episode 180
Power BI is Back on the Menu, plus The Importance of Usage Data, w/ Gil Raviv
episode 180
Power BI is Back on the Menu, plus The Importance of Usage Data, w/ Gil Raviv
Gil Raviv has spent years shaping how we use data. From his early days at Microsoft, where he played a key role in developing Power Query, to his work refining how dashboards serve their users, Gil’s career is a masterclass in making data tools both practical and powerful.
In this episode, we sit down with Gil to dig into why usage data is the missing ingredient for truly effective BI. He explains how understanding what users do with your dashboards can completely change the way you design and deliver them. We also talk about his journey through the world of business intelligence, the lessons he learned along the way, and why he’s so passionate about building tools that meet people where they are.
Whether you’re a Power BI power user or just starting to think about how data fits into your business, Gil’s insights will get you thinking differently.
Episode Transcript
Rob Collie (00:00:00): Hello, friends. We're back in 2025 with our first show of the year. But we're not the only ones who are back because today's guest is one of my favorite people, the one and only Gil Raviv, who is also very much back. Gil is one of those rare people who I look at and go, whoa, they've seen even more facets of the data industry than I have. He has worn many hats in the Microsoft data platform ecosystem. He worked on the Excel engineering team at Microsoft PROPER, with specific focus on the Power BI related capabilities, including the now forgotten PowerView. He then became a very popular blogger and educator via his site, Datachant.com, before delving deep into the world of consulting, eventually becoming the worldwide practice lead for Power BI at Avanade. So, he's been all over our Microsoft related game board. Then though, he spent a fair number of years working for Amazon, helping them try to play catch up with Power BI. Now, those are my words, not his, before recently returning to "our world."
(00:01:00): That's right, Microsoft, and Power BI and all those good things are very much back on the menu for Gil. And we are super happy to have him back as part of our community. We talked about all of those places he's been, and I think you'll find the human factors very interesting in every stop of his tour. Recently, he's rediscovered his software development roots and he has now built and is selling a commercial solution for Power BI usage monitoring called BI Pixie usage tracking information. If you think a tool like that isn't interesting to you, you might decide otherwise by the time you're done listening. Because honestly, even I got quite a bit more excited about it than I expected to be.
(00:01:38): The philosophy that drove him to create this tool is 100% relevant to all of us. Whether or not we ultimately decide we're in the market for such a tool. Because usage data is immensely important to anyone who produces a software product and Power BI dashboards are inherently software products. So, definitely stick around for the part where we dig into that. And yes, our good friend, ChatGPT, makes an appearance along the way. So, we still managed to get our recommended weekly allowance of generative AI. The philosophy behind this product and the product itself were so interesting to me in fact that I think we're likely to have a part two of this podcast sometime in the next few months. How's that for a teaser? So, let's get into 2025, shall we?
Speaker 2 (00:02:24): Ladies and gentlemen, may I have your attention please.
Speaker 3 (00:02:30): This is the Raw Data by P3 Adaptive podcast with your host, Rob Collie, and your co-host, Justin Manhart. Find out what the experts at P3 Adaptive can do for your business. Just go to P3adaptive.com. Raw Data by P3 Adaptive, down to earth conversations about data, tech and biz impact.
Rob Collie (00:03:01): Welcome to the show, Gil Raviv. It is fantastic to see you again. It's been too long.
Gil Raviv (00:03:07): Hi, Rob. I'm so excited to be here. Really, this is amazing experience for me. And to some extent, I'm closing a circle today with you.
Rob Collie (00:03:16): Yeah, so you've been on a heck of a journey, both in real world terms and in professional terms. I'm really looking forward to retracing that with you here today. I think a lot of people who are listening to this are going to recognize your name. You were synonymous for a long time there with Power Query, but you've taken a little bit of a side quest. You and I first, I think, started interacting when you were still at Microsoft.
Gil Raviv (00:03:42): Yeah, that was a fundamental chapter in my career. Hearing all your knowledge, following your blog, I think you visited us at one point and it got me a motivation to build something and show you. It was a LinkedIn report in Excel, I think even. And I was motivated to see how you would be impacted by that because you are our heroes. Microsoft as a product manager, any feedback that came from Rob Collie and Power Pivot Pro, was kind of a bible. Even if you cannot quantify it with something more measurements about data feeds or anything that... in terms of quality, whenever I could quote you on something, it saved me so much time. I didn't need to go through so many meetings and influenced my leadership about this is a decision we should take. So, that was the perception of what you brought to the table.
Rob Collie (00:04:41): Ah, glory days. Most people today who are working in Power BI have never heard of me. I believed in this stuff in a way, I believed in it before most people at Microsoft did because I had had the opportunity to go outside of Microsoft and apply it, and come back and say, "Hey, folks, that thing we were working on, it's really, really, really, really, really good. Way better than we even thought it would be." So, I was all in for a number of years there before it had gone mainstream. By the time it went mainstream, I was worn out from all the blogging and everything. I moved on to other things. And plus, I just felt like the world didn't need it as much. When I was the only person writing about it all the time, I felt like the world needed it. And by filling a need, it helped motivate me. Once there were a lot of people on board, that need wasn't really there anymore and it was a sort of sense of relief for me as well.
(00:05:36): The rest of the world is getting it now. And one of the stories I told about Ken Pauls for a long time was there was a brief moment in time when I knew something about Excel in depth that Ken didn't, and it was DAX. I wrote the book, and he read it and figured things out. Then he got better at DAX than I was and I was like, okay, that's the way the world's supposed to be. Ken's supposed to be better at everything in Excel than I am. So, yeah, I did. My wife and I, traveled over. So, you were living in Israel at the time?
Gil Raviv (00:06:02): Yes.
Rob Collie (00:06:03): And the beginnings of the Power BI cloud service were being developed mostly over there, right?
Gil Raviv (00:06:10): Yeah, it was funny thing where you need the team in Israel to bridge between the politics of two different teams, Excel and SQL BI on just... there is a little driveway across those two buildings and you need the people in Israel to have the dialogue.
Rob Collie (00:06:27): Yeah, other side of the world.
Gil Raviv (00:06:29): Yeah, that was kind of funny.
Rob Collie (00:06:31): Also, Power BI, the potential of it was so clear to Microsoft that they needed to pour resources into it, expand the size of the team really dramatically, in a way that they really couldn't do via incremental hiring. They just grabbed a whole campus essentially in Israel. That was a great trip.
Gil Raviv (00:06:53): Was it your first time?
Rob Collie (00:06:54): The only time I've been to Israel or really even to that part of the world. Met a lot of great people. The Euro soccer cup was being played at the time.
Gil Raviv (00:07:03): I couldn't know.
Rob Collie (00:07:04): You don't have the sports. The sports receptors weren't installed in you.
Gil Raviv (00:07:08): It's a big chunk in the brain. It's not just the receptors that is probably missing.
Rob Collie (00:07:12): So, how'd you find yourself on that team? What's your origin story that got you to that team?
