outline and tabular. So that's the foundation of how we will treat the totals, subtotals, run totals, columns, and row headers because they change entirely on those three structures. So if they are asking, "Hey, why do we need this?" "Why are you guys doing this list thing?" In the dashboard when they go to the next most then they can see that there is a dependency that requires that foundation. So that dashboard will tell you everything will tell you if you want to read the documentation about it, if you want to read blogs, if you want to... It will give users everything they need to be involved into this process. Full transparency.

Rob Collie (00:36:34): That's fantastic. Yeah, this definitely confirms for me that things have advanced.

Justin Mannhardt (00:36:40): Yeah.

Rob Collie (00:36:41): The new Rob Collie, the holder of my job is better than the previous holder. I would've never been that transparent.

Miguel Myers (00:36:50): Even Arun, I asked Arun, I asked Amir and many other ones, "Hey, I want to share all this information, but I'm scared that it might not be a good idea." And they told me, "No, go ahead. Engage your community. You're doing a great job." And that's as well giving me a lot of motivation. It's not only because I work for Microsoft, but the fact that they are giving me that freedom to not walk through the same path of any other PM and they're saying, "Hey, this is your job. Be creative. Show us what you can do." I don't think that this is something that I was able to find before anywhere else, and that's something that really every day motivates me when I wake up to continue putting the extra mile because I know that they're appreciating that fact.

Justin Mannhardt (00:37:35): That's the key though, right? You might as well get out there in the open with your customers and say, here's what I'm doing. Here's what I'm going to do. And if they love it, that's great for you. If they don't love it, that's great for you too, right? Then you can say, okay, we maybe need to think about what's going on here.

(00:37:54): I think Microsoft with Power BI especially, has done a great job being very transparent about roadmaps in general over years. But the level you're describing Miguel of, I'm trying to educate the dependencies and the interconnectedness of the major improvements we're trying to bring to the reporting experience, I think that'll be wildly fascinating to a lot of us just to understand what's really going on in the product. So kudos to you for doing that.

Miguel Myers (00:38:21): Something else that if you have noticed that we are trying to do is we are developing dashboards, like designing dashboards with the new capabilities. So then, hey, here he is, and people are really happy that once they see those reports and they explore what we did, they're like, holy cow, I didn't know that that was possible. I read the blogs, I know that this is a new functionality, but I didn't know that you can go that deep. So that's another section that I learned from the customer advisory team to educate and evangelize users.

Rob Collie (00:38:50): You mentioned when you're planning to share this, when are you tentatively planning to share this?

Miguel Myers (00:38:54): July. We have already the data. We have the design for the dashboard to make it super organized and easy for anyone with not even much experience using Power BI to work around the huge amount of information that is there. And as well, I'm already coming up with a new plan of how we will do documentation from now on. So that's going to be completely redesigned. So then it works really well with the dashboard. Whatever we develop, they can click on a link and they can go right into the bunch of settings or properties that we are adding, so then it's easy for anyone.

Rob Collie (00:39:25): When you say it's a dashboard, do you really mean it's a dashboard and not a report?

Miguel Myers (00:39:30): Oh yeah, it's a report. It's a dashboard.

Rob Collie (00:39:33): Got it.

Miguel Myers (00:39:33): It's a report. It's a report. According to Power BI is a report.

Rob Collie (00:39:40): Okay. Yeah. See, this is one of the problems, right? Eventually I just stopped calling them reports because the world requires you to. If you call them reports, people think they suck.

Justin Mannhardt (00:39:55): We don't want reports, we want dashboards.

Rob Collie (00:39:56): So my recent experience getting back into the visual layer, because most of my Power BI work, and I don't do that much of it in general, but most of the Power BI work that I've done over the last, I don't know, five years let's say, has been internally facing, sometimes even just for me. So I don't tend to care too much about the quality of the end user experience. But then when I sat down to build a set of, let's be clear, dashboards, reports, dashboards for my hockey league about all of our hockey stats, suddenly I had to put on my full consumer hat. I had to build on behalf of people who don't care about Power BI, have never seen it, needs to be really Fisher Price, easy for them. And that woke me up to a couple of things. Number one, honestly just how far we've come since I last built something at that level of quality, I was honestly super, super, super, super pleasantly surprised and impressed at how far the experience has come in terms of what I'm able to deliver.

(00:41:01): Even just little things like seemingly little things, the insane flexibility in the tool tip capability, like the tool tip field for visuals. I didn't run into any limitations this time. I mean, I was able to do things like on bar charts where sometimes there's a number, sometimes there's a label, sometimes there isn't. I've got a tool tip field made out of DAX that goes and checks certain things in the data and sometimes decides to display a text label and other times, and it's just so good. This particular application, you would ever understand why I would need such a thing. But once you get into it, yes, now you understand, okay, in this particular case, I need this weird, weird thing. And just over and over and over again, I was so pleasantly surprised at the quality of what I was able to build at the same time, just to get him exhausting.

