Austin Levine - Co-Founder at Power BI Consultancy

Welcome to another Founder Wisdom podcast. We have Austin Levine with us today. He's a CEO and founder at Power BI. Sorry, Case One it's called, and it's a Power BI consultancy. He's an ex-Uber. He's in Berlin, Germany. He's an ex-Adspace as well, which is interesting. Started at Deloitte as well. So interesting background. Today we're going to talk about dashboards, analytics, and how AI intertwines into this. So Austin, welcome to the pod. Tell us a bit more about what you're up to nowadays.

Yeah, awesome. Thanks, Charles. Thanks for having me. Yeah, I can start a little bit with my background and sort of how I got into the position that I am today. So over the last 13 or 14 years, I've been in every sort of data role you can imagine. So started much more on the on the business side as a growth analyst, data analyst. I was working as a consultant at Deloitte as my first job. Then I got into tech at Uber. So I was in San Francisco for 2 and a half years working on measuring the effectiveness of product changes in the Uber Rider app. Got into a little bit of data science there as well and data engineering. After that, I moved back down to LA where I'm originally from and joined Headspace. And I was a data scientist and a data engineer there. Meditation is also a big part of my life as well. So that was, that was a cool company to work for, working with a product that I used almost every day. I mean, same was true for Uber, actually.

Yeah, like it shapes your brain into thinking like a consultant and how to make money, and you're surrounded by high performers, right?

Yeah, definitely. I mean, to be honest, towards the end of college— so I went to college at Cal Poly in San Luis Obispo, California. I was an econ major. I wasn't totally sure what I wanted to do with my life at that point. A friend told me that Deloitte was coming down to speak, um, do some recruiting on campus. I decided to show up and that's kind of how I've— I fell into a data analytics associate role. I was always pretty numbers-driven but didn't know anything about SQL, didn't know my way around a database. So it was definitely— definitely jumped in with both feet learning like a technical field that I didn't know before. I was surrounded by obviously a bunch of really smart people. It's a very cool job to have at 22. I remember with one client. I was flying to Chicago every week from San Francisco, and the client was a big bank. There was, there was no budget whatsoever. You're staying at like the fanciest hotels and crazy meals and stuff. Um, cool experience at 22 and 23. But, um, yeah, where was I? So Deloitte, Uber, Headspace. After that, I really had an itch to to live and move abroad. And that's what brought me to my next role at another startup that was based in Cape Town, South Africa. So that was a long way from home. This is a startup that got acquired by GoDaddy. I was a, I was a data engineer there for about a year. And since then, I've been in Germany now for about 8 years. I met my current wife in Cape Town, moved back to Germany with her. Now we've been in Berlin, or we were in Berlin for about 7 years. Now we're in her hometown, a bit settled down, have a kid now and all that stuff. But professionally, what's happened in the last 7 years or so was working at a lot of different earlier stage startups as a one-man data team. So owning everything from helping define KPIs, building the actual infrastructure, data pipelines, how do we get our data from our source systems into centralized data warehouse. That was kind of a new idea at the time back then. And then of course, building the dashboards itself, this, you know, understanding the storytelling aspect of a dashboard, figuring out, figuring out who the consumers are of the dashboard, what they need to see, what they don't need to see, how to make that data legible and actionable, most importantly. And I did that at a few different startups while I was in Berlin. And, you know, the fun part of that was the first 3 or 4 months where you have a greenfield project and you really get to build, set up new systems, help really, you know, lay the groundwork for what the data landscape will become. And after that, it was a little bit of like maintenance work. It was not as Um, not as intriguing. It wasn't as exciting. And then I thought, hey, I could probably go into freelance. So long story short, I met a freelancer in Berlin. His name's Saju. I asked him, hey, um, you're doing freelance, how do I get into it? And I think a couple beers later at the beer garden we were at, uh, he asked me, hey, I've been wanting to start my own consultancy and I've been looking for a co-founder. What do you think? So that's how Case When was born about 4, 5 years ago, maybe 5 years ago at this point.

2021.