Gil Raviv (00:07:17): That's a good question that show how much my career was random. The number one reason why I wanted to join that team was the group product manager that led the product managers at that team, Igal Edry, his name, he was actually my manager in my first career when I was out of college. My first full-time career as a software engineer, he was my manager. He was a VP. We were a small startup for cyber security solution, and he was my manager at the time. And I worked with him I think around two, three years, and at several points he was so inspirational that... since then and then he moved to Microsoft. So, for more than 12 years, I just wanted to beat Microsoft just to work with him. He was just very amazing person for me in my career. Sometimes it's just a single sentence that I remember that was helping me think about how I should position myself and how to provide value. So, that was really important and that's why I wanted to be in that team just because of him.
(00:08:26): But it turned out that prior to that role, I was a product manager for seven years. Before that, the main focus was solutions that were about preventing cyber threat attacks, everything about cyber security, network security. But I turned out to focus more about the dashboarding, and monitoring and analytics. So, it was always like most of my career in software engineering and product management, data was part of the product. To be part of Excel with Datadog, it felt like, oh, yeah, I work with Excel a bit. I do pivot tables and these things and I think data is so important. And then because of my experience in dashboarding, I was the victim of being the product manager of PowerView.
Rob Collie (00:09:15): Oh, right. Right. I'd forgotten about that as most people have.
Gil Raviv (00:09:22): Around the nine months of my job role in Microsoft was to just prioritize things that we have no engineers to follow. And in the end, the most significant thing that I've done is a product manager on the PowerView is just to decide to write a spec of how we will just disable it and turn it off.
Rob Collie (00:09:42): Yeah, you got to put it to bed.
Gil Raviv (00:09:44): Yeah.
Rob Collie (00:09:44): So, the power of working with a leader who helps you be the best version of yourself, that resonates with me. If I heard it right, if I understood it correctly, you worked with Igal for two or three years. There were seven years in between?
Gil Raviv (00:10:01): Even 12, around 12 years.
Rob Collie (00:10:03): That many years later, you were still ready to jump at the chance to work with him again?
Gil Raviv (00:10:08): Yeah.
Rob Collie (00:10:09): Wow.
Gil Raviv (00:10:10): He was very instrumental in the way that I perceived the leadership principles. As a person it was so inspirational that I thought this is someone that I would definitely want to keep working with. And it's not like to work for, it's like to work with. And obviously Microsoft was a good brand even if in the more early days where it wasn't as sexy as Google, et cetera, but it still was a very good place to be. And I thought this would be a good option, so that's how I came up.
Rob Collie (00:10:45): So, PowerView is something much like most people who work in Power BI today have never heard of Power Pivot Pro, don't know what role I played for a number of years. Most people don't know about PowerView. They don't even know that this thing existed. It was a Frankenstein. For those who never saw it, imagine something that sort of a really, really, really primitive first attempt at the Power BI report surface that we use today. Imagine that being shoehorned into Excel in very rapid fashion. It did not take long to get PowerView off the drawing board into the product. As that product was being developed and you were the product manager for it, you probably weren't at the beginning thinking, "Oh, yeah, this is going to be a joke in the history of all of this," right?
Gil Raviv (00:11:32): So, I think when I just joined in the first three weeks, because everything was so overwhelming and I needed to learn so many things. Even acronyms, Microsoft internal things that I couldn't understand complete sentences. And there was all the technology as well. And the technology not of PowerView, but the entire Excel infrastructure, and components and how things are getting communicated. And not to mention features in Excel like DAX and Power Pivot that was so overwhelming. So, with all of those informations, I had then the impression that leading PowerView will be so cool because it's like the dashboards, and the visualization and the things that are really... I like that from my previous [inaudible 00:12:14], and I thought, oh, this will be a cool place to be. And then I realized that I was only responsible on getting the container operate with Excel and also all the interaction interoperability between that black box is PowerView and the other stuff that is Excel to do the wiring between those two. That was my responsibility. I would prefer to go and watch any spot that you would recommend instead of doing that.
Rob Collie (00:12:43): That's not an exciting job. So, the things that were happening in the surface of the PowerView canvas, were those being developed and designed in Redmond?
Gil Raviv (00:12:50): Yes, the actual add-in was developed in Redmond, and even it was integrated also with SharePoint as well. My responsibility mainly was to go through a backlog of many, many items of things that doesn't operate really well and start to figure out with a very small team how we would keep implementing them. Luckily at some point, the decision was that this is not going to be the BI solution of Microsoft. And I had another product manager in the team that was really super, super smart and talented, and got me the Power Query responsibility. So, I kept PowerView just for the replication, and started Power Query, and that was a completely different story just because it's so powerful tool. So, it is really fun to be a product manager and the tool that you are responsible for is so empowering to the users.
Rob Collie (00:13:45): And Power Query even to today, as a feature of Excel, I mean Power Query is still blowing people's minds right now. There's someone right now as you and I are talking that's discovering Power... there's probably 10 people discovering Power Query for the first time right now and going, "Oh, my God, why wasn't I told about this? Where have you been all my life, Power Query?" I'd forgotten that as well. So, what was your role with Power Query?
Gil Raviv (00:14:08): Yeah, so Power Query was quite a big role for me at the time just because as a new person in the office organization and Excel, there were a lot of things to work on. And specifically we wanted to release in Excel 2016, the power Query functionality as a built-in technology in Excel, meaning it's no longer a kind of first party add-in. It would become a built-in functionality which would require for Excel features to work better. So, I was still doing all the wiring between the container of Power Query and Excel. It's not that I was responsible for making new transformations or changing the user interface of the Power Query editor, but I was able to say to the Power Query team, "Hey, this checkbox by default is on." I have people on my side that are challenging us to change behavior, so I needed to be the negotiator, the moderator between different interests of simplicity in general, functionality in Excel, and the more advanced power Query capabilities that needed to be bridged.
(00:15:18): Lots of things that were advanced that allowed me to enjoy the challenges, the technical, the product management side of challenges were interesting enough even though it was connecting things between different products. Undo, redo, macro object. The object model and all the VBA references. When you copy paste a table, what should happen between all those things. And there was a significant team that needed to build all the integration between the two components and make them native. So, as a product manager, it was challenging and a good first role in Microsoft to participate in.
Rob Collie (00:16:03): Yeah, that's one of those where, because Power Query was being built into Excel natively as opposed to as an add-in, all the touch points, the reason that you want to integrate it into the product is so that it does have a thousand times as many touch points with the rest of the product as opposed to just communicating through a keyhole. And that forced you to understand the user experience, what it was going to be like for the human being through the entire life cycle of it. Like it or not, and I suspect you liked it, but it or not, you are going to be forced to understand what Power Query was for, how it was going to be used. Even though Power Query itself was being developed in another team at Microsoft at the time, you couldn't treat it as a black box. And you really liked what was in that box.
