Etienne Garbugli: My guest today is Patrick Vlaskovits. Patrick is the CEO and co-founder of Superpowered, a company that was acquired just last year by Splice. Patrick is also a three-time author, including two New York Times bestsellers, The Lean Entrepreneur and Hustle, and one of the first books that I’ve read and loved on customer development, The Entrepreneur’s Guide to Customer Development.
Patrick, welcome to the podcast.
Patrick Vlaskovits: Thanks for having me.
Etienne Garbugli: I’ve been following the progress of Superpowered for years. I think it’s a really interesting company, really interesting technology. Maybe we can start there.
So what was the origin there for the company and how did Superpowered come together?
Patrick Vlaskovits: The origin of the idea actually was my co-founder Gabor Szanto. Gabor is a Hungarian; he lives near Budapest. I’m American, but my parents are Hungarian and I speak the language. So every summer, not every summer, but usually every summer, we go back to visit relatives and have fun.
About seven years ago, Gabor and I were introduced by a VC in Hungary. I was sort of between projects, Gabor had been talking to some of these VCs about some of the technology he had been working on, and they’d said, “We know this guy is technically very capable. He’s a hacker; we needed a hustler.”
So Gabor and I met at a Starbucks in Budapest, and then he kind of described to me what he felt or what he sensed. And basically, it was that there was a big shift happening in the audio space where all sorts of tools and apps that had been available on laptops, so desktop-grade processors, were now moving to mobile. This is back in 2013 and the iPhone was about five years old at this time.
He said, “Look, most of the audio algorithm, most of the audio processing infrastructure was designed for the desktop environment.” Meaning that the assumption is that there’s an x86 processor, there’s always power, ergo it doesn’t need to be that resource-efficient. Whereas if you look at the mobile devices, they’re all on processors, much more resource-efficient, and, of course, power is always an issue when it comes to mobile devices.
Gabor had basically been experimenting and building with a new way to derive audio algorithms and very computationally-intensive audio algorithms for our processor. And he had figured out a lot. And I also agreed with him. I saw what he was talking about and we took his vision for creating sort of a new library for audio transformations processing, all the typical audio stuff, and bringing it to ARM.
We generalized a solution that he had come up with and then brought it to market. And then seven years later, we were required.
Etienne Garbugli: That’s awesome. What initially attracted you personally to the project? What convinced you that this had potential?
Patrick Vlaskovits: Good question. Two things: one is Gabor is a really special guy, really talented guy. And when we spoke, we spoke both Hungarian and English to each other. Then I asked them, “Where did you learn your English?” And he goes, “Well, I learned it by watching Friends and YouTube.” At the time, he couldn’t afford English lessons, but he knew that there was a lot of technical information locked up in English language white papers at Stanford and MIT.
So he’d download them, kind of figure out some of the math on them, but couldn’t read them. So he taught himself English so he could download technical papers to get his own technology better. I thought that was really compelling in terms of a guy who was a hustler and was trying to push forward and not give up. So I thought that was pretty cool.
I liked the fact that we were in a very technical space. I thought that was very interesting and kind of a moat.
In retrospect, well, we made many mistakes, but one mistake we made because audio space is a tiny space, our TAM was really small in a lot of ways. So one mistake I think we made is, early on, we should have expanded the library to image processing as well because you can use some of the same algorithms from audio processing to image processing. We should’ve done that earlier.
We actually never did that, but we should have done that and we should’ve done it earlier. So now I think it was a mistake because our small TAM really limited us in a lot of ways.
Etienne Garbugli: You’re saying that Friends actually taught someone to read technical papers? That’s awesome. How did you go about figuring out whom, specifically, the technology would be best for, initially, and what the core value was?
Patrick Vlaskovits: This is a really interesting story. Initially, it was better performance. And better performance, in this case, meaning lower latency and less power. What’s interesting about that story, our market positioning and that story, why we think that’s important, for the first three, four years, people saw it. They got it and they’d go, “Cool. I get it.” But they weren’t really thinking about this at the time.
So we were actually a little bit too ahead of the market, which is very painful because you meet with customers. They go, “Cool.” They nod, but they don’t really sign up. So that’s very difficult, but then around year three, four, five, people started coming to us and going, “Hey, do you guys have low power, resource-efficient technology that runs well on ARM?” And we’re like, “Well, yes we do.”