Yeah. And so what's been— what's happened between then and now? For the first few years of Case When, we were kind of generally a data and BI consultancy. And that sounded really cool at the time because it's like, wow, every company is a potential client, right? Because we didn't have any specific industry, any specific data stack, like we could work with everything. Um, it made it a bit hard to describe what we do in one sentence. Like even my, my parents didn't really know what I did. My friends back home didn't really know what I did. Uh, yeah, data, you know, building dashboards and blah, blah, blah. But what we saw, especially in Germany, was a specific need around the Microsoft stack and Power BI. So about 2 years ago, or a year and a half ago, we made the decision to niche down into Power BI specifically and really focus on that BI tool and the Microsoft stack around it. And niching down, I mean, you hear that advice from all sorts of founders. Like, I've consumed a lot of, uh, Alex Hormozi content in my time, and niching down was one of the, one of the best decisions that we made because now we can describe what we do in one sentence. People hear Power BI and they think of us, and the, the referrals and the work has kind of just kept flowing in, or it's been flowing in a bit more than, than it did before. Um, so that's been really great, and that's where we're at today. So, uh, we're a Power BI consultancy. It was just the 2 co-founders for a long time. We've doubled the team in the last year. I think part of that is credit to, to us niching down. So now we have 2 Power BI consultants on the team and much more focus about the type of work that we're working on. We could work way faster. we're clear on sort of where we're going as a company. So yeah, a little bit about—

And how does one consume data nowadays? I mean, brains have not changed so much, but what's changed is the ease and access to data and the ease of agents to filter that data and make sense out of that data so that our human brain can digest it, right? AKA AI's How should— how did that change how you built your dashboard since, I don't know, GPT-3 arrived?

Yeah, no, it's a good question. Building dashboards has definitely become more of a commodity. Like, AI can build not only Power BI, but every BI tool. You know, a lot of these tools are based on YAML files now. So as soon as you point that to GPT or ChatGPT or Claude. Lots of dashboards can be spun up for much cheaper than they used to. And so I think the, the important work is, is no longer around necessarily creating a nice-looking dashboard that tells a good story. Well, telling a good story is maybe not so easy for AI to do right now without some good prompting. But what's become much more important is the layer underneath, like the quality of the data in your data warehouse, in your data lake. Because yeah, you can hook up your credentials to Claude and ask a question and you're getting, you know, answers 10x faster than you used to. But are those answers actually correct? Can you trust it? You always had reports in your company that some— that leadership didn't trust. I mean, that hasn't changed with AI, right? They're just, they're getting maybe incorrect, confidently incorrect answers faster. So I think less of our work already, we see this already, less of our work is spent on building dashboards and it's more on the layer underneath. We're more selling the trust In the numbers, getting people to align on KPIs, making sure that the data is actually in a state where AI is going to give you correct answers when you point it to your data.

What does your dashboard look like nowadays? I guess it's different person per person, right? Depending on what role you occupy. But let's say most of the audience here is founders, CEOs. What should they look for on a daily basis? Yeah, what should be the format, right? I have a dashboard, there's numbers in there, it's projected on this TV screen, it's made by Claude Code. There's number of leads generated, there's close rate, there's sales cycle. What else is there? Monthly recurring revenue and a bunch of intriguing data. And well, I'm seeing a bunch of my agents' work, which is pretty fun also that's in there. There's a bunch of other things. What would you recommend? What do you see? On your end and what'd you build for your clients?

Yeah, I think the, the metrics that you're looking at, that example you used, I don't think those metrics are necessarily going to change, right? Those are the important metrics that measure your business, measure how well things are going. I think the interesting thing that will happen more and more is combining those metrics with all of your business context. So obviously your dashboard has the raw figures about what your performance is and how things are going, but it's maybe missing a lot of context from all of your meetings, all of your emails, what actually went on at that point in time. You're looking at a, you're looking at a chart and you see a spike or a dip on a certain day. What actually happened? Having an AI-based dashboard that not only can show you, you know, very legibly the raw figures, but can tie that back to business context and actually give you actionable insights, right? What do you, what do you do with those numbers? And so I think this is, this is the direction that we're heading, not just creating these dashboards more quickly and easily. And yeah, Cloud Code can create these pipelines, create a database, Create a report, but feeding that with your business context, I think is sort of the next logical evolution.

On your end, what do you see? What's important to you? Is it number of clients? Is it NPS score satisfaction? Right? Like what do you see on your end? You look at your dashboard? Yeah.