Gil Raviv (00:16:51): Yeah, the box was really amazing tool and I enjoyed building with it. And that's why I had the motivation to help on the engineering side and the definitions to break the add-in and make it a built-in functionality in Excel. It's also like if I would stay in Excel for many, many years, that was a good experience because I needed to work with so many different people to get the approval of making all the functionality be a native Excel. Even at the time in Excel 2016, me and another product manager in the team were the only people that bought a new buttons in the ribbon of Excel, that were significant in Excel 2016 at the time. So, it felt like very empowering to be in that position to influence and help shape that few buttons, new buttons in the ribbon.
Rob Collie (00:17:42): That's some of the most prominent real estate in the world, buttons on the Excel ribbon. And especially if you're adding those particular buttons. This isn't an upgraded bold button, an upgraded fill down button or something like that. I mean this is opening a whole new universe. Which of course is then very carefully hidden for the people who haven't discovered it. It's still sitting there getting transformed.
Gil Raviv (00:18:08): Yeah, I remember the first time where I had significant interactions with some MVPs, Microsoft MVPs. The field at the time was to tell them we are going to change it to get and transform and to explain why. No, it's not a feature name, it's just like a use case. It's a behavior. We put it as a label, but it's not a feature name. And it was like everybody hated me at the time, I felt like.
Rob Collie (00:18:33): That is an ongoing tension in the design of these products. There's a holistic thing where you have to make Excel feel like Excel. You have to make Excel feel like it's just a collection of really great functionality. And you can't have stars of the show sticking out because it clutters everything. At the same time though, we all know that getting transform on the Excel ribbon should have a marquee of blinking lights surrounding it. If you don't know about this, you really need to be looking at this. There's always that tension. So, let's fast-forward a little bit. We sort of met again when you decided to move to the United States. How long ago did you move to the US?
Gil Raviv (00:19:19): A bit more than eight years.
Rob Collie (00:19:21): Wow. Oh, my gosh.
Gil Raviv (00:19:24): Actually nine years. Nine years now.
Rob Collie (00:19:26): Wow. So, more than half of my 15 years that I spent in the Midwest, you were in the US. Jocelyn and I have moved back to Seattle.
Gil Raviv (00:19:35): I listened to your relevant podcast about it.
Rob Collie (00:19:38): Oh, cool. Very cool.
Gil Raviv (00:19:40): Yeah, you always inspire me.
Rob Collie (00:19:41): Aw, thank you. So, I forget, the reason to move to the US, was it related to your wife?
Gil Raviv (00:19:48): Yes, my wife works in the pharmaceutical industry, and we got a great opportunity to just move with the family for a global role that she had. So, I just moved. And as the spouse, without even the working permit, I needed to wait nine months before I could start working. And when I got the work permit, I think with you, that was my first official working relationship that I had in US.
Rob Collie (00:20:20): And when did you start data chant?
Gil Raviv (00:20:22): So, data chant started a few months before I moved to US. I already knew I'm going to be unoccupied for quite some time, unemployed, so blogging would be a good thing to do. So, I started data chant. As a product manager, I already had some content that I created. I had my first post in a blog was actually in yours with a LinkedIn solution. And you are very pivotal in my aspiration to follow some of the things you've done. I thought I could never achieve even 50%, but some areas, some dimensions were so inspiring and I try to follow. That was data chant.
Rob Collie (00:21:10): I think that was when we first started talking. When you realized that you were going to be moving, you were like, "Oh, I've seen this before." Microsoft employee moving to a new location, not being able to do the job that he was doing before. Like Gil, if we really set our minds to it, you and I could have a self-deprecation Olympics that the two of us could compete to minimize our own achievements. If there's a sport you could get into, Gil.
Gil Raviv (00:21:37): That's the only thing that I can say I am better than you. That's the only thing that I would...
Rob Collie (00:21:45): Oh, I don't know about that. Wait a second. How many years were you full-time independent blogger, consultant, et cetera?
Gil Raviv (00:21:54): So, nine months. I was just blogging a lot. And then I started working as an independent contractor, also helping your team with that regard and getting some customers. And then four months later, I decided I would need some full-time opportunity to work somewhere and needed also the money. And then I moved to a big consulting firm, Avanade.
Rob Collie (00:22:17): Avanade is one of the weirdest species of mutant companies that it was a joint venture between Microsoft and Accenture.
Gil Raviv (00:22:27): Yes.
Rob Collie (00:22:28): Avanade is obviously still around, still very large. What was it like working there? Was the Microsoft linkage really obvious?
Gil Raviv (00:22:37): Avanade was, and I think I hope still is, very special in the atmosphere. And the perceived reputation of the people that this is a company about very smart people that are super knowledgeable about the Microsoft technology. So, if you have in a typical company, consulting company, a lot of people that are just spending time on the fluff, and speaking about value for customers and you see that it's not detached to reality, in Avanade we were about the technology. So, once you have very talented people that knows, understand the technology, it felt as a good place for me bringing the knowledge on the Microsoft stack on the Excel and Power BI. I didn't feel like working in a regular consulting firm, at least the first year.
Rob Collie (00:23:22): 99 out of a hundred consulting firms give the consulting industry a very bad connotation.
Gil Raviv (00:23:28): Yeah. Yeah.
Rob Collie (00:23:28): Consulting is a punchline and a joke because so many of the outfits, so many of the companies are wired essentially to waste their client's money. All they need to do is just be sadly just a little bit better than not having hired them. That's the level of success that they're looking for, the level value they're looking to provide. But so Avanade with its Microsoft focus, has the opportunity to do that. Whereas a lot of general purpose consulting firms, they'll take any project, they'll work with anything. And so, that leads to situations where they're essentially bluffing where they don't actually know what they're doing. We have that in common here at P3. We work with other tech because other tech has to plug into things. But we are very much a Microsoft centric operation just like Avanade. And so, you have an opportunity to be elite if you have a singular focus.
Gil Raviv (00:24:17): Yeah. And also this is something that I think I was really inspired by you and your team. You preach from the get-go out self-service BI with Power Pivot and Power BI, can turn on value faster than in any other tool. And it's not like six digit projects. It can be simple. You go, in a week, you provide amazing value. So, all these things I came ready with the things that I learned from you and your team. And with the knowledge, the technical knowledge of what you have in Power BI, I knew that this is possible. And then what motivated me in Avanade was that many, many people, even if they are super talented and smart, they were used to work on the enterprise grade solutions that are not Power BI in Excel. And my excitement was to make them discover how much we can do better by using Power BI not just for data visualization, but for everything else, and to find those opportunities to make the difference for the customer.
(00:25:20): So, for those in Avanade that may be measured in the performance review by being a regular consulting firm where you have the chargeability and you need to sell hours to customers, for those kind of people, I would probably create a lot of challenges because when I came, I provided the value to the customers faster with sometimes just me in the team. And there were sometimes 10 times more people around me making a lot of effort on other stuff, and suddenly I tried to show the customer, "Look, you don't really need them. Here's the solution." And that was a fun place to be, the troublemaker for my team, but the savior for the customer.