So the gap between where we were and where the market was actually started closing. And then we also figured out that a large part of the value, too, was that a lot of developers were building cross-platform apps; so deploying apps to Android, iOS, even sometimes windows, et cetera. And our value is really about if you’re going to build cross-platform apps and you want one audio code base from which to work, we’re a really good choice. There are other cross-platform audio engines, but we are cross-platform and we’re very performance, which is very rare. So that was a big one. On sales calls, I always tell people, “Look, if you’re just going to do an iOS app and you don’t need anything special, just use the native iOS APIs that are free. You’re better off that way. But if you’re doing iOS and Android, then there’s certainly value. If you’re doing anything else, then you should definitely use us.
So, being able to, again, not just be cross-platform, but be very high performing in all those, and especially Android. Android is quite frankly a mess when it comes to audio; that was a big value proposition for us.
Etienne Garbugli: But if you’re saying you were ahead of the market, and for the first three years, not everybody got it, how did you guys know that this was the right positioning and the right value to put forward?
Patrick Vlaskovits: I think in some senses, we were wrong. I’m not sure what we would have done differently because if you compare it to the current offerings, our positioning is that we were designed for ARM and we were much better. But it’s really hard to sell a performance.
One thing we learned, though, is positioning around Android. We knew Android was a mess as we started selling, but Android is even worse than we had even imagined. So that part of positioning worked well. We actually released a lot of content around Android and audio latency.
In fact, one thing that we did that worked really well is we named a problem. So there’s this audio latency problem in Android and a lot of developers felt this problem. So they were trying to make low-latency apps, interactive apps, and their bosses said, “Hey, this latency’s not low enough. What’s going on here?” And they couldn’t figure out why they couldn’t make it work on Android.
And then we investigated and figured out what was wrong with the Android subsystems. And we wrote this long post about why this is happening, why it’s important, how we would fix it, and we called the Android’s 10 Millisecond Problem.
Now, the phrase “Android’s 10 Millisecond Problem” is now used. Like I’ve heard complete strangers tell me about Android’s 10 Millisecond Problem.
So in terms of a little marketing trick, if you could find an unfelt or unfulfilled need in a space that your company can help solve, even naming the problem can be of tremendous value. They feel this nebulous thing, like, “Android has this thing.” And you say, “No, no, no. It’s called Android’s 10 Millisecond Problem. Here’s why it exists, ABCDE, and here’s how to solve it.”
I think more and more startups especially should take advantage of the ability to articulate and name a problem that customers have and build credibility and quite frankly, build revenue that way.
Etienne Garbugli: At the time, how conscious was that effort of trying to name it and trying to own the…?
Patrick Vlaskovits: It was conscious. It was definitely conscious because I’m a writer. There’s tons of writers that have done this for other things outside of standards. Like you know Malcolm Gladwell’s 10,000-hour rule, and there’s a bunch of these little phrases that get associated with writers. It was certainly conscious and it worked well. It doesn’t always work, but when it works, it works well.
Etienne Garbugli: So in a way you were selling the problem to some extent, or at least articulating it in a way that people could relate to, “Yeah, I know what that is. That’s the problem.” And then you put this solution in relation to the…?
Patrick Vlaskovits: The basis of all storytelling and humor is that there is some problem that needs to be examined. So you’ve got to start front-load it with the problems typically first, and then the solution is after. And that’s how people tell stories and jokes.
I think successful content marketing is also very similar, especially for us where we’re doing a lot of low volume searches, but very high intent. So if you were Googling Android audio latency or terms around it, you should definitely be looking at our stuff and you probably should get on the phone with us. Does it make sense?
Etienne Garbugli: Yeah, for sure. How did you go about figuring out how to bring Superpowered to market? You have this technology, it works, you have an idea of what the value is, how do you figure out the next step?
Patrick Vlaskovits: I had done a lot of digital marketing before and I was pretty good at it. I mean, next up natural SEO and content marketing, a very natural next step is where you have developers searching for solutions. So building not content that ranks, but actually, that solves problems — we did quite a bit of that — is very powerful. And people would tell us on sales calls, they’re like, “I’ve read every page on your blog.”