(00:41:53): I don't want to make another report tab. I don't want to make another dashboard, not because the data work is going to be hard, but because the tedium, the exhaustion of getting everything lined up and let me shift gears for a moment and pick on a product that's not yours. This will be perfect. It's still a Microsoft product, but the other day I was so excited to try out co-pilot in PowerPoint and I went to it and PowerPoint has this great thing, the designer feature that just spits out some options for how your total slide can be formatted, and sometimes they're really, really, really good. So my expectation from co-pilot, given that they already have designer, is that I could do things like, Hey, I've just built slide three here, make slide three, format it the same way that slide two is formatted and PowerPoint co-pilot goes, nah, I don't make changes. No, I'm good. And I'm like, okay, so why is this being sold as a product? This is not ready to me.

(00:42:51): The most table stakes version of co-pilot for PowerPoint should be exactly that scenario, almost like format painter for the slide level. This is another one of those very, very, very difficult challenges for you. I actually don't know how you're going to do all the things that you have to do. I'm looking at your job going, yeah, I'm glad that's not me.

Miguel Myers (00:43:12): Or maybe saying it's so cheap to just dream about it, right?

Rob Collie (00:43:18): That's right. It's an idea.

Justin Mannhardt (00:43:18): Love that.

Rob Collie (00:43:20): When you said earlier in the conversation that you'd been pressuring for a while from the standpoint of your old job, you've been pressuring for a while that the report design experience would need to change in order to make all of this work, and then one day they turn around and said, okay, now it's yours. Was there even a moment of like, oh, no.

Miguel Myers (00:43:40): Yeah, that happens with even this one with the report experience. When they decided to first redesign the former pane, I was like, okay, it seems like this is good. Okay, I can try it. Not very happy with this, but I think I can deal with this. And then an object came and I was like, oh my God. And now it is like, all right, so now it's yours. And not only that, but we will continue the train for a little bit while, so before even putting your vision, this will continue moving.

(00:44:11): And I am like, yeah, but part of the vision is to stop this and rethink. I need to stop it, and if I let this train to continue, it will be harder for me to make changes. It can be challenging at some point when multiple things were designed in ways, and this is what we discussed at the beginning, people take decisions like the pivot table you said, the name, and sometimes to override that work is not possible. So you need to deal with what you have. And sometimes what you have is against your own strategy, understanding how can I move things in certain direction when the train is moving?

Justin Mannhardt (00:44:50): It's the grand balance between you want to establish a great foundation for a long haul effort, wherever the external forces come from, be like, yeah, but we're going to keep going, so it's tough. Don't envy you. I'm going to just be cheap and dream about it. I like that. That's my role here.

Rob Collie (00:45:10): The thing I was rounding the corner to is just another incredibly challenging dimension of your job, of the challenge in front of you. On the one hand, we can talk about all these visuals, all these visual controls and their settings and their learnability of where the settings are versus their flexibility of how many settings they have, things that you can do with them, et cetera. But then you have to zoom back and go, but in the end, the individual visual controls mean nothing. They're just the things that you use to assemble the overall dashboard experience. Is there a quick access toolbar in Power BI? Can I pin anything?

Miguel Myers (00:45:49): No.

Rob Collie (00:45:50): Okay.

Miguel Myers (00:45:51): Well, there is something important to mention regarding how to make it easier as well, and based on what we were discussing in the report experience team, one of the visions we have is to do something like presets. I want to have a top navigation and I want to have them aligned in this way. So you can open up a gallery with multiple navigations, multiple things that you can just click, drag it, and all settings will be there.

(00:46:18): With that in mind, one thing that we need to build that I'm excited as well to work on is something that is called templates. Companies, any user can create templates, and those templates can have multiple layouts. For example, I want this section at the top to have one long placeholder for a new card visual at the bottom. I want two placeholders, like two for visualizations, and then you can collect or create a bunch of wireframes with this collection of templates that you can just create a pack.

(00:46:53): I have seen websites where designers come up with these amazing packs and you can just buy them, and then you can get these amazing PowerPoint templates that has even a hundred pages that you can just go and select the ones that you want. So the vision I have is maybe even open up business for that allowing, just the same way as you find custom visuals and you get a bunch of visualizations that you can import in Power BI, add this new technology for templates and layouts and wireframes, so for people to create packs and then put them in some gallery where the people that they don't want to waste time making the tables look good or the slices. So the whole report, they just go there, connect that bunch of assets into the report settings, and they can just select, okay, I want to add a bar chart in this section of the layout pre-designed.