Rob Collie (00:26:09): There's a tension there. Being sort of the prophet of a new religion, spreading the new up-tempo style that the Power BI technologies brought. Again, most people, this is a testament to the fact that how effective this revolution has been, that most people working in BI today don't even know about the old way. That just occurred to me for the first time. Most people working in BI don't remember, weren't around in BI when it was so slow and so awful. The old enterprise technologies. Which of course, the reason for that is not that people have aged out, the surface area, the footprint of BI has expanded so dramatically. So many more people are involved in it today. You're opening people's eyes at this point in time within Avanade to the new way which is better, just fundamentally way better. And so, in that sense, you're like this bringer of truth, bringer of a new way and that's got to be very exciting. At the same time though, as you're pointing out, the way that you operate and the way that you're showing people how to operate represents a threat to their existing business model.
Gil Raviv (00:27:13): To some extent, yeah. So, it's more like for people that doesn't understand the potential, that can be a threat. But if you think about customer value as the main driver for making the consulting firm also successful, then that would not be the case. So, in many cases, I didn't utilize my special Power BI skills and excitements on replacing hours by other teams. It was more about the entry point for innovation. So, whenever you have a new project, you have new opportunity, you would use Power BI for the discover phase, and provide the higher value and then you can just grow on other technologies.
(00:27:55): So, you have AI, and Internet Of Things and so many other things that you would need to do on an enterprise level. Including even just migrating the data state to something more feasible on a large scale than just having slow Power BI refreshing multiple data warehouses. It may not be scalable for large enterprises. So, I was fortunate to bring the value in the early phases of strategic projects. And then all the other peoples that need to work many, many weeks and month to deliver something gigantic and monstrous, those things were still relevant.
Rob Collie (00:28:37): Yeah, that's true. I mean, enterprise consulting is going to always make its money on big infrastructure. If you truly had represented a threat to that, I don't know what would've happened. But because infrastructure work often does follow wherever someone like you or I goes, particularly in an enterprise environment, I could see that that actually was just almost like an accelerant. It helped pull through. I used to tell data warehousing people, "Hey, don't worry." Even though you don't need a data warehouse to get started, it tends to be that that kind of work follows where someone like me goes. It's just a lot more efficient when it happens and you end up doing many projects of that sort, each one more efficient than what they previously were.
(00:29:25): And in total amount of work in terms of plumbing is still probably expanding in the world. Even though people like me and people like our company are going around saying, "Look, it's all about the faucet, and we're going to start there and we're going to work backwards." It does not mean that plumbing goes away. It does not mean that we're annihilating that from the workspace. But I can still imagine that being very interesting tensions at various points in time. But the right thing happens.
Gil Raviv (00:29:51): The other thing that was interesting is that some of the key tensions will even between doing things on the Azure Cloud versus AWS or others. Many AWS. So, like when we worked with big teams, in many cases the data was already outside Microsoft, and they picked Power BI for data visualization because of the supremacy there. But as we came and do more than just... I came and do the Power Query magic, getting all the delays out, making better data quality, checkups with very large data just without waiting for some ETL team to do their stuff and just like, "Bring us the data will do it quickly," that opportunity was, I think, fundamental.
(00:30:39): Once we were successful in doing that, then our customers started to think, "Oh, this is how you do it in the Microsoft stack? Why would they stay on the AWS stack?" And then you have also, we were part of a bigger engagements with lot of people and teams from Accenture that developed other solutions and other projects that they owned on the other cloud. So, suddenly, we are creating a risk factor for everything that they invested. For years they invested a lot of things on AWS and suddenly, there is this little thing that they thought was shiny and sexy on data visualization, but it can do so much more. And suddenly, they started to saw people like us as a threat to the other business.
Rob Collie (00:31:26): That's right. It's a joint venture between Microsoft and Accenture. And hey, as long as you're using Microsoft Technology, which happens to be in most cases just better, that does represent a threat a lot of times to the other parent company. I hadn't even thought about that. I've only thought about Avanade through the lens of Microsoft. I hadn't thought about what it might represent to the other parent.
Gil Raviv (00:31:47): Now as a disclaimer, I've never seen a scenario where those kind of internal competitions leading any negative impact to the customer in the end. It was more about a big team of more than 100 resources that suddenly things are shaping up differently. And you have those internal conversations to keep the balance and the politics to result everything before things getting out of control.
Rob Collie (00:32:14): Yeah, that makes sense. When they started Avanade, they had to know that was coming. That's something that had to be baked into the charter, that we're not going to let those tensions hurt the customer. That's where reputation would be the end of the experiment if it were to gain any momentum. So, how long at Avanade?
Gil Raviv (00:32:33): Five years, starting with being the specialist in Midwest for Power BI, and then in the end ,leading the entire practice globally.
Rob Collie (00:32:44): Wow.
Gil Raviv (00:32:44): When I say leading, not the day-to-day work of every Power BI developer because it's a very distributed organization, but it was to be responsible for the capabilities of HR related 200 plus Power BI developers. But if you count also data scientists and data engineers that are also wanting to do Power BI. So, it was I think more than 300 people.
Rob Collie (00:33:10): Wow. So, that's a very different job for you than all of your other jobs to that point.
Gil Raviv (00:33:15): Yes, to some extent.
Rob Collie (00:33:17): That is people management on a massive scale. It's not a technical job anymore so much, right?
Gil Raviv (00:33:24): I was fortunate with that regard to have an organization that is very distributed. And with that regard, it's not that I'm responsible for the chargeability of a person in Canada or Australia. I'm not also responsible for them making the delivery for their clients on the day to day. I have no understanding of what exactly they do. However, in that multi-matrix organization, whenever they had either problems or whenever we wanted to take the preventative approach, the proactive approach to make them better, this is where I was responsible on the processes and the thinking, how we can do that.
Rob Collie (00:34:10): That sounds like a cool job.
Gil Raviv (00:34:12): It was. And I add around at least 50% capacity on just building things, creating things that others can use as added value to show the customers. So, demos, and capabilities and accelerators. Even just like an Excel calculator that will help you plan the migration of 10,000 Cognos reports to Power BI.
Rob Collie (00:34:34): It sounds like they created a job for Gil.
Gil Raviv (00:34:38): I think that was the case.
Rob Collie (00:34:40): There's no way that job existed or would've come about if they didn't have you. You can sort of see the outline of Gil in that job description.
Gil Raviv (00:34:48): Yeah. When you are in such scenario, perhaps you can retire with such a role, but you definitely cannot get significant promotions beyond getting into the top of that niche.
Rob Collie (00:34:58): Yeah, yeah, because everything above that is much, much, much more about the operational business and therefore, not terribly interesting to someone like you or frankly, like me. Then you made just a terrible turn to the dark side.
Gil Raviv (00:35:14): I betrayed my community. That was the feeling. The community, not my community.
Rob Collie (00:35:19): So, let's not sugarcoat it. One of the other big, big companies came calling. You're all the way up to your eyeballs in Microsoft for such a huge percentage of your career. How did they even reach out?