In fact, one time we outranked Android in Google for certain terms, I’m not kidding you. I don’t know if we still do, but there was a time when we were getting a good amount of traffic on Android-related audio terms. We were getting a lot of links and a lot of people writing about us and we were outranking Android itself. It was great.
Etienne Garbugli: So how did you know that content wouldn’t be the right way?
Patrick Vlaskovits: Because developers are always Googling for answers for stuff. They’re always looking for code snippets or how to think about something. Developers are always looking for answers to their problems, especially the more specific.
I remember seeing a startup I helped consult with. They looked at their search and they figured out that they were getting a lot of traffic on a very specific error code that only their software was throwing things. They were getting like 1,000 searches a month on that, and it really wasn’t leading the way.
So we told them, “You should build a page, optimized for this search that helps either the developers answer the question or lead them into a funnel or both.” So, the same thing for Superpowered. So, SEO and content market was very strong for us.
Etienne Garbugli: To figure out what questions people were asking and then creating landing pages/content pieces that help drive…?
Patrick Vlaskovits: I’ll give you one really good way and I don’t know why people don’t do this more often. In fact, I advised this to a company that was just acquired by a friend of mine; your support tickets have so much fucking information in them. And if you take them out, clean them up, you could build reams of good content at relatively low cost because you have a question or questions and then your support person has, typically, already answered the question.
So, generally speaking, you can just export it, depersonalize it, make a few tweaks, generalize it a bit, and you have this whole realm of high quality, very specific content that you can basically publish for very low cost as a knowledge base, or even as blog posts. Another thing I’ve talked about in marketing lectures is based on the questions asked, I can tell a lot about that person who’s asking the question, whether they have a budget or they don’t have the budget or things that they’re not even telling to us. So I think it’s always important to keep your finger on the pulse and support.
And early on, if you’re the CEO or if you’re the product, don’t just leave support. Don’t just go, “Okay. There’s a support function that’s siloed.” Make sure you’re in there understanding what people are asking for and why, and it can be a channel to not only derive new product features, et cetera but also to cross-sell into. There’s a ton of data there.
Etienne Garbugli: Yeah. It’s crazy how many people use support as a cost center when there’s so much stuff you could do with support content itself.
Patrick Vlaskovits: Oh, 100%. It just takes a little time to build that, but 100%. It is a cost center depending on how you’re set-up, but you definitely should be doubling down and figuring out what’s going on.
Again, there’s one company that was just acquired that I advised. Their support was a complete silo. It had nothing to do with product engineering and nothing to do with marketing. And there was just so much good data in there that they could have exploited and used, but they hadn’t made up their mind to do that.
Etienne Garbugli: So, once you guys figure out, this is the model, so we’re using content to drive people in and we’re trying to get them to buy two-user SDKs so that they use either cross-platform or for one specific. How much focus did you guys put on that specific model itself? How long did you guys remain focused on that portion itself before trying to expand as you were talking about?
Patrick Vlaskovits: I’ll kind of indirectly answer this question. One challenge with us was that the de facto competition, there was a lot of open-source stuff that’s “free”. It’s not actually free because it takes longer to use. It’s not nearly as good. Yeah, it’s free. You don’t pay for it, but most open source stuff out there is terrible. And I know that’s not very popular to say, and I’m not saying there isn’t good open-source software. There certainly is.
Most of the open-source audio stuff out there is terrible. It’s not supported, but you still have this sort of free thing. So, people go, “Well, so-and-so is free. Why should we pay you?” So you have to walk them through into, “Well, we are a going concern. We’re living, breathing, we continue to upgrade. We’re doing this, this, and this on the roadmap and we have support. And if you actually did the math, yeah, you wouldn’t lay out as much cash for the open-source stuff, but your time spent trying to figure it out and trying to support it would be a lot higher.”
So understanding how to sell there, and then also figuring out pricing where a lot of people were doing these perpetual licenses. And we did a few of those in the beginning where you pay us a few grand and you use it forever. And then it didn’t work. It just wasn’t working and we tried a bunch of different things. Like, okay well, can we scale with users? And the people said, “Hey, you want to do rev shares with us.” And none of them really worked, to be honest with you.