(00:47:50): But not only that, in the bar chart I'm going to add I have four different variants and I can see a preview of the things that will look, and then I can just drag the right conditions to make that visualization look exactly as the template. So then all those settings that we are in, you're not going to go and change them. The template and the settings already pre-established will carry them for you. That report experience will be the ultimate goal where all the ingredients gathered by the core visuals team and the restaurants set up by the core visuals team will finally give some revenue. In this case, talking about restaurants,

Justin Mannhardt (00:48:30): This made me think of an interesting fact about the world. My experience is there's a very small percentage of knowledge worker type jobs where that job should necessitate someone to start from the blank page or from scratch every single time. It's one of the most positive experiences I've had with generative AI tools is the blank page is not something, I do not need to win by starting from blank every single time. Anything that gets people to accelerate from blank. Because the blank canvas, I remember when I used to do Power BI trainings, you get a room full of students and you're going to teach them how to do Power BI, and a lot of them come from Excel, and the first thing you show them is this big white square. They're like, what are we going to do?

Rob Collie (00:49:19): We just get out our overlay with a grid and just put it over that white square temporarily. It's like this sheet of acetate, we put it over the screen and say, okay, everybody good now? Okay, blank blank. But with grid.

Justin Mannhardt (00:49:31): Yeah, blank grid. What you're describing is very different than what we have today as a Power BI template. Extraordinarily different. That would be amazing. So anything that just helps people not start from zero, this is a core philosophy of ours. The world moves fast, so great ideas come up and people want to attack these things and they're like, oh, it wouldn't be great if we had analytics that helped us with this problem. And then they get into it and they're like, oh, resistance, friction, confusion, uncertainty, not having the capability I need. And now the moment has passed and ugh. And so that creates a level of hesitation to try it on the next one. I want a front row seat to the new Miguel version of a Power BI template when that's ready.

Rob Collie (00:50:18): There's so many dimensions to this problem. Miguel, I assume you all have looked at PowerPoint's designer feature, and it's an obvious place to at least attempt to gain some inspiration from, but there's clearly some challenges with it. It's super, super frustrating to me. The designer feature generates all these ideas and I can press more ideas, but I can't point to a particular one and say, I like this one but... It just randomly decides if I just plop a bunch of images into the PowerPoint slide, it arranges them in a way that I think is very visually appealing, for example, but it sequences the pictures wrong and I want to rearrange the pictures, but the way that they've masked the images to give them blurry edges or jagged edges or whatever means you can't just move them around anymore. And it just kills me that I can't say that one but... Reorder the images.

Miguel Myers (00:51:11): And that's why the layout comes in place on the template because the layout will allow you to change the location and it will give you endless possibilities on how you can take that pre-designed page into something that accommodates to your own design.

Rob Collie (00:51:28): And so I'll just wait and see what it looks like. I mean, again, I don't want to get into attempting to design it with you here because you're already on it, and it's tempting. Let's grab a whiteboard...

Miguel Myers (00:51:40): Oh, yeah, but you're welcome. You're welcome. And the whole community is welcome as well. Actually, the entire [inaudible 00:51:47] formation for the report experience will be complete transparent as well. So we are not going to do this thing of here is the new thing, surprise, go and look at it or holding it into preview feature for two years. We're going to make sure that the designs, the mock-ups, and we want to explain the users what are going to be the first changes that they'll see, get some feedback or even sit in some webinars and have some open sessions where we can brainstorm is the path we want to take for the new report experience. So yeah, you're welcome anytime to bring some ideas.

Rob Collie (00:52:21): So in a former life after leaving Microsoft when all we had was Excel as our back in the power pivot days, I came to the sad conclusion that conditional formatting was formatting. So I would have these whole portfolios of workbooks and multiple worksheets per workbook, and there'd be pivot tables in all of them. And these sorts of rules, like whenever we have sales, format it as a data bar and maybe with some certain thresholds or whatever. And I had to eventually, for my own sanity write probably the most sophisticated set of macros I've ever written in my life, which involved defining brand new custom data types, like these structured data types that were brand new that I had to invent, whose sole purpose was to store the conditional formatting properties from all of the fields in a pivot table, remember what the measure name was, What the settings were, go through and inspect the entire conditional object for every last little setting. Remember that, and then go to another pivot table and replay those on another pivot table.

(00:53:32): So it was format painter, but at the pivot table level, respecting all of the settings I had made on each individual field for a conditional formatting. This is a tremendous amount of work, punishment. I don't really enjoy programming that much, so for me to be reading the documentation on how to create custom data, there was no way to do this with just regular variables. You had to create your own structured object from scratch to store all of this stuff. It was crazy.

(00:53:57): And it's just reflective of the same problem over and over and over again. Even though that was in Excel, this is sort of the same thing I'm running into is the format painter, really I just had this deep, deep, dark desire for the format painter to pick up the conditional formatting as if it was regular formatting. And of course, that's not the way it is in Excel either, but I don't have VBA in Power BI, otherwise I would've already written this.

Justin Mannhardt (00:54:22): And he is not asking for VBA in Power BI either, Miguel, just to be very clear, it's not a request from the community,

Rob Collie (00:54:30): And that's the last thing you need is another backward compatibility problem, right? When you move settings today, you have users getting angry about where they find them, but if there was a scripting layer and you move things around, you'd be breaking all kinds of customer scripts. I mean, I really badly want a scripting layer that allows me to build reports? Ooh, do I ever?

Justin Mannhardt (00:54:50): No, no, he doesn't.

Rob Collie (00:54:53): But I understand why you haven't given it to me, I would absolutely use that.

Miguel Myers (00:54:56): Let me see what we can do for that particular component.

Rob Collie (00:55:00): We could call this unfair feature requests, unfair in that we've got you on the spot. I've been in your shoes, right? So I know what it's like to juggle the things you're juggling, and I don't believe that I know better. I want to give you my oath, right? I need you doing what you're doing. I'm not going to be offended if my ideas never see the light of day, whatever. I'm just sharing my authentic experience as a user of your fine product. We have other unfair feature requests, by the way.

Miguel Myers (00:55:28): Oh yeah.

Rob Collie (00:55:29): Kellan, our president and COO wants to know if on the matrix visual, the plus sign indicators can be on by default. He made a very interesting observation was who wants those off? Who would prefer that?

Justin Mannhardt (00:55:46): You mean the expand collapse?

Miguel Myers (00:55:47): Yeah, the expanding collapse. Yeah. That's actually in our roadmap and what we are hearing from the community is that they want to have control that when somebody opens a page that contains a matrix with hierarchy levels, they want to have them all collapse or expand by default, or they want to have one category expanded i andrest collapsed and they want to save it so then users are not going through the whole process

Justin Mannhardt (00:56:17): Per user, yeah.

Miguel Myers (00:56:18): Yeah, so that's in our roadmap.

Rob Collie (00:56:21): Cool. There you go. Unfair feature request. What else do we have from the team, Justin?

Justin Mannhardt (00:56:26): Lots of praise.

Rob Collie (00:56:27): Absolute praise. Like the new card visual. Ed Hansberry is just a solid two thumbs up. Did he ever circle back Justin with what he thought was so great about it or he is like, no, I just told him it was awesome and that's enough?

Justin Mannhardt (00:56:40): We asked the team, things they would like us to ask you and ED just said, "I have no questions. I just want to say the new card visual is amazing." He just says "the ability to add a ton of info to a card. He's been able to get rid of massive clusters of old cards and text box and just have one visual and so you can change direction arrows all with conditional formatting." So he's just really jazzed about that and getting a lot of mileage about that. Maybe we hit Miguel with some of the questions from the team.

Rob Collie (00:57:09): Yes.

Justin Mannhardt (00:57:09): Do a rapid fire as a way to round this out maybe.

Rob Collie (00:57:12): Yeah, let's do that.

Justin Mannhardt (00:57:14): Even though you're responsible for the reporting experience and you still have to have a good answer to this, Chris Haas on our team, he wants to know what's your favorite DAX function?

Miguel Myers (00:57:24): Calculate.

Rob Collie (00:57:24): Are you aware that we did a DAX draft episode of this podcast where we took turns choosing functions and once someone chose the function no one else could use it?

Miguel Myers (00:57:34): No.

Rob Collie (00:57:35): Only they could use it, but we said from the very beginning that calculate was a community function.

Justin Mannhardt (00:57:42): It's off the charts

Rob Collie (00:57:43): No one could draft calculate, so if you can't draft calculate in the game, you assume that you already have calculate. What's your second favorite?

Miguel Myers (00:57:51): If.

Justin Mannhardt (00:57:51): Didn't Ed smoke you in that?

Rob Collie (00:57:53): Oh yeah, totally. I mean I was dead money when we started.

Justin Mannhardt (00:57:57): I can't remember the episode exactly. I just remember somebody drafted something like date add and it's like, well I guess since you drafted date add I'll draft sum X and this and then I'll be able to do date add my own way and I was just like, dang it Ed. So smart.

Rob Collie (00:58:13): Leanne, also, I think she got treat as, oh man, treat as.

Justin Mannhardt (00:58:17): That's a good one.

Rob Collie (00:58:17): Once you've needed treat as you don't need it very frequently. All right, so if that's fair.

Miguel Myers (00:58:23): Yeah, I think if is a DAX formula that really gives me good memories because it was the first one that I used to learn Excel. My teacher at that moment told the whole class, "Hey, you have to create this table and you have to use the if command." And that day I just took that task and I said, well, but this is boring. I'm going to make a video game using the characters I will draw and I will put them to fight one against each other, but each of these characters will have powers or special abilities that by using the if condition, I will cancel one and then it will tell you at the end, depending your selection, which one wins. And that was actually the first test that did that make me feel love with data Excel and DAX.

Rob Collie (00:59:06): It's also by the way, the reason why I repeatedly decided I did not like MDX the forerunner to DAX even if was really, really, really, really complicated, like astrophysics lesson just to write an if. I was like, no, this language sucks.

Justin Mannhardt (00:59:23): Yeah. All right. Here's one from Mitch that I'm interested in. Curious how you see Copilot impacting UI and UX and report design in the future?

Miguel Myers (00:59:36): That's a good question. People ask me that question very often and the way that I answer is this way. I think now the whole amount of features and changes that we want to do are more inclined to be used for Copilot. So I'm designing all these rich experience so then Copilot can do the job faster and easier for you.

(00:59:58): If right now Copilot is powerful enough and it can create amazing reports out of the box by itself, you won't get a richer experience or a very well-developed dashboard or report. But with this amount of work that we want to do, we want to make sure that Copilot is more powerful to bring these automatic creations without you putting so much effort, but it's keeping this technology the important keywords to create that solution.

Justin Mannhardt (01:00:29): Great. I like to just bold this because of my AI like hype train in a way, generative AI systems, it seems like a big leap is there's a middleware problem where the generative AI system can communicate with an intentional design within the software product itself to do things. And so that mindset you have of all the work we're doing, if we do it right, we're going to be able to actually make this come together. So that's good. Christina, she says, I love Miguel. Christina Ferrer says hi. She recently joined our team, which we're very excited about. Hi Christina. She says she's so excited for the allowance of multiple tool tips per visual.

Rob Collie (01:01:12): What does that even mean? It sounds awesome, but what is it?

Justin Mannhardt (01:01:14): What does that mean? And then when might one see this come down the pipeline?

Miguel Myers (01:01:19): Yeah, next week. And there were two problems that we were trying to solve at the same time when we were building this structure, and that is when you create any visualization, you have to bring the fields to grade it. For example, a line chart in the X axis you have to bring a field, in the Y axis another field, and if you want to split that line into multiple lines, then maybe a legend field or multiple fields in the Y axis. So then you can split multiple lines within the same X axis. So there are multiple ways to create your line chart. But those fields that you use no matter what in which way you want to set it will go automatically into the tool tip. So you will see in the tool tip the date or the fields that you use for the X axis, the Y axis, the legend, everything will be there.

(01:02:06): And multiple users are telling us, Hey, that information, I can see it even I can turn on labels and I can see the value there. Why should I see it again in the tool tip? I just want the tool tip to show information that are not there visible, that are more hidden, that can give extra insight into that particular data point. But the system doesn't allow you to do that. You're stuck with that.

(01:02:28): Not only that, if you have lines, imagine each line represents a country, but you are in Japan, so maybe the other lines are only referenced and Japan line is the only one that really matters for you to track. So the problem is when you hover into multiple lines, you can see a lot of tool tips moving here and there everywhere for each country and Japan is there unless you're able to find the right spot to hover it. So what this new feature is doing is going to give you full control on the fields that are going to be displayed, for which individual series, in the example for countries multiple lines, each line representing a country, you can say I don't want to show tool tips for any country, only Japan.

(01:03:14): So then even if you hover on top of Congo or any other country, nothing will show up unless you hover on the Japan line. That's the things that is coming, allowing you to have this amount of power. This structure is designed to go into a wild path and that is allowing you to have multiple tool tip types on top of each other. Let me frame this. When you design a line chart, you get a different tool tip than when you create a column chart. The way that we identify those tool tips is the line chart carries like the marker and then multiple series. And when you hover, no matter which data point we show you the entire column or category in that section. So then you can see all the dots or lines at the same time.

(01:04:02): So if I'm hovering Japan in August, the tool tip will display me all countries together along with Japan there are in August. So we call that one series tool tip. And when you convert the same visualization as a stack column chart, we do not show that. We show depending where you hover, that's the only tool tip that you can see for that particular country in that particular month. So that's data tool tip.

(01:04:28): I identify that we have more tool tips, we have the page tool tip or people call them as well, canvas tool tip where you design a page in any way you want and then embed that view into a tool tip. We have the text tool tip, which you can find in buttons when you create, go into the action and then open up tool tips and then text and then you can create a tool tip. When you hover that button, you can see a text popping up saying, Hey, click here and then it will take you to the sales report or whatever text you require. And then we have with the modern tool tip that we want to introduce as the new default, we have the action tool tip, which is a tool tip that when you hover and stay there for a few seconds, it displays at the bottom another section that carries actions like drill down, drill up, drill through, et cetera.

Justin Mannhardt (01:05:18): Ooh, love that.

Miguel Myers (01:05:20): Now imagine that you can combine the tool tip using these type of tool tips and have full control of how and where are they going to be displayed. So for the Japan, maybe I want to see the serious tool tip with a text tool tip describing some things that happen in that particular month and at the very bottom I want to have drill through capability. But for the rest of countries I only want to see a description of important information about the country, but remove the serious tool tip and remove the actions. So that's the control that we will give you.

Justin Mannhardt (01:05:52): Next week.

Miguel Myers (01:05:53): Next week you're going to have control over the settings, so you finally are going to be able to disable that data that repeats already in the chart. And the whole reason we are doing this, by the way, this is something that doesn't rely on the core visuals team, supposed to be the report experience team, but we need to move on the new core visual that somebody was happy about it. And the next step or milestone for the new core visual is to support actions. So then you can define actions for individual cards and say, if somebody hover on this visualization, I want to trigger a bookmark or I want to go to another report or to click on that URL hyperlink. So we need that component of customization from the tool tip to continue our roadmap for the new core visual. So that's why the core visuals team had to take that job and then we are just going to leave it there for the report experience team to continue the job, but we will move on the new core visual actions.

Justin Mannhardt (01:06:51): It just dawned on me listening to this, you answering some of these questions, I think most often when we're talking about Power BI, people are imagining companies using reporting and analytics internally, but ISVs and people trying to do Power BI embedded solutions, the work you're describing, they're about to get a massive boost to what they're trying to do. Because those operations, they're software companies, they're used to being able to do whatever they want to do and create whatever type of experience they're going after.

(01:07:23): We've worked with customers in this area and they run into this friction a lot. They want that experience you're describing, "Well, if I hover over here, I want something else to happen." I was like, "Well, no, you need to right click and then go down to this menu." And they're like, "Well, well wait a minute. No, I don't. That's not going to work for us." So that's another just call out on where a lot of your work creates a lot of opportunity and benefit for people is those types of applications as well.

Miguel Myers (01:07:48): Yep.

Justin Mannhardt (01:07:49): Kevin, Kevin has a question. What are your plans for improved design tools in Power BI such as better guidelines, fixed width and matrix columns, wireframes? I think you talked a little bit about that maybe with templates, but anything you want expand upon on that?

Miguel Myers (01:08:04): I think the most important part is this, those templates where you can just define the parameters and everybody else in your organization can use them. That's something that I'm hearing over and over again from big organizations where they need to create a set of templates with really well-defined structures and it's just not possible. Even with the template solution that we have it's not that great and the theme has some limitations as well that we want to start enriching. So I think the improvements that we want to do with theming and the new capabilities that we need to build for the report screens will finally get that page that everybody needs to improve their design experience.

Justin Mannhardt (01:08:42): That's great. Then I think the last one on our list is from Paulina. She was curious to know if there are any plans for a Microsoft official Gantt chart to come into the visuals list?

Miguel Myers (01:08:57): Oh yeah.

Justin Mannhardt (01:08:58): What other types of new visuals you guys are cooking up?

Miguel Myers (01:09:01): The first call is to start with partitions because the amount of settings that they share with, for example, with the Gantt chart, the Gantt chart has a secondary axis, which is the timeframe that you're going to scroll, and it requires the category side, which is all the task or categories that you are putting in the left side or right side, depending the part of the world you are. So we want to make sure that axis are refined, completely redesigned, and working amazingly to all partitions and as well the secondary axis and many other things like legends.

(01:09:37): We really want to polish the amount of components that partitions have so then when it's time to build the histogram, which is in our roadmap, I'll tell you the ones that we have, again, all this information will be shared to the community, so they will know better later. But we have the new waterfall that will allow you to have stacked columns, the new funnel chart that will be horizontal or vertical, and these are going to look like a real funnel chart like the one we have. This is going to look awesome, but requires the same components. We need to make sure that the connectors, the ribbons that connect one funnel to the next one works fine. But that's why we did a lot of transformations on ribbons and we will continue doing more.

(01:10:20): A histogram that relies on a secondary axis that create buckets, so a secondary axis will be redesigned and we have the stream chart and the Gantt chart as well. So there are plans. Please keep your eyes on the enhancements that we're going to do for axes, legends, shapes, lines, markers, so then it will give you a hint of what you will expect for the new charts. As soon as we reach these components, it will be much easier to handle new visual decisions.

Rob Collie (01:10:51): That's great. I realized I have an unfair feature request, another unfair feature request. Unfair in that, it's a fair request, it's just unfair to put him on the spot. When I multi-select two visuals, I want to still be able to use the arrow key to move them. That kills me.

Miguel Myers (01:11:07): So the report experience team have ideas of how we will need to improve the keyword experience, and that includes using arrows, includes if you want to select something like for example, you have this column chart with a lot of labels, total labels, data labels, and maybe you want to change data labels for one particular series. So just by selecting control and clicking that label, automatically we will show you in the format pane all the settings related to that selection. You can backspace to remove it, you can select multiple visuals and move them up and down with the arrows. The kind of control that you should have with your keyboard should be very intuitive, familiar and accessible for people to use, including that one that you were saying.

Rob Collie (01:11:54): I asked you for something incredibly simple, but you've got way more coming than that.

Miguel Myers (01:11:58): Yeah, it's more than that. Yeah. It's not just the arrows, it's a set of shortcuts or interactive actions with a keyword.

Justin Mannhardt (01:12:07): Well, that's one of the hallmarks of people that compete on Excel is like you just don't need a mouse.

Rob Collie (01:12:11): Yeah, I like the mouse for selecting, but then when I want to move things, I don't trust myself with the mouse.

Justin Mannhardt (01:12:16): Right. You need the pixels.

Rob Collie (01:12:18): Well, I want to maintain that vertical or horizontal, one or the other, right-

Justin Mannhardt (01:12:23): Yeah.

Rob Collie (01:12:23): ... their alignment with something. And so I know that the arrow key will be very, very, first of all, very gentle. I can move it like one unit in which I can never do with the mouse, but I also want to make sure that I'm either moving it just horizontally or just vertically. I've never even tried if the shift key gives me any sort of locking like that. Sometimes you can lock whether you're scrolling vertically, horizontally, whatever. But yeah, the arrow keys, as soon as I multi-select, sometimes I've actually multi-selected a bunch of controls, grouped them just so that I can move them with the arrow key as a group and then ungrouped them. I don't want to have to go through the group and ungroup.

Miguel Myers (01:13:02): Yeah, of course. And with these moving things with arrows, there's another idea we have for the canvas itself, for the report page. I don't like to call them canvas for some reason. I like to call it report page.

Rob Collie (01:13:15): Dashboard page.

Miguel Myers (01:13:16): Dashboard page. Yeah. It is better, right? Even better. So for that one, we want to do something that is called grid lines. That already exists, but the grid lines we have is useless. It was just an image placed in the background that resizes and moves and it's not great. What we want is to have some square grid line that you can control and say, I want two hundred lines vertical and five hundred lines horizontal. Each of those lines together will create squares, and when you move your mouse up and down, it will go from square to square. So then you can guarantee spacing, margin, all that. Having control with arrows in groups will be very handy.

Rob Collie (01:14:00): Just those two improvements, the grid line thing and the move with arrows, I'd crawl over broken glass to get to that version of Power BI.

Justin Mannhardt (01:14:09): You don't have to.

Rob Collie (01:14:11): He's just going to give it to me and he is going to put it in his dashboard and his roadmap and everything. Here's the controversial question. It's not even really that controversial. It was an ongoing fight that I had with a corner of the internet for many years. I'm going to give you no middle ground. You have to choose one extreme or the other, not extreme, you have to choose one stance or another is the rules of this game, which is dark background dashboards as opposed to white background dashboards. Are they perfectly acceptable or a crime against humanity?

Miguel Myers (01:14:41): They are perfectly acceptable.

Rob Collie (01:14:43): Yes, that's the correct answer. Okay.

Justin Mannhardt (01:14:45): He can stay.

Rob Collie (01:14:46): We're on the same team. For a long time, there were a number of people very vocal on the internet, but this is a core part of their professional identity is insisting that anyone who created a dark background dashboard didn't know what they were doing and it was dumb. I never understood this. The reason I thought that ask this question, Miguel, is because I briefly glanced at your LinkedIn today and I saw you sharing someone's work, they were doing some sort of Street Fighter report design. First of all, I just saw the animated GIF go by, I didn't have time to engage with their post, was that Street Fighter report, was that something being done in Power BI?

Miguel Myers (01:15:25): Yeah, of course.

Rob Collie (01:15:26): Oh my gosh, it looked amazing.

Miguel Myers (01:15:29): Yeah.

Rob Collie (01:15:29): Really, really, really good. We're going to have to link to that person's LinkedIn post in the show notes. I loved it, but then the second thought I had was when I saw that, I'm like, oh, is Miguel sort of tacitly approving of dark mode dashboards? If he is, that's a huge win for Team Rob. Team Rob.

Justin Mannhardt (01:15:48): Congratulations. Miguel. You have been promoted to Team Rob.

Rob Collie (01:15:52): We already knew he was Team Rob. He was Team Rob++. He's doing the same job I used to, but it's a harder job and he's doing a better job of it.

Miguel Myers (01:16:00): Thank you so much. I like dark themes, dashboards, and I have designed multiple of those ones. I even designed a dashboard you can import the dark theme or light theme and automatically the entire report changes. Everything. I won't use the same blue for one and the other one, I will change the hue a little bit to make sure that the blue for the dark one will look better than the one for the bright one. Everything will change, title, size, everything, so then it looks good.

(01:16:30): Obviously there are challenges for both. When you create a dashboard that is light or dark, you need to know how to play the game, and that's why I think people are coming with that thing saying, "Hey, it's a crime for humanity," because what happens is they saw somebody trying their best to create a dark mode dashboard, they didn't nail the rules that they're supposed to use.

Rob Collie (01:16:56): I think that's probably the excuse. I'm going to be a little less generous with the crowd that was so... Because these people have gone away, they've hidden, they no longer say these things, right, but was a hill they were prepared to die on back in the day.

Justin Mannhardt (01:17:11): Maybe they did.

Rob Collie (01:17:12): That's funny.

Miguel Myers (01:17:16): That doesn't surprise me. When I just joined the team here at Microsoft, I had to design multiple reports that were live. One of them is the skateboard dashboard. You can even see it now in multiple blocks, so I designed that one and other ones before that one too. And you have no idea the cruel messages that I got from people saying that designing a report was insane. It was waste of time. It's a shame for somebody to come up with designs like that when dashboards were supposed to be simple and easy.

Rob Collie (01:17:48): So they came for you too.

Miguel Myers (01:17:50): Yeah, yeah. At that time, it was more often for me to hear that feedback in blogs and forums than people saying I think it's great that we are applying design to dashboards. Now, nobody says that. Now, it's the other way around. People are engaging others to think about design when they design those dashboards. So I think it's over time people change their mind.

Rob Collie (01:18:14): Yeah. The way I looked at it at the time, and still to this day really, is that even if there were some truth, let's say that in the Edward Tufty sense of the world, that there was an objective, let's say 5% advantage in terms of user's ability to perceive information properly on a white background dashboard versus a dark background dashboard. Let's allow that there's some small advantage, even if there were, which I don't actually really truly think there is, but if there were, if people are going to engage with the dark theme more readily, if they're going to like it, if they're going to be drawn towards it more, that's part of the perception funnel. It's just one step upstream. Getting people to engage with something is part of the perception chain, and I just felt like these people that were arguing so vehemently were missing the point, that engagement was part.

(01:19:08): You have to get people to engage first, and if people are going to be more compelled and interested in a dark theme in certain scenarios, well don't fight it. Get them to the table and then...

Miguel Myers (01:19:19): That's right.

Rob Collie (01:19:20): Yeah. I'm really glad. I was even one time, Justin, this was before you even at the company, I was presenting to a room full of Microsoft sales reps in Redmond in a two-day session, and I showed one of the dark mode professionally designed dashboards that we had and someone raised their hand and said, that's a badly designed dashboard because it's dark mode, and I lost my cool.

Justin Mannhardt (01:19:40): I don't believe that for a second. No, sir. Good golly. Poor soul.

Rob Collie (01:19:48): And that wouldn't happen today because I'm glad to see that that's gone away. Anyway, I just... Fine, it's a victory lap.

Justin Mannhardt (01:19:55): Victory lap.

Rob Collie (01:19:56): I like that we've evolved in that way. These things can be fun. They can be visually compelling.

Miguel Myers (01:20:01): Sometimes they can even be necessary.

Rob Collie (01:20:04): Yes.

Miguel Myers (01:20:05): One time it happened to me with one customer when they asked me to help them with the design, the whole report design, because they were having issues with the usage and the users. I did the whole workshop, gathering requirements, creating some flows to understand how they will use the report, coming up with the architectural information, everything in a professional way. But when it was time to do the wireframes, I came up with the conclusion that we needed to design a dark dashboard. The leader of the team was really offended by that, that person said it was a bad practice, but I was able to demonstrate that based on the information that I got from the stakeholders, they are looking at their dashboards nonstop because they are not only analyzing, but they're monitoring because it was a live report and they need to take actions and looking at light screens. It was like harming them.

Rob Collie (01:21:02): Eye fatigue.

Miguel Myers (01:21:04): And they were saying over and over again that they were trying to find something else to look at it because it was just too intense. So that person finally changed their mindset that it's not a design, luxury stuff just to make it fancy, depending on the design requirements, sometimes dark mode are a necessity instead of just a silly suggestion.

Rob Collie (01:21:30): So this is why all of the CIA movies and whenever you see the interior of some sort of navy ship or something like that, all the screens are dark mode is because of eye fatigue.

Justin Mannhardt (01:21:42): Yeah. They're staring at them all the time.

Rob Collie (01:21:42): There's a real reason for it. It's not just because it looks cool.

Justin Mannhardt (01:21:45): It does look cool.

Rob Collie (01:21:46): It does look cool, but there's an actual reason for it and I'm glad that we're now accepting of that. Well, Miguel, I really, really, really appreciate you spending the time with us.

Justin Mannhardt (01:21:57): Absolutely.

Rob Collie (01:21:57): I know our listeners will as well, and it's been a pleasure to meet you. I mean this sincerely, I am really glad that there's someone like you in the position that you're in. I feel very good about our future capabilities with someone like you. Again, you don't have the authority to do everything, but you have all the responsibility and I like that it's you. Our future is bright as a result. Yeah.

Miguel Myers (01:22:20): Oh, thank you so much.

Speaker 3 (01:22:21): Thanks for listening to The Raw Data by P3 Adaptive podcast. Let the experts at P3 Adaptive help your business. Just go to p3adaptive.com. Have a data day.