Gil Raviv (00:35:30): It was Amazon. They reached out probably one of my managers in Avanade that was then moved around a year later to AWS. Prior to that, I quit Avanade and decided to find something to do. And one of the things that I explored was to just start my own consulting as a freelancer. Because I thought, why would they need all the middle management, mid-tier, all the fact layers of people getting salaries just because of my talent? I would just provide the talent directly to customers, charge them accordingly. And that was the motivation at the time.
(00:36:09): And as I was thinking about what to do next, the realization of being independent was also very scary. And I started to feel the first silence of nobody is reaching out or almost nobody is reaching out, and I felt do I want emotionally to be in this position as independent? And then suddenly, they reached out. To be more accurate about it, I decided that I would probably have one thing that I would want to do, which is either to go back to Microsoft in the Power BI team or to find something that will be valuable in terms of the financial aspect. That was the case.
Rob Collie (00:36:48): Amazon came along and made you a Godfather offer, the offer you can't refuse.
Gil Raviv (00:36:54): That was a, "Here are your golden cuffs."
Rob Collie (00:36:56): Yeah.
Gil Raviv (00:36:57): "Would you accept them?" And I said, "Oh, that amount of gold. It's so shining." And now there was the cognitive dissonance of convincing myself why this is the best decision to make, and I made it.
Rob Collie (00:37:13): Well, what did you end up doing at Amazon? What was the nature of the job there?
Gil Raviv (00:37:17): They hired me to start a new practice for the competing product of Power BI Amazon QuickSight. And I was about to hire around 30 people to start a new practice. So, that was to some extent what I did in Avanade, now to do at AWS and their professional services team in a product that is relatively newer and has a lot of potential to go. And I thought, okay, this is going back to Microsoft 20 years back, and have the opportunity to go and start a team with 30 people that can then evolve to something even better. So, I thought on the managerial potential, this could be a good experience for me to try.
(00:38:00): And around when we reached a team of nine, the world changed in terms of the economic situation, and there was a hiring freeze that led around a year later to just realization that the specific team cannot exist anymore with the current objectives. And from building a global team of professionals, I needed to decide, do I stay in the organization as a delivery lead just being responsible for the ability of people? Or to find something else. And then I thought that's perhaps a good time to move back to product management like in Excel Power BI, just to do it on Amazon QuickSight. And I ended up in a role in Amazon QuickSight helping on the visions of making this service as powerful at some point to Power BI in terms of the get data experience. So, power query kind of things with different technologies, and different experiences and mindset, but to bring these use cases into that solution, that was how I found myself there.
Rob Collie (00:39:10): Interesting. So, how long did you work on that?
Gil Raviv (00:39:14): Close to two years, a bit less. And that was a tough time. It started with honeymoon like in anything that you're doing in life and career, but ended up with me having the trouble of doing my role as a product manager because I felt pity for the users. And that's something that I couldn't shake off from being an expert in Power BI and doing so much things with Power BI. And now I have a tool that doesn't match the features, I felt that I cannot get out of that equation.
Rob Collie (00:39:49): It's tough. I'm trying to be also very respectful and sensitive of the fact that you worked there, you worked with a lot of people there that I'm sure you liked and respected, and you still do. But there is a fundamental fact that it is very, very difficult for a software organization to come up with something on par with something like Power Query or on par with something like the DAX and VertiPaq engine.
(00:40:16): Most software, when you see it and you see that it's successful, is easy to copy. It's more like, oh, people like that. That's the thing that you gain from looking at someone else's software and go, oh, well, that's easy. You can copy that in a day. Not in a day, but we can. It's very straightforward. Like something like Outlook for example is an easier thing to replicate. Something like Word is an easier thing to replicate. But unless you have the exact right kind of team, and the size of team and the amount of time that it takes, you need all those things. You need the right talent, enough of it, and time to develop something like that. You just don't have that.
Gil Raviv (00:40:52): Yeah, it's impossible. And then you end up also with the mediocre copy.
Rob Collie (00:40:56): Yes.
Gil Raviv (00:40:56): And lucky for me, and also to be fair with so many talented people at AWS, this is not how Amazon operates. It's more about customer obsession. My main challenges were fighting against myself because I kept mentioning use cases and elements with Power BI. And every time they heard Power BI, I got a penalty on how they perceive me. Because when you are working in a company like AWS with all their leadership principles, you are not supposed to copy another solution. You're supposed to think, work backward, think about the customers. And then your product may evolve to be in many things very similar just because the use cases were important for the customer. But not because I knew how Power BI worked.
(00:41:45): So, my challenge was not to mention Power BI, but my fun and excitement was that I should take out the features, and look just on the use cases and what customers are trying to achieve, and try to innovate on how to deliver that, detaching myself from the knowledge of the other product. I was compensated and evaluated by taking it out. You need it as a benchmark perhaps in understanding which use cases you miss, but you are not supposed to think of copy catting the other product. And that was more challenging for me because how can I just forget all the things that I did with the other tool?
Rob Collie (00:42:23): One of the biggest problems with big companies is that you need rules like that to run a big company. You need guidelines. You need principles. We don't copycat. Then that runs up against your experience. You can't unsee the things that you've seen. And the things that you've seen work and that customers absolutely need. That principle is now being used against you in a way is like no, you need to not know those things. How the heck are you supposed to explain a use case and the fact that you've seen it work? You're not just conjuring it up, it's not just some good idea. You've seen it in practice, which is the most valuable thing. There's proof there. It's not a theory, it's not a hypothesis.
(00:43:06): But in order to share that you've seen it, and to leverage that real world experience and that real world evidence, you inevitably have to mention the context in which you saw it. And now you're the broader organization with its rule that it's now over applying. It's applying this rule to its own detriment and ultimately, dissatisfying you. Once companies get big enough, they almost contain the ingredients for their own failure because these rules have to be applied without much thought.
Gil Raviv (00:43:33): So, the day-to-day challenge was to just based on things that they've done. So, we are talking about problems and solutions, not about how to solve them in a specific technology or how a feature should look like. There is no copy catting here. It's only thousands of touchpoints that I built with Power BI, and I knew how I can resolve all kinds of things, not just data Power Query, even on the data visualization.
(00:43:57): Now I wanted to do that also with Amazon QuickSight. And I tried, and I squeezed and I learned a lot to be able to achieve. But then I find, okay, I cannot do that. So, now the conversation is, here is the use case that I really want to solve, can we prioritize that as part of the solution? And there came all the frictions because on the day-to-day, the Kind of persona was a bit different between Power BI persona and Amazon persona that is building more centralized solutions and embedded solutions, not the business people just building their own stuff. That's where I felt that I am very frustrated on making a big change and impact with the day-to-day frictions on, "No, this use case are important." And I said, "No, they are not." "We never heard about them. You are the first one telling them." So, that was too much on my side to fight against this team.
Rob Collie (00:44:54): I could imagine that being very emotionally straining and difficult over time. So, that's why you're back.
Gil Raviv (00:45:00): Yeah, that's why I'm back. So, I knew quite some time that I want to go back and with all the challenges that I had. And by the way, I kept building stuff with Power Query and Power Pivot in Excel because I was allowed to use Excel.