The rev share doesn’t work because anybody who offers you rev share, typically, isn’t making money at that point. And then there’s this whole audit requirement and people don’t want to do that. It can be really tough. That didn’t really work.
The user stuff didn’t work because people felt like they were being punished for success. So, we just had to make a decision to say, “Look, we’re going to do this tiered pricing where it’s a fixed price per year. It’s annually, recurrent SaaS. It lets us continue to invest into the software and any of our clients get the benefit of it, and let us grow the company.”
And that actually was sort of a minor innovation in and of itself because the audio processing software was not being sold like that at the time. There was a lot of open-source stuff. There’s a lot of perpetual licensed stuff.
Then also, we positioned ourselves as a premium product because that’s what we are. The biggest thing also, I think, is we learned how to identify people. So when I sell, I always look for a fit between our solution, their problem, and their company and their budget. And one thing that I tell anyone who’s doing early-stage company, the sooner you figure out who you cannot serve and then learn how to identify them in your funnel, and then help them find another solution, the better you are.
I call those people the anti segments. The anti-segments or anti-personas are fucking deadly. They waste your time. They’re never happy. You literally cannot make them happy. It’s just a fundamental misfit. And even if they want your stuff, for example, for us one of our weaknesses was we didn’t have really good documentation. We knew we didn’t have great documentation.
We knew you had to be at a certain level of a developer to get a lot of value from us. For us, it was a resource constraint question. It was, well, if we have resources, do we develop more cool software or do we develop documentation? And 99% of the time, we said we’re going to do more software.
This is, admittedly, a weakness, and this is something I would even say on the calls. I’m like, “Look, our documentation is not great. If you have a development team that has heavy documentation requirements, we’re probably not a good fit for you.”
It’s not that we didn’t have any, but we didn’t have any long-form narrative type. So, early on, we made the mistake of selling to people who needed a lot of handholding, and then you gave them the hand-holding, and they still messed up. They just weren’t happy.
So now, part of the sales process and this has been going on for awhile, and I think any good sales person will tell you this is. You identify them. You figure out if there’s a fit, if there’s not a fit, you gently push them out and say, “Hey, I don’t think we’re a good fit for you. Here’s why. I think you should go here, here, and here.”
Gently and diplomatically push them away because if they get in there, they can make life hell for your team. They give you bad information. Not because they want to give you bad information, but just because they’re trying to fulfill their own personal needs, like anyone else. So, understanding who your anti segments or anti-personas are, who your segments are, who you can really help, the sooner you can do that — and it takes a while. You have to iterate a while — but the sooner you can do that, the more powerful your growth engine can become.
Etienne Garbugli: So, to that end, how do you differentiate what should be an anti-segment versus a growth or an expansion opportunity? The same thing you could have said, like we were going to improve and make it better.
Patrick Vlaskovits: One example is, like I said, you needed to be a decent developer to use our technology. And I’m not just saying this as a bragging point. We knew that we were living in ourselves because we did not have documentation. We understood this, but at the same time, you had to be a competent developer, and it just was really tough for beginners.
So identifying people early on, either the way they’re asking questions or how they’re asking them, and then literally saying, “Hey guys, we don’t think we’re a good solution for you. We think that this competitor over here is a better solution for you.” So that’s one. Budget was another one where we had a lot of people; a lot of really huge companies that you know and you’ve heard of, massive companies come to us.
We closed a number of deals that were very lucrative. That was great, too. But we also had big companies that were like, “Hey, we’re a huge company. We’re a monster in this space. Because we’re a huge company, we want your stuff free. Maybe we’ll let you talk about it.”
Initially, we were very flattered and it’s really flattering when you talk to the managing director at one of these big companies. They have a huge, massive budget and you’re like, “Wow, maybe I can get a deal done,” and learning how to say no to those guys, too.
And just saying, “If you don’t have a budget for us, that’s fine. But you’re not getting our stuff for free,” because it doesn’t make sense.