I mean, obviously we're tracking monthly recurring revenue. We're tracking close rates in terms of sales calls, stuff like that. Internally, I'm tracking a lot of things around speed to, for example, get a new hire up to speed at sort of 80% of the quality that I would expect from myself or my co-founder. Like a lot of speed and efficiency metrics. It's a little bit qualitative. That's sort of hard to track on a dashboard. Um, but as a consultancy, obviously utilization in our employees, um, to figure out our, our profitability and our margins as well. In terms of like marketing and sales, we don't have a huge presence right now. Like, you know, we're running some cold email campaigns, LinkedIn campaigns, but we're Trying to go all in— trying to— we're going all in on, uh, YouTube. So this will be our next big channel that we're going after. Don't have a presence there yet.

Why are you going all in on YouTube?

Because I think YouTube, a high-quality YouTube video, is like, as founders, we want to be Visible to our potential clients in order to sort of build familiarity and a level of trust. And being, being the face of the business in a video is not easy to just automate with AI. So I think there's an added level of trust there that we are real people and you can trust us and you can trust our business rather than going all in on, you know, LinkedIn or a blog, which we also have, and that stuff is much easier to automate. But man, I was scrolling through LinkedIn these days and it's just, you can see how many posts and comments are just, you know, obviously AI generated. And, you know, the whole dead internet theory, it's a little bit disheartening sometimes. And you think, oh man, so many people are doing this. So many people are flooding my feed with with content that it sounds good, like it sounds right, um, but it doesn't feel human. So I think that's, that's part of it. We want to, we want to give people the feeling that we are human, we are trustworthy, there are real humans behind the brand, we know what we're talking about. But to be fair, or to be honest, um, video production is, is something new to us, and I know that there's a little bit overwhelming thinking about how much there is to learn to make a high-quality video. Because there's also a risk that if you put out low-quality videos, that would hurt your brand more than if you just didn't put out any videos at all. And long-form video content, we can slice that up and create short-form video, cross-post that everywhere, take a long-form YouTube script, create an article out of it. So kind of Long-form video, long-form YouTube content feels like a great core piece of content that we can use to create, you know, auxiliary pieces of content from.

Yeah, yeah. Um, your clients, what precisely are you doing for them and how do you stay competitive in an age in which technically they can go on Floodcode and create their own dashboards and automations?

Yeah, yeah. Like I said, I think it's becoming cheaper and easier and more of a commodity to create dashboards. What we are really helping to— like, what we are really selling is the trust in the, um, the trust in the numbers underneath. Like, the trust that if you use, uh, if you use Claude or ChatGPT to point to your database, that you're going to get the right numbers. And for almost every use case that we've seen, every client that we've seen, they can't do that right now. You know, you have multiple definitions of revenue for different departments. So depending on who's asking the question, they're potentially getting a wrong answer and no one's going to trust that, right? It's, it's still the same issue that we've always had with building dashboards. At least with a dashboard that you've built, you can set some guardrails for what people can see, the type of answers that they can come back, you ensure that they don't see a wrong answer. I think right now, if you just point Claude code at it, at your data, um, right now these guardrails don't exist, or as strict of guardrails don't exist. So this is, this is really what we're selling. I mean, we are definitely also still building dashboards. Um, but what's changed is that internally we are, we're building dashboards with, uh, with the help of Claude Code for sure. Like, this makes everything so much faster. Like, we're delivering faster for clients. We, what we build costs less because we do it faster. So, uh, we're obviously passing on those savings to them. It allows us to learn faster, learn how to develop in this age of AI more quickly, create nice templates and Claude skills and all this stuff to make the next dashboard even faster. So your question is, what do we, what do we sell when people can just use Claude Code and create a dashboard? Yeah, it's the— you still have to have someone who really understands when the AI gives a wrong answer. So you have to have someone who understands the fundamentals of data modeling, knows all the ways that a question that you ask in, in free text can be— can possibly show a wrong answer, like understanding all these different nuances with how their data is. Um, so it definitely still needs an expert's eyes, but for sure, I mean, we want to get to the point where we're really just selling the governance and the trust because obviously A BI consultancy will not just be creating dashboards in the future. Like, it's already happening. So we know that.

And so far, I mean, YouTube, I'm also experimenting stuff to get lead gen there, right? It's hard. It takes a while. Shorts are good to attract traffic. And then there's a bunch of other strategy with like AEO and SEO, right? Like, so that your videos get seen. Um, how have you been getting your leads so far? What's your number one lead gen channel and precisely how much leads has that been generating on a monthly basis?

Mm-hmm. Yeah, to be honest, we are almost, I would say 90% is referrals. So, and that even increased once we niched down into Power BI. So we're still a small shop, right? We, we're trying out cold email on LinkedIn. We've gotten We've gotten several leads from both. I would say we have probably like 10 high-quality leads per month as a mixture between— and high-quality leads is really like they have intent and it's the right ICP, which is sort of a medium-sized business in manufacturing or real estate or logistics. CFO or controlling controller or head of data. So yeah, between cold email and LinkedIn and referrals, these are our main channels right now. The long-form YouTube is— yeah, the idea with that is not necessarily that it's the channel that will bring in the most leads in the shortest amount of time. For sure not. It's more about building the long-term perception of expertise.

Yeah, and you and your co-founder, how do you separate roles? Who does what and what does the day-to-day in your life look like?

Mm-hmm. So I think, My co-founder and I made the mistake for many years of really just trying to do all of the same things but just split the tasks, meaning everything around sales, marketing, client delivery, creating internal systems to speed up, you know, our internal processes or speed up the processes for how we deliver to clients. And so we have since split those responsibilities a bit more where he is more on the revenue generation side. So he's doing more of the, the marketing and the sales and the networking. He's also German, so that helps in Germany because he can, uh, you know, speak at these, at these meetups and conferences and, and network with people a bit more easily than I can. I mean, my German is I would say like conversational light. But, um, and then I'm focusing more on client delivery. So I manage our consultants, I manage the systems that we use. So my day is, I would say like 90% of my day takes place in Cloud Code, in VS Code, just different terminals for different repos that are working on a different topic, whether it's a client project or an internal project. And just like everyone, I'm trying to learn the best way to use AI every day. Things are changing so much, like which model do you use? What are prompting best practices? How do I make sure that all the context that, that occurred in the last 3 weeks is still in my context window or is retrievable for the next session that I start with Claude Code? So this is where I spend most of my time. I'm definitely much more the technical, like, data architect. So if we have a client, I'm designing the system, managing the employees who, who build more of the front end. Right now we have, we have 2 Power BI consultants who are mostly on the front end. I'm kind of the only back end guy on the team, so the only data engineer. So I'm doing the pipelines and creating databases and stuff like that and doing the data modeling. Yeah, this is sort of how we, how we split our time or our roles.

And do you feel being in Berlin's an advantage? How's the startup scene there?

Yeah, the startup scene in Berlin is actually really cool. It's, it's close-knit, like it's pretty small, but it's close-knit. Um, there are a lot of great— especially in, in the data space, there are a lot of great meetups. And met a lot of people through, you know, Slack groups and meetups and stuff like that. So it feels a bit more— yeah, it's much more of a community vibe, I would say. I mean, a lot of big startups are not only startups, but tech companies are based in Berlin these days as well. So it's only getting bigger. But obviously coming from San Francisco, it was— everything is different from being in San Francisco where, you know, you're walking down the street and everyone's talking about tech, um, which got a little bit— I got a little bit tired of that after a while. But if you live— if you love tech, you live and breathe tech, like San Francisco is the place to be. Yeah, but yeah, I'm not in Berlin anymore. I'm actually—

I don't think I would stay in SF full-time, uh, because it's I don't know, this, like, it's the vibe is a bit, uh, too stressful and hypey. Um, yeah, and I think you need like recovery time, and SF is probably not the best place to recover, right? Uh, if you're always in the center of noise.

Uh, totally. Yeah, from the ages of 22 to 25, it was— or 26, I think it was good. But now, you know, I'm I'm 35. I don't want to be in the middle of a, of a big loud city anymore. I definitely need to recover a bit more. But as a, as a young professional fresh out of school, um, yeah, it was a great time. But I totally hear you. And to be fair, it's been a long time since I've been in San Francisco, so the, the vibe there could be a lot different from how it was in like the mid-2010s.

Yeah, dude, fully understand. Where can people find out more about you and the company?

Yeah, so you can visit our website, it's casewhen.co. Case when, by the way, is— if anyone is familiar with SQL or SQL, which is like the standard language for querying databases, uh, a case statement is, um, kind of a core function of SQL. So Case When is our name.

That was Austin and Charles for another