Rob Collie (00:45:16): Of course, yes. Yeah. You can't rule that out. They might want to. They might want to outlaw Excel at Amazon.
Gil Raviv (00:45:23): There is actually the Salesforce Quip, if you know that. It's like there is a spreadsheet, like a Google spreadsheet, but just like 10% of Google online that is so, so minimal with capabilities. And many people just use that, but still because it's very dysfunctional, we ended up with a lot of things in Excel. And that was allowed not the online version, the desktop version.
Rob Collie (00:45:52): It reminds me of the bomber era of Microsoft where it was not okay to have an iPhone and work at Microsoft.
Gil Raviv (00:45:59): The wind does not clear.
Rob Collie (00:46:04): Yeah. Yeah, you got to have a Windows phone, just no apps. You don't get any apps. So, how long have you now been back in the light with us? In the Microsoft ecosystem, you've got a new product that you've developed. Is data chant a thing again? Are you back to blogging? What is today? What are we on like Gil 9.0?
Gil Raviv (00:46:23): 9.0.
Rob Collie (00:46:26): Yeah.
Gil Raviv (00:46:27): Yeah, probably 1.2.
Rob Collie (00:46:30): Okay. So, what is Gil 1.2 up to these days?
Gil Raviv (00:46:34): I am independent since June last year, so it's six months now. And this will probably bring another interesting conversation about AI. But as I decided I'm going to live with the emergence of all the generative AI capabilities that I cannot follow since the day that they were introduced, I knew I can now do more technically with those tools. And all the knowledge that I forgot from my early days as a software engineer just came back. I left Amazon, and with all the things that I already build in the past with Power BI, I thought, okay, now I have better technical tools with the help of ChatGPT and all the others to get me ramp up on the things that I forgot, and be able to build things and write code to build solutions. And as I was starting playing with that and building stuff, I came up with the realization that I can create a product around a user engagement and adoption for Power BI.
(00:47:41): And within two months ago, I was writing Python code full time with it on automations of Power BI and the new file format in Excel. When you are a product manager in Excel and you know that there is a file format standard that you need to comply with, and now you understand all the challenges that this kind of standardization entails. Now looking at Microsoft releasing the ability to change the file format in Power BI made me think, okay, now I can build automations that will change those file formats. What would be the best thing that I can deliver? And then I started with the ability to measure engagement with Power BI reports from end-users.
(00:48:22): And as I built it, it turned into this kind of like amazing, gigantic magical creation that was in one end, super complex to build. But on the other end, when you show it and demo it, it feels like a magic to me at least. And I thought I must do a product from that. So, initially when I started it, I thought thinking like in Avanade as building accelerators, I thought this will be an amazing accelerator for me to offer consulting on top. But as I build it much more extensively, I thought I can be an ISV now selling that for customers enterprises. So, I evolved into not trying even to sell any hour of my time in consulting or making training, and just build, and build, and build this solution for six months now.
Rob Collie (00:49:11): So, there were times in your career where you were a software developer, but then you spent a long time as a product manager and as a consulting manager. So, not writing real code.
Gil Raviv (00:49:25): Except of M.
Rob Collie (00:49:26): DAX and M are real code.
Gil Raviv (00:49:27): Yeah, kind of. Yeah.
Rob Collie (00:49:29): When you think of software developer, you don't think of those languages. So, with the help of ChatGPT, you are able to very quickly, not just get back up to speed as a software developer, but you were able to get up to speed on things that you'd never been doing, because the world of software development has changed quite a bit since you left it. Just as a brief side question, have you spent much time with ChatGPT writing M and DAX?
Gil Raviv (00:49:55): I tried that a bit. Still today, if there is some complex DAX that I would spend perhaps five hours to achieve on my own, I was curious sometimes to see how ChatGPT would solve that. I must say I was never very satisfied with the results of that. However, in things like Python, because Python you have much more training data on the internet, the results are more promising. You get value faster through that interactions with ChatGPT.
Rob Collie (00:50:32): You ever tried it with M, with Power Query?
Gil Raviv (00:50:35): So, actually what I thought you asked me earlier, what are my plans with data chant? So, I thought if I'm going back to do perhaps blogging on sharing knowledge, I would just go back on every article that I shared on things to solve with Power Query using the UI and M, and try to see how ChatGPT would address that, and then compare. I think it may be better, but I'm not very optimistic about that use case. From what I observed so far, if you know already so much, it's just save you the Google search. There is a syntax that I forgot a reference to how to change something with dates that I forgot. I can use ChatGPT and it may be accurate on finding the example. So, it just sent me the Google search and reading an article about.
Rob Collie (00:51:27): What we've found is even just taking your M code from the advanced editor and handing it to ChatGPT and say just rename the steps for me.
Gil Raviv (00:51:36): That works. Things that are basic would definitely work. I would even say if you, let's say you have someone in the team that wants to do Python, instead on Fabric from getting a data flow gen one to something on Spark by Spark, ending the M code there, I'm sure ChatGPT will have good enough logic that is equivalent to Power Query. So, these things are good.
Rob Collie (00:52:04): The way I look at it is there's nothing truly special about DAX or M that's going to make it somehow immune to being solved to the same degree as something like Python has been mostly solved. And we've had a guest on this podcast, Brian Julius, who is very much in the camp of, yeah, not only will it be super useful with DAX and M, but it already is. You just need to know how to finesse it.
Gil Raviv (00:52:29): Yeah, the same time kind of of effort you would've to let's say read my book on back Power Query to be very strong with the UI and a bit of the M code, you can use instead ChatGPT and it'll generate the code. But you need to learn how to do all the prompt engineering to make it better. And it's the same time in the end. I'm not saying ignore it. It's definitely important tool to have, but it cannot replace a good knowledge of the technology and the UI, and how it translate into the M code. So, you don't need to write it.
Rob Collie (00:53:05): Yeah, you still absolutely need to know what you're doing. And I think that's going to remain true for a long time. The way we're looking at it is that it will accelerate, even someone like you or me, it will accelerate what we can do. So, let's get back to BI Pixie because I've long looked at this. The usage, and monitoring and all of that. Microsoft calls it solved because they have usage reports. You go and you look at it and you're like, this is not solved.
Gil Raviv (00:53:29): You have a very shallow perspective of how your users are interacting with your reports with out-of-the-box functionality.
Rob Collie (00:53:37): So, what are the top three examples of things that you get with BI Pixie that you've built into that product? I'm a product manager at Heart as well. So, I looked at their usage reporting and went, oh, there's a huge opportunity here. It isn't necessarily a great fit for our business, so we didn't go and pursue that. But I completely understand and salute the effort. So, I'm really excited that this is something that you saw as well. If someone says to you, "Why do I need this thing that Gil has built? Because Microsoft's already gave me usage reporting." What's the canned answer to that?