Sometimes, getting logos does make sense. But for our case, a lot of times they wanted to pull us into these projects, use our technology at scale, and have it supported, but not write us a check. And there are a few companies, I’m not going to name them, but there are a few companies that are famous for doing this — bullying young companies into helping them because they’re just massive.
Again, in the moment you’re like, “Wow, man. If those companies did use us, we would be in a hundred million devices overnight, and we can tell people that. It’s a famous device that everybody has in their homes.” But we just had to learn how to say no to those people.
And that’s another thing, too, I’d advise as well with any early-stage startup; figuring out how to say no, and not chase every opportunity. Don’t get me wrong. I chased a lot of opportunities and I don’t regret even the ones I didn’t get, but sometimes, these opportunities will sink you. They’re not really the opportunities.
Etienne Garbugli: So in a sense, you are always refining that idea of what the right customers are. Partly, what you were mentioning, support as well, that’s a good way to understand the patterns as well there.
Patrick Vlaskovits: 100%. If you understand those people, you can figure out if you’re fit and you’ll see patterns. For example, one of the things that we say now in sales calls, there’s these smart hobby guys, Joe in the garage. We’re a good fit technically for those guys, but we’re not necessarily a good fit when it comes to the budget. And I tell them straight up.
And a lot of times these are very artsy projects. Like one project was a set of flying drones that were to fly in unison and play music and do these interactive things. It was just an awesome robotics project some German team was putting together. They liked our software, but they’re like, “We have zero budget.” I’m like, “Well, I’m sorry.” It didn’t work out.
So if I sense there’s an artsy techno project, we’re not typically a good fit for those. But my takeaway, though, is to figure out who you’re a fit for. Make that fit, sell that fit, and provide that value. Figure out the people you aren’t and move those people aside, and help them find your competitor or another solution or whatever it is.
Etienne Garbugli: So just to get that focus?
Patrick Vlaskovits: 100%. You need to have the focus where there’s the three or four segments that you can identify, you can deliver value. They get it and they know they’re getting value. If there’s so much noise; you want to filter that noise out.
Etienne Garbugli: Okay. In that case, what were the triggers that led you guys to expand from audio to websites and apps to network libraries and to cryptographic libraries? When did you get the inflection point where you’re like, now is the right time, or we should do this?
Patrick Vlaskovits: Well, one thing we saw is that we had people taking our SDK and then integrating it with third-party networking tools like OpenSSL and things like that. And then they had some trouble and they said, “Look, we like your stuff, but OpenSSL is not really working.” Basically, it was a pretty simple adjacent feature where they need to encrypt some media and then just send it somewhere.
We actually built some cross-platform networking technology that basically did that. This was almost directly from user feedback where people kept saying, “Hey, help us do this thing. Otherwise, we have to pull in OpenSSL to do this and it’s a big pain in the ass.” And that was, looking back, a pretty apparent pattern pretty soon.
The stuff we do now is actually a little more out there. It’s really interesting, though. And I think it’s going to be tremendous. It’s a WebAssembly-based audio we’re doing. I don’t know how well you know it, but WebAssembly is a technology that’s being supported by the major browser vendors right now.
The idea is that you can compile code to run as if it were on a hardware device, meaning that you can have really interactive, low latency interactions; whether it’s a game, whether it’s a Photoshop tool, whether it’s some sort of musical device. Right now, the reason we use apps is basically there’s a better connection between an app and the hardware device than if you went through the browser. WebAssembly, in a sense, is changing that.
And if WebAssembly keeps being developed and being supported by the browser vendors, you’re going to have a lot of app developers move back to the web and there’s a number of reasons.
First and foremost is economic. I don’t know if you remember a few years ago, Fortnite launched an Android app, but not on Google Play. And someone asked the CEO of Fortnite, “Why did you do that?” He was like, “Well, there’s 50 million reasons why I did that.” They call it economic efficiency, but basically you make more money.
So think about it. If you had a franchise like Photoshop or what have you, and you had to send 30% to Apple every year, as soon as you could figure out a way to not do that, you would do that, right?
Etienne Garbugli: There was a case this year with the Basecamp. They were complaining about that 30% they got up to management at Apple and all that stuff.
Patrick Vlaskovits: It happened for Basecamp because those guys were smart and loud, but that’s not going to happen for the average developer.
Etienne Garbugli: For sure, but there is a problem.
Patrick Vlaskovits: Yeah, exactly. The WebAssembly stuff was interesting so we actually built, things may have changed, but we were the first real-time, time stretching in the browser. Time stretching is actually a very computationally intensive audio algorithm. It takes a lot of CPU. And basically, you slow down or speed up the tempo of a song, and then you also can do pitch shifting where you maintain the pitch so it doesn’t sound weird. Mathematically, it’s actually very difficult and computationally intensive to do. And latency is critical. You have almost no latency. We brought our knowledge of the web to do this. And one of the reasons we were acquired was we showed this to the Splice guys and the Splice guys were like, “Oh shit. We thought this was like three years away.” Splice is building cross-platform products, basically. They want to be all available on every surface there is, so Native, Web, what have you, and then Superpowered technology allows them to do that.
I think Gabor told me, “Man, we should have launched it even earlier.” People are licensing it; people are using it. It’s still early days, but I think you’re going to see more and more people building progressive web apps or some stuff like that. I think you’re going to see more and more of that.
What’s going to be interesting is if Safari will continue to support WebAssembly. Right now it does, but if Safari endangers the app store, I think you’re going to see a situation like what Steve Jobs did to Flash, where they killed Flash on Safari. They said it was because of security, but I think it was because of the interactive features.
Etienne Garbugli: Yeah. Probably the same thing; they’ll find another reason to justify it. It’s super interesting. The reason why you guys went with that feature set or those specific feature sets for WebAssembly, is because that was one of the best ways to showcase the value that you guys provide? Or is there another set of criteria you guys were looking at for that?
Patrick Vlaskovits: That’s a great question. We wanted to show that the big promise of our technology is how efficient it is and how little power it uses. And anybody in the audio algorithm industry understands that time stretching is really hard. So unlike Reverb or Echo, those aren’t particularly interesting and they’re not particularly computationally intensive. In time stretching, you’re basically saying we can do everything related to audio on the web.
When I talk about audio, it should be clear to the audience, I’m not just talking about play audio, pause audio, what have you. We’ve done interactive audio, where if you’re a DJ, you’re listening to the music, and then you’re changing the music in real time. Or let’s say you’re playing a musical instrument. If there’s a delay between hearing the instrument or even like a Skype call or a Zoom call we’re on right now, if the delay were too high, then it would be hard for you and I to converse.
And that’s a common problem when you’re on a Skype call and you keep talking over each other because the latency is too high. So lowering that latency, at least on the client side, there’s not too much that we can do on the network side. Now there’s some latency on your computer. There’s probably some latency in your earbuds as well; probably not a lot. There’s some network latency and then there’s my computer and your computer latency.
And probably the greatest latency right now is in the network. There’s nothing that Superpowered can do about the network per se, but we can certainly drive down latency on the client side.
Etienne Garbugli: That’s super interesting because, initially, that’s what you guys saw as the core value. And even a couple of years ago, you’re still using that core to figure out what are the other things you can do.
Patrick Vlaskovits: Well, to be fair, there’s only a few characteristics of audio that people care about. There’s the subjective audio experience of anything. But it’s well understood that audio has to be interactive and low latency or interactive audio has to be low latency. And then it’s bit depth and things like that. There’s only really a handful of characteristics that are used to describe audio technology in terms of when you give it to a developer.
So bit depth, I guess power consumption matters to some people. Then the subjective, for example, Gabor tuned our time stretcher to have a musicality to it. That’s a subjective thing. Some people might say that’s great. Some people might say it doesn’t sound good.
People actually think our time stretcher sounds really good, but there’s a fair degree of subjectivity to that too.
The point I’m trying to make is in audio, you can’t really innovate too much on how it’s judged or the parameters in terms of people are always going to look at the same parameters and compare it to their solutions. We innovated on pricing; also, ease of use another one.
Etienne Garbugli: It’s super interesting. I used to work at LANDR. Think you guys were involved with them at some point. And we would get feedback on the mastering and you would get the same mastering, same quality; some people would think it’s amazing. Other people would think it’s not amazing. And you would get that gradient of quality perception to some extent, which, just like you’re mentioning, it’s subjective.