Gil Raviv (00:54:12): The main vision here is that any professional BI developer, if you build dashboards on Power BI, Power BI reports for organizations. And business decisions are made out of your dashboards, you need to think about that element as a digital product. So, we bring the product management experience that we have and we want to be able to successfully measure the impact, the effectiveness of our dashboard because it is a digital product. So, if analytics and dashboards are needed to save money and provide significant value to organizations. Without the analytics tools and solutions, they cannot make informed decision. They can lose money, they can skip opportunities. And you want to measure all the business operations. Why don't you want also to measure the success of your BI solutions? Because those are attending a lot of business and value on the other things that they measure. So, you need also to be able to measure themselves. That is the motivation.
(00:55:17): And when you look about any BI developer, wanting to be professional and more intentional about measuring their success and impact, you end up with out-of-the-box capabilities that are very specific to basically page views. In the digital product, you cannot just tell things about your users just by them opening a page. There is a lot of interactions inside. What they do, where they click, how long do they stay in those reports? What they select, which bookmarks they click on, which values? What are the top values that they select in their bar charts and maps? All this story is something that you would want to be able to measure if it would be a successful web solution, web app or a mobile app. So, now as a BI developer, if you want to be able to measure those things as well, BI Pixie today is the only solution that allows you to do it.
Rob Collie (00:56:13): It's funny that all of that information, basically every click is captured in the grand system and it's stored. You can get to it, but you have no chance. Be like exploring a database with your eyeballs instead of having dashboards over it.
Gil Raviv (00:56:29): It's even worse than that by the way.
Rob Collie (00:56:30): Is it?
Gil Raviv (00:56:31): It's not capturing everything. That's the problem. So, there are two type of problematic or challenging gaps. The first one, as you stated, like Microsoft store a lot of data, but it's very difficult and complicated to find the important stuff out of the Raw Data there. That's one use case. The other use case is that actually lots of things are not even being measured at all. Now, theoretically, perhaps Microsoft can have all this data. It may be stored internally in some systems that Microsoft have, but they're not exposing them to the end users because of priorities, and compliance, and privacy and customer data that they would want to protect.
Rob Collie (00:57:13): So, what's an example of something that Microsoft either isn't storing or isn't making available, that's something that you think is important?
Gil Raviv (00:57:21): So, for example, if you are now on a Power BI report, and it's not like embedded report where you may have a little bit more options. So, it's a Power BI report. It is consumed in different ways, but you don't have a JavaScript implementation on top of it to control some stuff. So, you cannot know if people are clicking on bookmarks, which visuals they click on, how long do they stay in those reports and pages? So, specifically, the duration where a user stays on the page, on the Power BI page, is something that you don't have any data on. So, you do have, if you use for example, log analytics from Azure, you can get every calculation on the VertiPaq engine that comes from clicking. But you cannot differentiate between where you clicked versus what was just all the other visuals that have the interaction, and now you recalculated the data for them. You cannot now differentiate between what was clicked and what the other visuals were just affected. So, the story of users clicking on specific things is completely lost.
Rob Collie (00:58:32): Do you have a solution for that in BI Pixie? Do you capture more somehow?
Gil Raviv (00:58:36): Yeah, so the approach that I took was in addition to collecting the data that Microsoft provide out of the box to include also the ability to instrument reports. So, basically when you use BI Pixie, you will run an automation that will scan all the reports and add to those reports, little invisible tables. And those tables are acting like a web tracking mechanism. They use DAX measures to go to a web service, and send contextual information about the report, and the page and the things in that page to a web service to collect this data. Now, that web service is deployed on the customer side. So, BI Pixie is not a SaaS that collects lots of things about the users.
Rob Collie (00:59:23): Perfect.
Gil Raviv (00:59:23): The customer will have that it's private and they don't lose this kind of compliance and requirements.
Rob Collie (00:59:31): So, the product manager nerd and me is fascinated by this. So, let me make sure I understand. BI Pixie, you have automation that goes and basically adds... it's a hidden visual?
Gil Raviv (00:59:42): Yes, a hidden table visual.
Rob Collie (00:59:44): The built-in table visual, or is it a custom>
Gil Raviv (00:59:47): Built-in. Everything built-in. I didn't want to make custom visuals. A simple little table visual that is a one-on-one pixel size single cell.
Rob Collie (00:59:56): Okay.
Gil Raviv (00:59:57): Invisible, also that you don't see it.
Rob Collie (00:59:59): Genius.
Gil Raviv (01:00:00): You can open the selection view and see all those elements grouped together if you have the edit options. But they're all just grouped together, invisible, and provide all that context for you.
Rob Collie (01:00:13): Okay. And that visual, because it participates in all the cross-filtering. Like if I click on a bar on a bar chart, that's sending that click essentially to the table visual so that the table visual can also respond and filter accordingly. Is that how you capture what was clicked?
Gil Raviv (01:00:30): So, basically, let's say you have a bar chart. So, now I'm creating a table. This, I call it the pixie, a little magical little thing that is invisible. It is there, and it'll only operate when you click on that bar chart. So, the instrumentation also define the interactions. So, there is a single interaction between the bar chart and that table. So, only when you click on that bar chart, that table will emit some data. And the data that will be emitted from that table will also include which columns you selected in the bar chart and what the values that were selected on the dimensions. So, now you can know not just that the bar chart was clicked, but what the user clicked on.
Rob Collie (01:01:11): Oh, my God. How excited were you when you figured out that that could work?
Gil Raviv (01:01:15): So, at the time, I figured it out in a manual effort. Four years ago. Actually five years ago. I blogged the technique, the vague technique on one use case, but it wasn't automated, so there was no value to doing. It's a crazy thing to do it manually because you have so many things that you can do. And now with the ChatGPT bringing me now to do automation and write code, I could just scale up the business logic of how I can instrument things in what kind of logic, including bookmarks, including slicer in the visual.
Rob Collie (01:01:51): This pixel-sized invisible visual also captures bookmark clicks. Is it also the thing that helps you measure lingering time, like the time users spend on a page? That seems like you'd almost need something custom for that.
Gil Raviv (01:02:04): So, the duration time is achieved later on when... So, you have lots of logs for those activities with the contextual information of them. And now I can sessionize those elements together into a duration. So, what is a session duration basically when you talk about it for a business user? You will define a timeout. So, if I have a lot of logs from the same user in the same report and page, if for let's say two minutes or five minutes, I don't have anything else, I can say, okay, I can wrap up now a session and I can now calculate the statistics of that data. So, all this business logic is done outside of the automation. And I used of course, power Query magic to build those elements together. And eventually I can scale it into more like now PySpark and all the other fabric things as well when the data becomes bigger.
Rob Collie (01:03:00): So, even in something informal in my personal life. So, I built a set of Power BI dashboards for my hockey league back in Indiana, and I've handed off ownership of these. They're being refreshed today. I got a bug report yesterday that I didn't have a chance to engage with. I'm going to help him debug that maybe after we're done recording here today. And of course, I use Publish to web, to share this with the world, so it's like there's no way that I can even get the normal usage statistics.
Gil Raviv (01:03:33): You can with BI Pixie. Publish to web use case is also supported.
Rob Collie (01:03:34): All right. So, I badly, badly, badly want to know, as the creator of these dashboards, I badly want to know what the interactions look like. What are people actually doing with these things? And I don't know. I have no idea. The only insight I have into how these things are being used is that the link that I've published for everyone to click on to go to is a bit.ly link. So, I get bit.ly usage statistics on clicks. Even there, if they happen to just take the URL out of their browser after the bit.ly link resolves to the Publish to web URL, if they take that and favor it, I don't even get that. I don't even get-
Gil Raviv (01:04:13): So, I think tomorrow you'll be able to have that. I can help you.
Rob Collie (01:04:17): Okay, look, I know this is your product, but I also know this is your time. So, if you're open to collaborating on that, and so I could start getting usage statistics about the hockey thing, oh, my God, that would lead to all kinds of conversations. We could do a follow-up podcast.
Gil Raviv (01:04:32): And you can show how you see all the usage and-
Rob Collie (01:04:36): That's right.
Gil Raviv (01:04:36): You can even see who is the most popular player based on the slicing experience. You actually have also customer satisfaction. There is a thumbs up, thumbs down feature that I can add to your report, so people can give you a thumbs up, down [inaudible 01:04:56].
Rob Collie (01:04:56): Oh, so cool. So cool. Yeah. We've probably had three or four episodes of this podcast that revolved around the things we've learned or that I've learned in the process of building these dashboards. Because these dashboards are built for the definition of the end consumer. This isn't even a business case, but yet people are keenly interested. These get 50 clicks a week even through bit.ly. They are trafficked. And the homepage that renders is just bookmarks. It's just a menu of reports. So, there's by definition, no value delivered on just first glance. People are going to have to interact. But I don't even know.
(01:05:34): Someone commented the other day on Facebook said, "One of the only reasons I would ever come back to the league..." this is someone who hasn't played in the league for years. "One of the reasons I want to come back to the league would just be to keep myself from falling off the bottom of the Hall of Fame report in terms of games played. I just want to cement my legacy." I'm like, this is someone who's not even in the league. They're looking at a report that I wouldn't expect anyone. I wouldn't even know anyone was looking at it.
(01:05:58): Anyone that builds a dashboard or a set of dashboards, Power BI reports. If you're not keenly interested in this sorts of information, you're not really thinking about it properly. I completely respect what you said. These are digital products, these dashboards. They are software. And the ability to see how they're actually being used changes everything. Your expectation as the person who built it is going to be wildly different from the actual usage over, and over, and over again.
Gil Raviv (01:06:27): Without it. You can only learn about things based on just feedback, and shouts, and complaints from users, but you don't know a quantitative, empirical way to measure what's good, what's not.
Rob Collie (01:06:41): People have a disincentive to provide you feedback, too. To provide you feedback, they have to admit that something's not working for them.
Gil Raviv (01:06:49): And they're busy. You may not get engagement from them to get the feedback.
Rob Collie (01:06:54): I mean, I've been asked questions about our dashboards like, "Hey, can you add this to the dashboards?" And I say, "Well, it's already in there." Especially in a business environment, there's also a disincentive to admit that you don't get it, that you don't understand. And so, people will sit on questions or failures, the things you don't see. The data doesn't lie. Data will show you that no one's ever making it to that drill through that you know is the most important thing there.
Gil Raviv (01:07:22): And if someone just opened the report and then hate it, you will lose the opportunity to tell them different things to make an opportunity to change.
Rob Collie (01:07:33): Gil, this thing, I mean, I haven't even seen it yet, just from your description of it, I know now what you've built, and it is genius and it's inspired. It's not just from an administrator standpoint. You can think of a BI admin standpoint depending upon how big your estate is, how many deployed dashboards, how many deployed reports you have, maybe. Maybe I'm interested, maybe I'm not. And that's from the usage statistics perspective. This isn't just usage statistics plus. This is instrumentation. And every software company in the world that's cloud software, now 100% relies on detailed instrumentation of their software in order to make their products better. Software has gotten much better in general because of detailed instrumentation that informs the product teams about what's being used, what's not being used, et cetera. And when we deploy a dashboard, we unfortunately don't get to just sit there looking over every user's shoulder seeing what they do. And BI Pixie basically does that. It gives you incredibly comprehensive. I can't believe the things that you've solved. It's so cool.
Gil Raviv (01:08:40): Yeah, so imagine now you can have a heat map that will show you all the visuals in all your reports, and where they are clicked on to improve different things. And it's endless because when you have this data, the combination of things that you can now provide. For example, if you as a customer already agreed to use BI Pixie and allow me to instrument your reports, now I can gather statistics about reports in terms of the visuals, how many slicers bookmarks, how many table visuals you have, how many interactions. And include that in the calculation of, for example, perhaps you have low usage in your reports because those reports are super complex.
(01:09:20): So, you allow BI Pixie to measure what is the KPI for complexity and correlate that with engagement. And now you can know, "Oh, here are my reports that are extremely complex. I know what I can do to make them simplified and increase the adoption." Or, "Here are reports with no slicers and no bookmarks. I know now that the bookmark here and here will get me more traction on the third page of the report that has an important use case that I want to allow my users to discover."
Rob Collie (01:09:50): That's just awesome. Okay. Well, I am really looking forward to digging into this, messing around with it with regard to the hockey dashboard. I think there's more to this than just that. So, to be continued. And Gil, just how great it is to have you back. We missed you.
Gil Raviv (01:10:05): I missed you as well. And again, I wanted to say, and I will say it as many times as I can, you were so pivotal in my new career analytics and consulting side. And everything you made in your own journey was something that I used as a benchmark for, oh, this is super cool to do. I really want to see if I can follow some of those big footsteps. So, thank you for the inspiration and the opportunities. Even my knowledge, my technical knowledge today, started with you providing me a test to join your team.
Rob Collie (01:10:44): That's right.
Gil Raviv (01:10:44): And that was super complex, and I needed to learn so much things and try. So, it's like before the test, I knew very little after the test, I read your book and I saw, oh, I can be much more determined on solving things that I don't know enough. The idea that there is no, no, I can't. It's always like, yes, I can do. It's like what I watch a bit about with your grandfather, and also my condolences. But this kind of attitude of solving things, I got it basically from being a part-time consultant in P3 for months. Solving things that previously I couldn't know how to solve, but now I got an opportunity to just die hard, sketch my head and walk out on solving. That was really instrumental on me getting into the big consulting with the attitude of there is no challenge that I cannot learn how to solve. Not because I'm smart about it, just because I know that with the determination and access to knowledge like you shared with the community, I can eventually solve it if I will be stubborn enough.
Rob Collie (01:11:52): It is super, super gratifying to know that I help someone like you do what you've done. Because I look at the things you do today. I'm like, I can't do those things. So, I'll allow myself to feel some small percentage of that. To be continued. Thanks for being here today.
Gil Raviv (01:12:07): Thank you so much for having me.
Speaker 3 (01:12:08): 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.
Sign up to receive email updates
Enter your name and email address below and I'll send you periodic updates about the podcast.
Subscribe on your favorite platform.