Patrick Vlaskovits: Yeah, a lot of it is and that can be tough. And then learning how to filter that out and proceed is important as well.
Etienne Garbugli: You mentioned in an interview from many years ago that a startup is an exercise in resource constraints. I thought that was really interesting even today. I think that’s still probably true. How do you recommend entrepreneurs factor in their own constraints?
Patrick Vlaskovits: I don’t know if I have a thought-out recommendation, but I think 100% it’s an exercise in resource constraints. You only have so many hours; you only have so much energy. You only have so many developers. You only have so much money. You only have stuff that you’re good at and you just got to figure out how to make progress with what you have.
A few years ago, there’s a lady professor, I think, in Virginia, she talked about effectuation and that was her big idea. And part of what was really interesting is entrepreneurs make do with what they have at hand. I think that’s a part of it too.
In my case, Gabor is like a 100X developer, and I’m not just saying this to exaggerate. He is more productive than like 10 dudes put together and he’s obsessed about audio. He also knows chip design and hardware and audio acoustics and API design. So, a lot of times, I would literally go, “Well, do we hire someone to write documentation? Do we have Gabor write documentation? Do I try to write some of it?”
Basically, what you come to is like, “Well, there’s only one Gabor and it’s just better if he produces awesome technology super fast. And then we’ll figure out the documentation stuff.” So we knew we weren’t perfect, but we had owned what we couldn’t really fix. And time and time again, we’d take a break like, “Man, we really need documentation.” And we couldn’t find anyone that we trusted at a reasonable rate to write it for us at the time.
I’m not saying that we don’t have any, but the point, coming back to research constraints, is it would just be silly for us to have tasked Gabor and our other developers to write a lot of documentation when, in the same amount of time, we can have improvements on existing features, or even build new features.
Etienne Garbugli: No, that’s a great point.
Patrick Vlaskovits: It’s a really tough part of being a CEO or a founder.
Etienne Garbugli: Yeah, making that first determination that this is a competitive edge that we have. We have this super great developer or whatever else and we’re able to do this is already something people don’t seem to think about oftentimes.
So having been through the eyes and loads of startups several times, talked about startup, wrote books about it, worked with a lot of entrepreneurs; what are some of the signs and signals that you feel typically indicate the need for a pivot or a change of strategy?
Sorry, big question.
Patrick Vlaskovits: It’s tough to think of easy ones. I think we would’ve done better had we pivoted slightly and gone into image processing. I think it’d have opened up a considerably larger market for us. I couldn’t convince Gabor. This is also a resource allocation question.
It was tough for me to convince Gabor that that would be a smart move. This is a tough question. I’m trying to see if there’s something generalizable about our experience that would make sense to the listeners of this.
Etienne Garbugli: You mentioned a few times during the interview how in hindsight or you’re mentioning it as having done that would have been better. Are your negative instances where you’re like, “We spent way too much time at that stage waiting for X or waiting for Y.”?
Patrick Vlaskovits: We developed a solution for OEMs to solve the Android latency problem. And we actually demoed it successfully. But we couldn’t convince any large OEMs to license it from us. And had I known how tough it would have been to sell into an OEM without a champion… I don’t know if that was the smartest move. I mean, it was a hot topic audio space of time.
We had very large companies, OEMs, come to us and ask for a demo. They wanted to see that we could prove our claim. We actually got down quite a way with a Chinese OEM actually until they pivoted radically one day. And then showed this solution to other OEMs, and we spent a lot of time on not only the solution, but actually trying to sell that solution and we couldn’t do it.
I don’t know if that was us tilting at windmills or we just couldn’t sell it or I couldn’t sell it, but that was one where, looking back, it was like, “Maybe we shouldn’t have wasted time with this.”
Etienne Garbugli: Yeah. That fits what you were mentioning before about finding these anti-segments or whatever.
Patrick Vlaskovits: Yeah, exactly.
Etienne Garbugli: Is there a way for founders to quickly figure out what those should be?
Patrick Vlaskovits: I think it comes down to the problem. I think the biggest thing when we were selling to these OEMs; it was basically OEMs that are producing inexpensive Android phones that, at the end of the day, they didn’t perceive it to be as a problem like we did. Although one of them, we got pretty far down the way with them. So I guess the mistake was not verifying that the pain was really there. That was probably a mistake.
We demoed technology that would have been relatively easy to implement and you could have updated billions of Android phones overnight and they would be almost real time audio-capable. Basically, making them as capable as an iOS device.
There are multiple challenges there. One is, do they perceive that to be a problem? Two, if they do, does the team think they can do it themselves? We ran into an OEM that was like, “We can do this ourselves on the weekend if we wanted to,” which is completely incorrect. And that’s common when you’re selling to developers. Anyone who’s ever sold a development tool knows this. It’s tough selling to developers. It’s not easy.
Etienne Garbugli: Yeah. I think there’s a little bit of, I don’t know if arrogance is the right word, but there was a little bit of that reaction of, “I could have done that.” That’s super interesting. So, maybe more generally as the last question, as your perspective on starting businesses evolved over the years, how do you see B2B entrepreneurship evolving, moving forward?
Patrick Vlaskovits: If anything, I appreciate now what people have told us before, and I didn’t appreciate as much: have a big TAM, which is basic advice. It’s like 101 where people go, “Hey, make sure your fucking market is big enough.” And you go, “Yeah, yeah, yeah, yeah, whatever.” And now looking back, I’m like, “We should have listened a little bit better.”
We had a good outcome; we did fine. We had a good count. We got acquired. It was good. Everything’s fine, but just having enough surface area and having enough shots on goal I think would have been better. I think we could have grown faster and been more successful.
Etienne Garbugli: Yeah, for sure. Is there a way to look at it a little bit differently where you could say maybe this was your beachhead market; your entry point was that market?
Patrick Vlaskovits: Yeah, for sure. What I liked about it, initially, it was very niche. What I liked about it is it was so niche; it was easy to identify customers. It was relatively easy to identify customers and then market to them in terms of building content.
Etienne Garbugli: There’s definitely benefits to that, right?
Patrick Vlaskovits: Oh, tremendous. However, audio is just a weird market, too. If I knew what I do now, I don’t know if I’d ever do audio. Audio is just a really interesting, weird, strange market that has all sorts of idiosyncrasies. Every industry has its idiosyncrasies, but audio is especially weird and strange.
A lot of people that really love it are really fanatical about it and they are obsessed with it, and they’re great people, but it’s very different from the typical B2B, SaaS tool doing accounting or marketing or what have you.
Etienne Garbugli: Initially, four or five years ago, when I saw what you guys were working on, I was like, “Whoa, this is cool.” I wouldn’t be a customer of Superpowered, but I was like, “This is really cool. It’s a very focused play.'” It’s not something that you have 16,000 companies that are looking into it, which gives you the ability to gain ground on the market.
Patrick Vlaskovits: True, but then the other side of that is you also have a lot of open source stuff that’s not that good that draws people away. We’ve had multiple people say, “We need to go do the open source.” And then, I’m not kidding you, they came back six months later and they’re like, “Okay. We wasted six months on open source stuff. We found some budget. Can we make this work?” This happened multiple times. And I’m sure it happens in the generalized B2B world as well.
Etienne Garbugli: For sure. Thanks for taking the time. That’s really appreciated. Where can people go to, to learn more about your work, your company, and everything?
Patrick Vlaskovits: So superpowered.com is still up. Like I said, we are a part of Splice. You can find me on Twitter at @Pv. Then I’m on LinkedIn, Patrick Vlaskovits. If someone wants to send me an email; vlaskovits [ @ ] gmail.com, just my last name as it’s spelled.
Etienne Garbugli: I’ll add those links to the post and thank you for taking the time.
Patrick Vlaskovits: Cool. Thank you.
More on Anti-Personas
- Company Focus = Growth: Here’s How to Focus Your Business to Win Your Market
- The Importance of Focus in B2B Startups
- How Loopio Grew from 0 to 300 Customers by Focusing on Relationships
Download the First 4 Chapters Free
Learn the major differences between B2B and B2C customer development, how to think about business ideas, and how to assess a venture’s risk in this 70-page sampler.
Enjoyed this Episode?
Subscribe to The Lean B2B Podcast for more: