Hey there, this is Eric Wright, the host of the disco posse podcast, and welcome to the show for newcomers. And thank you for coming back. If you're a repeat listener, you're about to hear really great conversation with Kirk Marple. And Kirk is the CEO and founder of unstruck data. They're solving some really, really cool problems, some crazy difficult problems, we get into how they do it, the choice on the platform a lot more. But before we get to that, I like to sell big problems. And I like to tell you about the great sponsors that helped me to help you solve your problems. That's a lot of help. That's a lot of problems. first problem is making sure that you have everything you need for your data protection needs. And it's simply easy to go no further than to V e.com. forward slash disco posse. And you can find out everything that the fine folks at beam software have to offer, whether it's on premises in the cloud, virtual stacks, cloud native and containerization, holy Heck, even your SAS stuff. So don't forget to back that sass up, which includes things like your office 365, your Microsoft Teams. That's right, just when you thought it was all protected, because it's in the cloud. Cloud is just somebody else's computer, and no one's backing it up except for you to get it in. That's beam, v e dot m for slash disco posse. All right. Speaking of protection, the one thing you want to make sure you do is you protect your identity, you protect your data, and you protect your traffic. Privacy is a human right. And the fine folks at ExpressVPN they know that. I know I use it, especially because I'm traveling around and you're in places where you don't trust the Wi Fi because you shouldn't trust the Wi Fi. You shouldn't trust it ever. So go to express VPN and find out how to protect yourself and your privacy is very easy. Go to try expressvpn.com forward slash this capacity, and they'll hook you up. It's a real real cool thing. I highly recommend it. Alright, one last thing. One more thing. Seriously. But wait, there's more. You want great coffee? Go to diabolical coffee, calm. That's it. Amazing coffee, devilishly good. And also wicked cool swag. All right, let's get into this is Kurt Marple. Kurt Marple is the CEO and founder of unstruck data. We talked about his own history of building a platform building a product development. Being the the first salesperson, it's, it's a really cool lesson in how to start and found a business and being a technical founder. I had a really, really great time chatting with Kirk. And I think you're gonna be able to tell in this conversation. Let's check it out.
Hi, everybody. This is Kirk marbles, CEO of unstruck data. And I am here on the disco posse podcast. You've you've done this before. Correct? professional. It's funny. I was a DJ in college. And so it's like it feels like going back to those old days. Oh, nice. That's actually the throwback. That's how the whole like intro started. It was like hearing folks that Yo, yo, hey, this is Christina from Motley Crue and you're listening to WNEF ny. So like, it's, it is a lot of fun. Everyone's smile, someone says like, I don't think that you would ask the CEO of like some company to do that. And like, Oh, actually, no, I would. That's the funny thing about this. It's good. But Kirk, thank you very much for joining. I know this is an area that I'm passionate about. Because you are solving a complex problem. That really is I get excited by complex problems. Maybe it's the the nerd in me that loves to look for, you know, where what a lot of people think are generally intractable problems as well. And you've, you've got an incredible sort of history and leading up to what we'll talk about with unstructured data. And also just the way you're running the team, and a lot of the stuff around the the founding of the organization is really cool. So I want to really thank you very much for sharing the time with us. And if you don't mind, Kurt, let's give a quick intro and a bio of yourself. We'll talk about unstructured data, and then we'll, we'll go from there.
Happy to Yeah, I mean, at its simplest case, now I'm a career software developer who ended up starting a bunch of companies. And so I still still write code pretty much every day, but I've bootstrapped companies sold them. And now I guess I was working on this, this concept for about four or five years, just kind of nights and weekends had this sort of itch to think about other ways to manage what we call unstructured data. I mean, in a way we called it back in the day entertainment world. I mean, we just call it media. But it's really not about eyeballs. It's about the data itself for industries. And that's really what we're all about today is figuring out platform and tooling for a wide range of industrial and commercial customers, I mean, from a port to a public safety group to a manufacturing plant. They're all using images, videos, 3d files, documents, everything like that today. And what I mean, my thesis was was there just isn't a good set of tools and platforms for this, and people are really either having to build it themselves. So it's what we see commonly are, I mean, hire developers to write a bunch of Python code or something that index all this data and make it usable, but I had this concept where I thought we could do it automatically, I mean, get you 80% of the way there and kind of an easy button, and then, and then hopefully have a platform to build on to kind of finish up last 20%
Well, and this is, the first thing you have to unpack for people is just even the definition when we talk about unstructured data because, you know, I quite often you tell people about unstructured data and they're like, You mean like no SQL and like, well, now that's, I get where you're at in it. But this is a very specific thing, not not just unstructured data and attaching metadata but complex data, because it's not easy to categorize
it log data is the other one that people always go back to have them in the data dog. So the world and stuff like that, but yeah, we're, we're definitely opinion I mean, it's, it's what you would think of is like in a Google Photos, or iPhoto, for consumer, images, video, most of that kind of stuff. But it's in the industrial workplace, which is an order of magnitude, I mean, 3d use big CAD drawings, all that kind of stuff.
When this is where the, the scale takes you to a difficult place, quickly, in this world, right? Andchuman factor for us is, I mean, we always say like, if you're dealing with like, 100, or 1000 photos, or whatever, we're not the right, like, we're overkill. But I mean, when you start getting into hundreds of 1000s, and millions like, and you have to ingest that data and think about storage, and think about performance, that's where it gets really tricky.
When we're in an interesting spot, where even more people aren't understanding how to exploit and I'll say, exploits the nice way, the harder way of saying leverage, but like exploited, meaning that you ultimately extract every possible, you know, inch of value out of this, this content and this data. They know they need to collect it. And they need to store and they need to. So we've kind of got the first phase, which is ideation like Well, we've got this potential to capture this data. And then, okay, let's start. And then let's try and get insight out of the data. And that's, so a lot of people really did start probably, you know, within the last decade, especially doing a lot more just raw collection, but then they don't know what to do right in there. They're not really you get on particularly. And that's I mean, that's really where we are really being opinionated about is data. ingestion is where we're starting. And just people have data sitting in s3 buckets and having a blob store to have it in Dropbox or something like that. Getting it into something a platform and indexing it is I mean, it's been done in a sense, but it hasn't been done at scale very well, for and across all the different formats. I mean, everything from a point cloud to a CAD drawing. That's the first stage and then data organization of auto organizing that data. I mean, we have a one of our investors is a an oil and gas company. And they have, I mean, hundreds of 1000s of videos of the undersea floor of where their oil pipelines run. And I mean, what we've heard is they just sit in an s3 bucket somewhere, you know, in a storage container somewhere. And they're not well indexed by time and and geospace and stuff like that, or even about what's in the videos themselves. And so we really see a platform that can just get access to the data and organize it as a first step. And then it's about exploring the data. I mean, what's interesting, I want to filter by that region of the North Sea, or that region of the Pacific Ocean. And I want to look in this month in 2017. And only look there like, that's really the next part we're pretty opinionated about. And then the visualization. And I mean, because a lot of times we can give you information automatically, but waiting has come from the data that is unique or something that only matter you would only see we can't train machine learning to be that specific goal. And then exactly what you said is data insights. That's the next phase and we're getting to and our version one is not going to cover that yet. Really by It's really the next fast follow, because that is really where we need to get to is tell people about their data automatically. And augment.
Yeah, this is the, I'll say sort of the, there's a panacea of what we hope to reach. And there are a lot of islands of real incredible innovation that we have to traverse to get to this great place. And it's, it's neat when you talk about, like I said, the industrial use cases you've got, you know, again, the first thing as humans, we think, is like, oh, I've got to get, you know, photo data. They think of fairly simplistic use cases, the first thing you think is like, Oh, I can tell the difference between which which of my kids is in the photo, like, it's like, we immediately try to attach it to a thing. But like you said, the sea floor, the first thing you think is like, this very hard to differentiate, and you've got to attach your topological metadata about it, you've got to attach location data. Like it's, there's a fantastic Yeah, well, that's right. Yeah, it's not even moving. And how do you like, you know, okay, it's that picture. And I see Oh, wait, there's a rock on the on the pipeline right now. But when did that rock appear? I mean, what is it in the last set of scans? And how do we know it's the same part of a pipe? Like, that's really where it gets tricky.
Yeah, the, and this is the problems that are being thought about in these industrial arenas, you know, again, you get into mining, you get into, you know, and just topological and GIS work, that's, that's happening. It's incredible, because they've, a lot of them have should have just ascertained, like, that's it, that's the best we got, right, we we've got satellite, we've got certain things. And we're going to have to do some kind of level of manual insert of human managed classification and storage. For the longest time, it was really seen as, like, the scale would be too fast, it can't be done in any kind of real time. So they just kind of threw their hands up and said, like, okay, let's just fire it up into an s3 bucket, as he said, probably exposing it to the world by accident, and then we'll deal with it later.
It's interesting, man. And we've heard this time and time again of I mean, number one, it's either siloed where the maybe the data has been captured, but it exists in, in separate sort of systems. And so they're having trouble or having trouble kind of connecting that up and seeing holistic across that, that's one problem they have, scale is another one. And then honestly, just I mean, getting as we get into different file formats, like 3d point clouds and stuff like that, it's it's a technically more hard, it's a harder problem. And I mean, dealing with 50 million points in a scan of a facility is a lot different than dealing with, I mean, one gigabyte, like JPEG, or something like that. So it's a big, big difference.
Yeah, while the rest of the world is trying to tell the difference between a Pomeranian and a blueberry muffin, we got bigger, bigger things to solve for. And we'll look at autonomous, you know, autonomous driving, the amount of data points that are continuously updating in real time. And that's, that can't be aged data, that's where they need to be able to have rapid access to classify, measure, apply other machine learning. And it can only be done once, they actually sort of set the source of truth as close to real time as possible. The data itself right,
is really true. And I think even worth seeing is there's a kind of sifting mode of you have so much data, you just kind of need to get to the I mean the needle in the haystack a bit, but you don't really want to, there's not ever one thing you're looking for, you're just trying to get to kind of sift it out smaller dataset. And then maybe you're using manual techniques using machine learning or whatever. But it's, and that's what we're looking at is I mean, okay, I just want to carve my eye this massive data set on a carve it down by like by tags by time by geospace, and then maybe, maybe I'm left with 1000 images or 1000, whatever's and then that's what I want to do labeling on machine learning on and things like that. But you can't, you got to have a good organizational way to start at that massive data warehouse first, and then just get to what you're trying to work with. And then we're also adding is I mean, the concept of data collaboration of your teams need to discuss that data in a way that is localized to the data and actually literally have like almost like a Instagram feed commenting history on an image of the sea floor. I mean, it's like a I mean, there's consumer patterns that we're trying to bring a little bit to have your team's discuss the data right in, in in place. And then maybe even leverage that like, and when you start to see, we can run machine learning on that say, hey, why does this? Why are these assets getting a lot more chatter than than other assets, things like that. So I mean, it's not, that's not day one for us. But I think once we start to get that flow of data working, that's where I really see it getting to
go in to talk about the team, because and your own background, obviously, this is not, this is not something somebody says, I, I'm a general front end kid, I think I've got this crazy idea like you know, you you've got some real history in where this come from. So talk about how you began and where the suddenly became obvious to you that there was a problem.
I mean, it's so I mean, as I said, software developer, did my masters and CS and just ended up at Microsoft doing, I mean, a bunch of really interesting things around like 3d virtual worlds. And I mean, worked in the media, Windows Media Player group. So started cut my teeth on metadata. And that was really where I started to see the concept of media metadata around that. And then really, it was starting my own company that was doing video transcoding, you didn't we did audio transcoding, things like that we're working with metadata from the record labels actually trying to collate that into a feed back to me, this is pre Spotify. But essentially, the same thing with the record labels would basically publish it almost like an s3 bucket. But it was like on an FTP server, here's my data, here's my audio like my WAV files, my metadata, and what essentially we were doing transcoding for a company at the time to burn CDs on the CD kiosk. And that was really where I mean, it's it's very similar in a way to what we're doing now. But it's taking that metadata, taking unstructured data in those audio and creating some structure from it. And we were like, canonicalize, the metadata and doing something or we could create a feed. And that was I mean, I had that company for about almost a dozen years and sold out. But I mean, I learned immense amounts, just dealing with video at the broadcaster's dealing with audio. We even had an image archive we built for one, someone. And so I've always been kind of cross media type in that way. And really, I see this as a second generation of that, but not for media entertainment, companies, like studios and stuff, but really, for everybody else. And and by incorporating the 3d side, the documents, CAD drawings, all that is, it really opens up to a lot more stuff. And I mean, a couple years ago, I got interested in knowledge graphs, and that was really the turning point of just on my own, just trying to learn and I saw how this application of taking media, taking the metadata and mapping it to a knowledge graph. I was actually working on a podcast discovery platform. And this is my kind of passion project of that was what I was trying to build nights and weekends. And then I realized everything kind of synchronized this last year, I was like, maybe it just makes more sense to map this to industry and commercial and kind of instead of consumer. I still think that idea has legs someday. But it's a mean, was able to get funding for this idea of basically taking the same approach. I mean, but unstructured data, media data to a structured knowledge graph, and make it searchable and all that. I had a team of I do have a team of folks that I've worked with, we've worked with multiple companies together. I mean, they, they came along, and we're happy to join the ride. And so it's a great bunch of folks that product, QA, design, engineering, everybody. And yeah, I mean, so we so I had the backend of the product, mostly done like 80% done before we started because I had already been working on it. And so we've mostly been working on the front end. And so we've been basically started in like, March of this year is when we got funding and what's that about four months? Yeah. We're, we're probably about six to eight weeks away from getting into the hands of customers. So moving very fast. And the products already looking great. I mean, it's, it's been a bit of push. But uh, yeah, I mean, we have design partners, we're starting to talk to you, we just hired a new head of sales. And we're just right on that cusp of sort of proof is in the pudding now, like, we have a product and we made a opinionated kind of thesis about what it should do and why it's doing it. Now we just got to prove it and, and make make customers happy is what matters at this point.
With your understanding at real enterprise scale, you know, how important was that in your ability to now have a go to market with this product because this quite often is a really tough battle when you go especially the companies at the scale that are going to leverage this. They generally have a very strong engineering team or set of teams. There's a lot of NIH not invented here. So it is So I'm interested in where you know that you can, you know, leverage both relationships as well as just general market understanding to speed that GTM
it's a really good point. I mean, I think, I have assumed that even like I was at General Motors for a time working, that was where I first got exposed to kind of industrial media as you want to call it, or industrial unstructured data. And it's like, I'm not even sure if they would be a customer for it, because it is a bit of, like, it's a bigger ecosystem. I mean, they have a lot more engineers, like there's we are competing against what I call DIY, but he and he kind of sort of NADH in a way, but it's, it's, I mean, there's, there's a slice of customers or companies that I'm just not that worried about. I mean, like Uber, I mean, Tesla, like, they're not going to use our software, we're gonna they're gonna build it themselves, they already have tools for for stuff like that. But if you take a notch down, there's a huge swath of industrial companies, from manufacturing to chemical to any I mean, other robotics, OEMs, and things like that, that, I mean, I believe, I mean, can really make use of this, and it's not going to solve every problem they have. But hopefully, it's a catalyst, let them do the hard problems. Like we kind of want to be an easy button, really more about the data management. We're not going to solve every machine learning problem in the world. But we want to plug into other platforms for to let people invent and be more bespoke specific parts. I think media initially, our goal is not to like take over the entire machine learning world, it's really to solve a problem that a lot of people have to do themselves, and then build a platform for people to integrate with. That was in in my old world, no company that was a big jumping off point where we had an API that some of the broadcaster's could plug right into, and they stopped even using our front end web UI, and just talk to the API. And and we just became like a core component in their system. And that's I think, that's the path I see for us is, you got to start somewhere, we have a really nice UI and with data visualization, and essentially data analytics, soap UI, but the platform's there for the taking. So once we get people kind of flowing with data, we have a graph qL API, we can say, hey, you want to do anything interesting on your data that we already have? Here's an API for you. And so it is a two fold kind of go to market?
Well, it's important too, because you know, even where there are folks that have gone down the road, in fact, sometimes the best thing you can have is a team that have made a run at it. And they've realized the the challenge, because then they start to think about lifecycle, even if they do it, well, they now own and operate that in that app infrastructure forever. And you bring them a solution that effectively sheds a layer within that stack for them. And then you now just, you can handle lifecycle of the application and just simply present an API that can very cleanly deprecate for them, they don't have to worry about any of that stuff. As changes occur, the underlying infrastructure, you can tap into different backends. Yep, it's, it's important. And because of the scale of folks that you're, you're going to, they'll probably have seen some of the limits already. Yeah.
Even things like making best use of cloud storage and putting unused data in there, cold storage, now, stuff like that, it's like, I mean, that's a sort of edge case that I mean, what every vendor, every company in the world can have to implement their own multi tiered storage. I mean, that's, that's like, we can just plug that into the platform, and they get it for free. It's like, hey, we'll save you a little money by I mean, pushing your data out, or archiving it or whatever. That's the kind of stuff that we can do easily, and give them insight to say, look, I mean, hey, here's an analytics view of all your data, you have 47% images, and 50% of that is JPEGs. And, I mean, 80% happened to be in this geospatial region, and like, we show them clusters, like, that's where I want to get to is showing them those automatic insights on their data that they would not be able to find themselves manually. And, and even I mean, deduplication of data is important as well, I mean, they may, who knows, I didn't realize that someone, I mean, copy their whole s3 bucket into another s3 bucket. And it's actually the same data. We can automatically like we index even down to like a, what's called a C for ID. It's like a unique hash of the data that we can then D dupe on. So there's really I mean, that's the kind of stuff that sure, the core of this, maybe other people can do, maybe not at scale, but they can solve enough of their problem. And then it but the more features we can add to be like, oh, wow, you get all this for free. It's it's managed by someone else. That's where I think it gets pretty interesting. And we have I mean, we've talked to customers who tried to cobble together a solution in some of these spaces, but it's like, they're not the right I mean, they did a great job at it, but to maintain it for the next five years. It's not It's not their job title, like, I mean, they're not they're doing on the side kind of, but they have their own day job simply they're doing as well. So that's where those are the people we're trying to help.
Yeah, I may enjoy changing the brakes on my car, but I don't want to be responsible for all for the cars every season, right? It just, I may get a little hankering to give it a whirl again, like when I was a kid, but I don't want to own it forever at that point.
It's exactly true. And I think I mean, that's where we do want to play well with the ecosystem. And when people do want to have their hands deeper into it, I mean, we can they can access the data and do their own compute on top of that, but I mean, I just believe I mean, especially at scale, I mean, once you start to get to hundreds of 1000s, or millions, it's, you just can't do it manually anymore. And so we're we're trying to just build that tool, those tools and platforms to make that easier.
But it sounds definitely that based on your understanding of where the I'll say the floor is on the appropriate customer, you know, right away, you can basically qualify out to make sure that you're not wasting anyone's time, as you say, like, here's an example use case we've got it doesn't sound like this is where you're at. And they can say like, Oh, no, no, no, we're not, we're definitely not there yet. It's. And also, because you've lived the actual experience of running the environments, you've got a better insight, to be able to have that conversation, instead of
developing that go to market strategy a bit better, we have a new guy we've just brought on is awesome. He's been here about a month, but we're just still trying to, like define the storytelling part of it. And just like, okay, who is the sweet spot kind of worked backwards in the customer journey, because I think like we've had some people show interest in the Knowledge Graph side of it. But our version one is more opinionated about geospatial, like, if you don't have geospatial data, we're really not at the sweet spot right now. And so because that's a lot of what are you I mean, has a map view and has other stuff that it's, you wouldn't be getting the benefit of it. So we have those people, we're gonna have to turn down just because it's not our, it's not a perfectly aligned yet, give us a year or two, maybe we can stretch to that. But right, it's, I think that's where we just need to kind of stick to our guns a little bit and say, Look, I mean, here's what we're focused on for go to market for this year. And really, just, I mean, I want to get really sticky customers who just start putting data into the platform and grow with us too. And as we grow with as we can grow with them,
when and this really brings to you, given your development and product experience, and that you've you know, been to the you've been to the well before as a startup founder. That idea of like ruthless pragmatism when it comes to features that you bring into whatever you bring to the market. Again, if you found that you've got a good handle on it, how's the rest of the team? Because it's very easy for us to be like, Oh, if we just added one or two more features, we get this customer. But you don't actually know that that's the case. They're just saying that because they're trying to defer the conversation half the time, right.
I mean, that's the truth. And I mean, that's we're pretty much right on the cusp of that right now, because we're just about the show, but we've taken in a lot of customer input, but until they get it in their hands, it's a different story. And so we're just about to go into a private preview kind of early access mode in basically a month and a half or so. And that'll be really where things get real is I mean, have we infer their right, their problem set correctly? And what are we missing? I mean, I'm sure we're missing something. And we've actually left a good bit of time in the roadmap for the next quarter. So just to not to be super aggressive on features now that we got to this point, too, but to know that, we're going to get a ton of ideas from the folks that actually start using it. And we have to have capacity to be able to respond to that. But it's also I've been in that situation to where you start to get into everybody is a unique like, like, need something separate. And you can get really randomized by trying to over respond. So I mean, I hope I have the pragmatism, and that's something I think about a lot is to say, look, I mean, 8020 rule, like, okay, who do we who can we hit 80% of these people with with these features. And you got to be able to say no to customers and say, Look, it's just not in our wheelhouse. And I understand what you're saying, I understand what you want. But like, we're not going to build that. Like, it's just not something we can respond to. But that's where I hope to get to, is give them an API, give him access to the data and be like, hey, you want to go find a vendor who wants to build this as like a plugin? Go for it. I mean, here's an API to integrate with, or here's a web hook that will will tell you when something happens. That's my goal is you always have to have that escape valve of the API and things like that.
It really I often tell people, especially in you know, my own sales organization, anybody to talk to you in an advisory You've got to at least have that option there. Then the onus moves to the customer to really define if that's a need, or if it's, you know, I get asked all the time and say like, oh, you're cuz every cloud providers, yes. You know, you know this Well, except we don't do as much with Google. And you know, and they'll still often get asked that. And in the case of many of the folks that I talked to, they are just asking because they want to know not because it's a real sticking point for them to not buy in, it's an intellectual thing, especially as a as a nerd trying to defer the sales conversation. I'm going to throw these things out there. But when I hear what Yeah, we've got a, you know, an API available. So you can plug other third parties into it, we can get web hooks in and out, you hear these things you're like, Okay, so there's flexibility there. Yeah, I've got options, I feel okay, taking this first step, knowing that there's many ways in which I can go to the next phase. And I think it's a
difference between kind of, I don't know, the right term, but sales and sales engineering, like a salesperson would hear that. And they might think it's one to one, like every request is about is a is a must be able to train your sales team properly. The filter that I think is also key, where it's like having someone with the technical inputs, they look, just because they asked about AWS, like to your point, or Google is not a not a requirement that we need to go to the product team with. It's I mean, we got to dig into that a little bit more into Look, I mean, if I mean, in this way, I mean, I love that part of the technical sales side, because I can go then dig into it, and kind of listen for those questions, and then lean in. I mean, we just, I mean, we've had the same question about like, on prem, like, I mean, do you want to run all of your stuff on prem, and we're actually managed in Azure today. And we can deal with data on AWS or Google or on prem, but our code, like it's literally running on Azure, and using a lot of managed services and for us to repackage and be portable, we could do it, but I don't, I don't know the value out yet. And, and, and we have had people ask about can you run on prem? Can you run on AWS? And my answer right now is Yeah, I mean, theoretically, we can I mean, someday, but we essentially have to swap out our database and swap out our whatever like I mean, our Kafka like our event hub for Kafka or something and but it's it's a it's a zero sum like it's a it's a sideways move at that point where I mean we're not getting any value from that but we just have to make sure the customer is and they'll then they they get that value to me like would they be a customer otherwise it's just like a like to have or it must have
Yeah, this is really where and this interesting you know you're you're obviously you're you're a technical founder and you're but you're a you're a sales versus leader as well which is a rarity that you can have this sort of diversity of capabilities at the business and tactical layer your you truly are a unicorn knows of humans correct because quite often I find this you know, really strong technical founders, but they probably wouldn't pass a Turing test if you and then you don't want them in front of customers. So we have this really chat so it's it's very interesting to me that you've been
trained in a lot of areas it's honestly I mean, a lot of it's just the Ross Moses I mean it's I don't have an MBA I just have a master's in CS but it's like, I mean, I learned a lot just from the scars along the way I look at and I had a great business partner and my first company who was kind of the relationship guy on the sales like it was the kind of hacker hustler model of it and like Yeah, he I mean he had a conversation with anybody It was really I mean everybody loved him and I think I kind of learned from that a bit where I was like super heads down building product like i was i mean coding and I I kind of had blinders on to that side of the business a bit and over time I sold a company and then I worked I was kind of CTO or VP for some places and I mean I was an exact so I had to interact with sales and marketing and stuff like that and at VC backed companies but I think now taking another go at this it's like yeah i mean you have to you really have to be customer for a customer lead and and really think about the customer and be pragmatic about it but i mean i'm also I mean writing back and code still so it's like I'm having to wear those two hats but in the in the daytime it's I mean yeah I mean we got to it's it's to make money you got to make customers happy and that's that's where it all starts
with and really especially in a in a very technical sale the CEO is the first CRM as well right? You have to be customer facing you have to be the head of sales, alongside your relationship sales, you know, person or team. But it is a there's definitely a period where the handoff has to be slow and methodical and proven before you You can know that because there's I'm curious, you know, with a very technical sale, they've got interesting sale cycles, too. I know like the enterprise stuff can be really long, it's hard for a lot of folks to be able to weather the storm, a long sales cycle. So you kind of know what you're heading into with that, though.
Yeah, we're trying to follow I mean, we've been opinionated about this being more of a consumption billing model. So it's more like a Dropbox kind of model where the more data you put into it, the more money you pay, right. And we have, I mean, there's got to be kind of more margin multipliers, because we're like doing machine learning on it. And we're running computer vision algorithms like, so it's not gonna be like at cost for storage, like a like a Dropbox or Google Drive. But that's what we're seeing is, and it makes it easy to get into because literally, you just couldn't get an account and start dumping data into us in 15 minutes. And so I mean, five minutes early. And so it's, that part is really we're being we're taking almost a bit of a consumer prosumer kind of bent on it, say, look, you find us on the internet, like you hit a, you get a free, you can basically sign up, get a free trial, and get, we're essentially doing storage quota as our trial, like, we'll give you a couple 50 gigabytes for free. We're still trying to figure that out, but and just play with it and see you mean and give us I mean, if you like it, give us feedback, and then expand from there. Like we want to get into more of I mean, an organizational level sale. And so it's more of like a land expand upsell model as one sense. But then we also have an outbound sales focus where I mean, there's the bigger companies that we can go talk to and work downwards from, like various innovation groups, and things like that, as well. So it's, it is a bit different. I mean, I, I've been in that long term, enterprise sales cycle before, and we want to try and stay away from that, like, we can't, we can't support like six month sales cycles for sure. We just want to have people just try it out. So I guess, product growth is like the term I guess he said,
right? Yeah, and this is important, because you're in the end, of course, we always are ultimately selling to a CIO, CTO, like, the person that ultimately signs the bigger long term deal and understands where you become core to their ecosystem is in the executive team. However, to be able to get in and get practical, and get sticky, they require like, I love that you've got this ability to just get in there, have them try it, have them understand it, and then say, Okay, yeah, you know, like the matrix, just more.
I mean, I mean, I listen to a lot of podcasts, but just like startup founders, and folks like God, I'm hearing the ones that, I mean, are like getting in there, I mean, getting just kind of getting their hands dirty, like, what's the one, the one product product, dirty one, but it's, I think it's like they they are, they're selling to developers, but then they have to sell to the team, as well. And so like I have to grow into that into into some of those is, is really tricky. And I think for us, initially, we're going to sell into a line of business users and try and get them using it, like, Hey, we're just trying to automate your day to day workflow. And that's the first goal. We're not as focused on like machine learning, data engineers yet. But once we open up the API a bit more, that'll I think, be the focus, we're actually gonna bring somebody in the house to kind of own that from a thought leadership standpoint, have somebody hopefully bring it in later this year? I think that'll then open up that whole workflow of okay. Data labeling, I mean, training, all that kind of sucks, cuz I'm not a data scientist by trade. I'm a platform guy, but not that world. And so I think that's what we want to expand to. And I think there's, I mean, but you have to talk their talk as well. And so that's a totally different type of cells.
Yeah, you need the evangelists, you know, somebody who can be very much from that ecosystem, whether it's, we call it talk about developer advocacy, or whatever, that we the name change events. Well, it's like one of the old g technic technology evangelists. They're like, Oh, isn't that kind of closer to like sales? Like, no, it's developer advocacy, before there was developer advocacy. I just didn't change the title.
It's I mean, it's a big it's, and that's where I see like, I mean, we're kind of focused on 2021. Right now, we have a big set of goals for 2022. It's like an order of magnitude more, but it's like, we got to get that stickiness, get some feedback, and then really put the pedal down. And that's what, from a fundraising side, we took in a good bit of seed. But it's, I mean, I think there's a lot more like to get to that next level of growth. I mean, we want to show traction first, but there I mean, I have so many other things about, I mean, we're actually building a mobile app for data capture. And we want to do things like live data capture, connect to edge devices we want to do. I mean, I've been thinking about augmented reality to like see that change over time tracking, in almost like a ghost image as you're walking around. Looking at your facility, we have the data, we know where you are in 3d space, we can essentially show you that as you're walking around the data from the previous inspection and things like that almost an augmented world. I think there's so much high ceiling, I'm cool stuff like that, that would, would help the line of business user side of cyber thing. So yeah, I mean, we're, we're looking to hit it pretty hard next year, and, and hopefully, hopefully, really, really grow this,
especially if you think of like industrial flow, like that is something where to be able to understand true, you know, path optimization in the physical sense, it's, there's no way to do it, because they, unless you're going to get every single person to wear a sensor, you know, and then build a system in which you can measure their place at any point in time in a 3d location. Just building the collection system is like, that's a huge problem unto itself. And then now it's really classifying and then making sense of that, that database,
that was the first customer signal that we really heard that was new, when we've got started is I mean, everybody, like they're saying, they're not at their desk very often, they're actually going around the facility, they're very mobile people, per se, and six hours a day, they are capturing data, but then they have to go back to their desk and figure out how to upload it, how do they get it off their iPhone? Where do they save it, and if we can just streamline that part of it, the data captures collection side, that's really what we're focusing on right now. So we're building a pretty cool kind of session based data like capture session based UI tool, and then what are you gonna expand that out to like 3d Point Cloud capture with LIDAR, and I'm gonna do iPhones and, but it's going to be basically image audio and video at first. And but the cool thing then is we're doing like speech text extraction. So you can do a voice memo of saying, hey, equipment for like, part ID, ABC 123 has a problem, here's an image of it, here's a video of me walking around it, and then just automatically sync it. And then we can now correlate that with, I mean, eventually, when it correlate that with their database, say, hey, you spoke the words of this equipment ID, we can give you like all this other information of the, I mean, the history of it, and all that kind of stuff, as well as see, Hey, what did this look like three months ago, because we have that data as well. So I think that that holistic view gets really interesting,
because then you're combining both sort of spatial data as well as time series. And then each and each unto itself is its own complex problem that people are trying to solve. And then to put them together, it's basically you've got a physics problem now like, that's the level at which you're working where, unless you're friends with Eric Weinstein and his career a tough nut to crack.
That's where I think that's what gets me excited is we call it the triad of like time and space and metadata. And we extract data in all those three axes, and then organize it and visualize it in those three axes. So the team is now that's like a term that keeps coming up almost like a meme internally of, as long as we kind of bring it back to the triad of those three axes, like we're doing the right thing. And I think it's, we just got to get everybody's data aligned to that. And then we can start innovating around around those axes.
When thinking about the, the sort of future availability, now you've got two ways in which you can apply a lot of work that your platform can do, because number one, of course, you capture. So add ingestion, there's a certain amount of work that can be done. But the beauty part is that once the data is at rest, it doesn't take away your opportunity to then augment, do additional things. So even as a person takes on the platform early, they'll be able to realize additional benefits down the road, because that stuff that you may do it, ingestion eventually can still be done at rest. And then they can sort of gain from the rest of the community
levels of data at rest to where we have the metadata we've extracted, and then the source data. And we are looking at, like, we can act as an archive, or we can keep around their source media, like keep a copy of it. If we want to backfill and run ml algorithms, you got to have the data. And so it's I mean, the metadata is useful for some things, but you can't I mean, I want to go I don't know we there's a new version of a model that was better at identifying grace or something you got if you want to go and read evaluated against all your data, I mean, you have to touch the data. And so I think that's an interesting part. And then also the metadata. The writing algorithms, just on the metadata gets interesting of, we're looking at doing like similarity search, or clustering and say, hey, why, I mean, why are all these pictures clustered together via the same tag, and that's where we can pull insights out. That is not obvious. And so part of it's just visualization but part of its actually writing algorithms to to glean more insight. So those are the two areas that are pretty cool.
You've really got, you have also like a an infrastructure platform challenge that you have to sell just in being an abstract data that in and of itself is incredibly complex and you've chosen a talked about for about a year, right, you're running in the as your platform. Now you must be doing stuff in there that they are excited to watch.
And it's I mean, it's like I think, I mean, we're, I kind of had this where I'd kind of been dabbling in the side and going through a couple iterations of like, I was using what was called Service Fabric, which was kind of like almost a Kubernetes cluster model. And then I moved things over to Azure Functions to be purely serverless. Because I was paying for all the infrastructure myself, and I was like, I don't want to have to pay x $100 a month just to keep this cluster up, that if I just move it to serverless, then I'm only paying for what I need. And so once I kind of taught myself that whole model, but also, I guess the the event driven architecture has always been kind of natural to me. And so we're all built on that. So it's everything is asynchronous, everything is event driven. But that actually makes us lets us work at scale very well. I mean, you could dump 100,000 things assets in there at the same time. And you're I mean, your farm just bursts burst out, the database can scale up, I mean, the function, the compute scales up. But there are limitations to like, we don't have access to GPUs in a serverless architecture, which is a limitation. So as we're doing some of the 3d work, we actually have a limitation day where like, we can't create thumbnails of the 3d geometry, because all the libraries I could find required a GPU. And so it's things like that, where maybe we do need to have a separate like Kubernetes cluster with access to GPU to farm out some of those pieces or like video compression GPUs faster, in some sense. So we do get into an infrastructure, like, how much do we need our database to scale, I mean, we're using essentially two database types of more SQL, like a no SQL document, and a graph database, as well as a search index. And so over time, we may try and collapse those into I mean, look at other technologies and things. But I pick I mean, I really like Cosmos dB, just personally on Azure, just because of the scalability of it, it's very easy to use, you don't have to worry about indexing. And that was, I mean, it's not the most perfect thing in the world for every case. But for what we're doing, it works really nicely. And but it's also we've abstracted it enough that if we need to swap in other databases in the future, we can and so I have been forward looking enough to say like, yeah, we might need to move them to four j or Tiger graph or something like that, in the future. Or maybe a customer has a reason they want to use that or something. We can, we can swap pieces of it. And now
it was the recent thing, which was totally new for J. They it was whether it's like a funding camera, or what was a recent announcement that was basically like, okay, graph is real, like, this is a thing now like, it's so funny that we went from like traditional tag as the like, de facto metadata of, you know, the future. And now once we saw how we can leverage graph, it's like, oh, okay, this is literally from crawl to run, as far as capabilities. And now we get to, you know, figure out how to bring a lot of those applications forward to, and also, just like any technology, probably doesn't maybe doesn't need graph, like maybe the traditional structure is just fine and dandy.
For us, streaming we, I mean, I've been, I mean, I wrote a lot of SQL code back in the day, and then kind of started just doing more just classic, NO SEQUEL, but it's, I mean, the thing I like about graph is you can kind of almost invent your schema on the fly, like, we come up, we're like, oh, actually be really useful to have an edge between this node and this node. Okay. I mean, it's like, like, there's no, there's no backfill of schema and schema migration and all that kind of stuff. And we do have to think of kind of like, doing it in place, and how do you backfill those things. But it's, it's really I mean, it's almost like an indexing model to me. And we were kind of using the graph as an index. We're using the just a document store, kind of as our raw JSON store. But then we're also doing things like full text indexing and stuff on top of that. And I do think I mean, there's, there's a chance to, like, what I kind of quote invented is this kind of hybrid of those three models. I do think could be we could invent our own database that was perfect for that situation. But I don't know if that's really the place for us to be I mean, I think it makes more sense of, I mean, there's so many new graph vendors like Katana graph and Tiger graph and all these that I'm going to assume there's a better mousetrap out there. But but there are things we're doing like integrated search and large JSON blob storage, which most of the graphs don't support today because they're mostly property. key value based, right? Yeah. So, I mean, I haven't come up with an alternative to kind of the architecture that I kind of came up with. But I do think we could get more of a spoke about it later. And if I mean, if performance is key, and if we're all gated on, like our, quote, database performance, we'll probably have to rewrite it in a more native fashion. But I love having managed services, and I don't have to think about, like, auto scaling. And now, or and all that, like, I'm a big fan of managed services, I would prefer not to do that.
But it's very much like the same reason that your platform is powerful. You can tell people like, Look, do you want me to be writing the database? Do you want me to be writing capabilities that that meet your needs, right? Because ultimately, it's abstracting the cost and your infrastructures abstracted from them. But of course, because of technical sale, you'll often we all kind of love to say like, gosh, how's it all built? And, you know, I'm curious freedom, I'd love to hear like you talk about invents, and you know, how much of the stuff that you have is, is wrapped in IP and patents? Because it's a very interesting space that you're, you're in?
Yeah, I mean, so I mean, all, let me say, I mean, the majority of the IP other than managed services is all less I mean, it's all built from scratch, like, and I'd written most of the, I mean, pretty, I wrote basically, the backend all myself with a couple of contractors I had early on, I was working on more the billing and customer onboarding side, like I started a startup company back in 2015. I guess that was just, I worked on that basically full time. And that was just the kind of SAS onboarding pieces of what we've already and it's still in the box, like, it's part of what we're doing. And then I started building layers of like media management, and then the graph stuff evolved and all this other stuff. So that's all homegrown. The places we're using third party is like around. I mean, we're using some open source technology for some of the parsers that can parse different image formats or media formats, we're doing, obviously, all that Azure infrastructure and things we're taking me we're using Event Hub, Azure Functions, cognitive services, cognitive search. And then I mean, I think the other part is, I mean, the UI is totally, totally, but we are licensing some, like react components, from some folks that there's some, like, really, really nice react components out there that are, are just better than we could build in the time ourselves. We, I mean, those are paid for. But yeah, I mean, and pad wise, it's, it's not something we proceed yet. I mean, I'm a little, I don't know, I kind of have mixed feelings about that whole I mean, market I've, I have some patents from when I was at Microsoft, but I think as a startup, we're more concerned right now in time to market then then the whole patent world, but it is something we may come back to.
Yeah, that very much comes from it's funny, as I work with a fellow I'm an advisor to one company and, and he very much came from, like a patent strong approach to things. The first thing that they do is they eventually go and sort of, like protect a lot of the IP, because in like, his particular business, and his successes in the past have often been the product goes you know, to whatever level but then they license it the patents and that's actually part of his chosen business model. So I like that you're just say like, that, you're sort of focused on this is where I know I can be successful. Let's go and hit the road. And this project became I just realized, I can't tell how much sound microphone I've got someone powerwashing the side of me I can't hear. I can't muting myself out just in case of like, oh, there's other words from the contract just decided to come on the day you're recording?
No, I actually just did that last point on patents is a buddy of mine who has another company, he they're much more patent heavy. And so they are looking at technology licensing. I mean, so it's not the wrong decision. It's just we're so hyper focused on just we want to scale we want to get get data and get to market quickly. And I mean, and a lot of the stuff it's sure it's it's stuff maybe I've done before other companies have done before, but in a different market. And so it's just not for this, like media entertainment has some of the basics I mean, we've already been covered, but we're doing it for really a different market segment and in a very different way to
wait and quite especially when you get into the sort of the patent building, it's a huge amount of effort and time that you invest towards it. And part of it is that you may want to find that, you know, the method which is ultimately what you're patenting may evolve as you bring on the first number of customers and then at that point, it may make sense that Okay, let's go down the road and, and, you know, protect this, you know, through that means, but in the meantime, you know, you're you're solving an incredibly complex problem and that's where you want your engineers to be focused. Know it on city. I'm down with a sea of lawyers to write down, you know, drop nice pictures of the method.
And I've been I've been there and yeah, it's I mean, it is distracting and i think it's it's not wrong, but it's just like, I think for us right now is, it's, we're not like inventing like really key, like gnarly algorithms at this point. Like, it's really more about the process. And I mean, developing workflows, and just ease of use and all that kind of stuff. So, I mean, there are definitely some places that are very, very unique, like the whole Knowledge Graph architecture is completely out of my brain. And I mean, maybe something we should we should look at doing some production in the future. But it's, I mean, have a little bit of time. We haven't I mean, we haven't launched yet. I mean, so we're still haven't really shown it publicly. We may come back around on that, but it's, yeah, right now, it's, it's really just, I mean, we're more focused on customers at this point,
if nothing, you can use these podcasts as prior art. Now, actually, that so the Knowledge Graph piece, maybe for folks that are fresh to this, give a bit of a description, maybe an analogous description, on, on where Knowledge Graph comes in, because this is a really amazing area of work that's been sort of research based, as a technology, but we're seeing real practical implementation. So good, open some folks up to it. I
mean, I think it's really interesting. And it, I mean, there's a theoretical side to it. I mean, I remember in undergrad taking, like graph theory classes and stuff like that. But at the end of the day, it's I mean, it's a simple enough case of, like, most people are used to sort of a tree of I mean, you have a parent, you have children, I mean, you have grandchildren, I mean, that kind of concept of, I mean, you have a hierarchy, some kind of hierarchy comes up in a lot of applications, or a lot of use cases. But what gets really interesting is when you can connect those dots of, Okay, this, no, this parent over here actually has a connection to something in this other sort of tree over here. And that's where we're creating those threads. So you kind of have these sort of separate trees that are all related to your your media, like I took an image and we actually create thing today, it's about a dozen or more, quote nudes in the graph from one image. So we're, we're devolving that out like a decent amount of data in terms of the files, we're tracking the metadata we're tracking, but then we were putting in about another concept just of tagging. And I've been involved in metadata metadata standards for for a number of years now. And we decided just to say, look, let's not worry about like a custom taxonomy and really getting too crazy about it. But let's just call them tags, let's be very simple. Tags are I mean, labels are sometimes called and map everything to a tag structure. And so sure, you lose a bit of like, I mean, is the word. Perfect, good word. But it's like, is it I mean, it there could be different context or different perspective on the same word. I mean, it can mean different things. But you can infer a bit of that by what it's connected to. You don't have to have this massive hierarchy of like, I mean, okay, is I mean, when you say Seattle, does that mean, is that related to a sports team? Is it related to the city? Is it related to Chief Seattle? I mean, like that, we're not getting that specific about it, we're trying to write simpler, but then you start to get really, I mean, if you have this taxonomy, or applying this tag set, you can then be like, Well, okay, grab one tag and pull on it, and see, well, what is it connected to? And what are the kind of what's the weight of I mean, why, like, why are all these assets over here connected to this tag, but not all these assets over here? That's where I think it gets really, that's where you got to get the data into the graph. And that maybe isn't the really exciting part. But it's then well, what can you do with the graph? And can you enrich the graph? So where do you go, typically a graph, you're talking about entity extraction, and entity enrichment? And so entity extraction, in my mind is more the, how can I create those nodes in the graph and put them into the graph from unstructured data? So you're kind of extracting some structure, graph structure from unstructured data? And then there's going to be enrichment, which is kind of a loop to say, oh, once I have it in the graph, can I enrich it, add more data to it. And I started when I was working in the in this podcast idea, I was looking at, oh, I found a reference to a podcast in a webpage. And then I would enrich the podcast by going to grab all of the episodes of that podcast, and pull them into the graph. And so I would create, I would basically unroll all these links into more nodes in the graph. And then you could say, oh, there's a guest in this episode. And you could create an edge between that gas on that episode and other episodes they were in. And so it's I mean, and that's kind of where my brain was starting this whole process. And then I took that model and applied it to industry realized Oh, wow, there's a conveyor belt that I see in a picture, I have somebody say the words conveyor belt in an audio stream, and I am finding that in a document somewhere, and you essentially are extracting those nodes and then creating edges between them. And that's I mean, a lot of the same application for other non business use cases. But that's how we're applying, applying that same technique.
If you could solve the podcast problem, I know a guy that has real trouble with statistics. And let me tell you, there's no greater lie than one backed by podcast statistics. But as far as the end that that really is the idea, right? You said, not enriching the graph for anybody that's worked in traditional database architectures. And then we move to like the idea of like, oh, let's add key value stores as another way to attach metadata to it. Like no matter how you slice it, anybody that's had to like press go on a on a database migration in a sick years, a massive amount of data, that just like, no matter how many times you do that, you've had to apply a human understanding of this data. And you've had to map it out, we used to have DBAs. And they'd have these giant sort of UML diagrams and all these different neodi RDS. And then in the end, you are only as right as understanding the output that you needed. But if it suddenly changed, you were fundamentally wrong at this point, like you immediately were wrong. And like, forget about your Congratulations, you get data normalization, like a good students. But now what you then you add to extend the database versus graph is really, truly a living, breathing way in which you can evolve the relationships, which is huge.
I mean, I think part of it part of my reason for I mean, I started dabbling, probably around 2015, with what was called document TV back in the day, and now it's Cosmos DB on Azure, because of I had my old product was all SQL Server, Microsoft SQL Server based, and having to release new product and do database migrations were on prem software, so like, we would have to run it on their cluster, and it was their managed store SQL servers. And it's just such a pain in the butt, and, and so unreliable. And so that's why I was like, Look, I want to really learn no SQL and this whole JSON, I mean, Doc database model, and started to really like the, okay, I mean, you're going to do something, essentially, it's a patch methodology of I want to like patch this JSON into this other JSON, I think, and I was just like, Wow, this is so much easier to use in a lot of cases. And then I kind of learned the graph world a couple years later, and kind of came up with this, like this hybrid of that. But I mean, we're still we started looking at doing some database, migrate or not, I should say, database enrichment, we don't have to do database migration, because essentially, the schema is still pretty generic. I mean, we have a know, maybe a dozen entity types in the graph. But the majority are really like, there's four, or five, or six and get used a lot. And they're somewhat generic in the sense of, I don't anticipate doing ever really, truly migrate those, like, those are never gonna go away. But maybe we invent new entity types in the future. And we just add to the graph, so it's more of a aggregation model than kind of a transformation.
When the differences in in graph, the data becomes the model, not the other way around, versus having to create the model and back the data into it. You have, you know, like I said, the relationships were suddenly you see, it's like the Six Degrees of Kevin Bacon would have been a lot easier if we had graph back in the day.
It's really, I mean, sometimes the graph surprise you and that's where we're actually going to be adding in a graph visualizer tool into the application. Maybe not, I mean, it's it's more about data exploration, initially, maybe it is even more manual. It's like, Hey, you can navigate the universe of your data. And I think I've, I've seen that even when I was doing this podcast thing as you learn things that weren't obvious. I mean, you would never have thought to query them. But your eyeball would be like, wait, why is everything centering on this one node here? That's weird. And I actually mean, visually caught bugs, just because I was like, I could see these all these almost satellites around my data that weren't connected anything. And I was like, That's odd. And like, build data quality algorithms to figure that out. In the tooling, or even my unit to my testing, was it was actually way easier just to visually see the problem and go, Oh, okay, now I know what's, what's broken there. I'm missing edges between these things. But to like, I would have never thought to query for it necessarily, like I would have had to write data validation, just assuming the code was working because it was working in other cases. So that's where I
write in the writing heroic level SQL queries to try and hunt down data across the structure right?
So that's where I think that gets really cool. And I'm really excited about building that out in the next couple of months. We're just getting, get everything else ready. And then we're going to put this in, it's kind of almost a dashboard view of here is your universe of your data for exploration? And then I think there's, I mean, we'll just keep building on that and leveraging that for other for other stuff, too. And, and then really, our futures then becomes all about data integration and data collaboration of, hey, can your organization communicate over this data, triage it, prioritize it, and then use it for other things. That's, that's really what I see is kind of the next step. I mean, that's where things get really interesting.
Yeah, and it's, I like that you can look at it because we ultimately always sits from both sides, because there's the pure sort of the data, the human exploration of the data and understanding a business context or a functional context that they want to apply. And they can visually explore and walk this path of the data in a way that you may not have understood. And then on the bottom side, of course, you can then do inference, and then you can apply other now you can bring machine learning into it. And you may find interesting edge cases that are now real core use cases from either end. But not neither one is pure. It's not that the human exploration of it is wrong or bad. It's quite often that's the only way we do it, right? We it wasn't a UI versus UX, right? The classic like a square lawn with a cut path or diagonal across it, right? Like it's, we don't necessarily understand what the behavior will be until you observe it. Right? Right. And x, allowing an explorable interface and seeing people then explore the data, which is what graph unlocks in a way that we couldn't have when we had to, purposefully builds the structure in the query. It's a, I'm excited, I'm with you, I, I cannot wait to see, you know, the sort of next phases as what you're doing get unpacked on to the world,
that's gonna be expanding and having more eyes on it. I mean, I know that the team that's building right now you got that kind of like, feature blindness a little bit, because you've been staring at the same view for so long. And it's we've started to roll it out, like we just had a bug bash with our sales team this week. And they're using it in a very different way than I've ever used it. And it's just as you get it into more eyeballs, we go, oh, man, I never even thought of that. And that's the part. I mean, it like, I guess, maybe some people are annoyed by that. But I love that part. Because it's like, I know that I'm not gonna be able to see it. And I'm getting blind to it, because I look at it every day. But I'm super excited to get it in the hands of other people. Because I mean, we're going to get great feedback. And we're right on the cusp of that right now.
With the I probably anger my team, but there's a weird combination of just like, man, like, they just kind of get mad when you discover stuff. But they're also super happy. Because, as you said, you get this thing where you're locked into, like they know the use cases like test driven development effectively, they said that they only build in the tests that they know of. And so I'll come in and they'll be like, yeah, look at this, they could show you this really neat thing we're doing in like in graph exploration. And I'm like, Well, I can pin all these nodes over here. Like that's because that's the way you've shown the demo said, but I can't get rid of them. Like the first thing you do now is you then take this to time series visualization. Like what if I wanted to, like change the nodes that I've pinned over here? And they're like, ah, like, as they they just thought, like, I know, the use case, I'm going to test and it works in this beautiful workflow. Unless you need to back out somewhere in the middle, and then you're stuck.
It's so true. It's so true. I mean, it's a classic thing, but it's, I mean, I think that's where I've been, I mean, having been through multiple product releases and all this kind of stuff before it's like, I don't, I don't it doesn't stress me out. But I think for the some of the more junior developers or other people that haven't been through that, like, it's like you expect to get it day one, correct? Correct. And my expectations are a lot lower, I guess where I know, I know, we're gonna like iterate. And we also we we've been intentionally give ourselves time to iterate. Yeah, that's where you can get burned if like, if we just try and power through the rest of this year, and give ourselves no oxygen to just like, that's how we're gonna fail. So it's a Yeah, so we're trying to be pragmatic about them.
What and this really is, given that you've been, you've been a founder before, you've understands the merger of like the engineering impact of changes that you make in the way the platform works? how did how do you define like, all say, from your previous experiences, almost every founder tells me what they thought was the MVP was way beyond the MVP. It's really hard to know you're right. Everybody in hindsight goes. Yeah, we waited too long.
And it's it's I mean, we're we've had those debates like we, I mean, we haven't gone that long. I mean, what it's about for four or five years.
Yeah, I'm talking as if I've, I've had you going for four years, you've actually you're in pretty MVP, but you're very You very evolved in the MVP, like you are pretty far along on.
We and it is a prep ma'am. I'm trying not to overdo it, I catch myself like I'm there. I mean, sometimes just getting in the weeds a little too much, and I gotta just shut up and just be like, okay, let's get this out the door. And it's kind of my message to myself this week is, I mean, look, I mean, remember, this is the first version, like, it's never gonna be perfect. And it's over. I mean, it's, you just get into, like, starting to give commentary about the product, but it's like, No, no, you gotta wait for customers. Like, we got to like, we've been pretty opinionated from our initial customer discussions, but we don't have the customer usage discussions, you know, the customer feedback yet. And so that's where I think we just got to, and we put a little bit more in the box. And we were originally for MVP. So we decided like, I mean, there were a couple feature areas we're like, like, I mean, we just, we got to just get that in the box first, and not wait. And so, I mean, hopefully it's the right answer. I mean, I think, I mean, from a funding perspective, we're fine, we have the time, but I don't want to miss the window with customers either. So we're right now and given that we're we're in July, right now that our focus is kind of right after Labor Day, like people get back from vacation, let's just be completely ready for whatever we're going to get by then. And hit it hard. So we got from Labor Day to Christmas, is our focal point for this year. And then we just want to, it's not even necessarily about revenue, it's just about feedback. And and writing. Getting people I mean, using it and and proving some of our thesis points. And so yeah, I'm excited. I mean, it's just we're, we're about six, eight weeks away from that or, or so. And, I mean, we have got a bunch of people that are ready to roll with it. And now, now the rubber meets the road.
Yeah, yeah, no, it's it's such a fantastic, I'm incredible respect. Kirk, you've like everything I think of like a throw at you. You very, very thought you've thought this whole process through in a lot of ways that when I talked to founders and Korean builders, it's very, they there's always stuff missing, you've, you've done a really fantastic job of understanding every edge, you talked about the fact that you built onboarding as part of the early part of it, also a huge area where people like, oh, let's get the product to work, not realizing if you've got a a friction, a high friction, onboarding, then your sales team is going to be in real grief as they go from first customer to second customer. And I really, your experience and understanding of the customer side and the customer led approach is it's coming through in the way
you put it together. What's interesting to me is Yeah, that was that was essentially a company I was trying to start for just the onboarding. And I was was sort of like a, I mean, if you look at like Twilio or companies like that, they're just very API driven companies like I wanted to have something like that. And essentially, I was building an API to do that. grungy onboarding, like tie billing together with provisioning together with authentication and all that, because that was my initial pieces is like, that's the part that everybody has to build. I mean, even if you're using auth, zero, even if you're using stripe, or even whatever, just tying all that up into a nice epi bow would have us. And I mean, maybe I was a little early it was it was kind of hard thing to sell by itself. And also, I didn't have like a business side co founder. And so but what I did is I essentially took the IP, and just rolled it into everything else I did. And so it's like, I use it for my own projects. So it was a good, I mean, it didn't go away. It's still in the box right now, even in unstruck. But it's funny, I mean, that's basically I took all that IP that I built over the years and that became the starting point for what we did. So it I mean, it the startup didn't pan out, but the code ended up ended up living on for a long time. So
quite often, I mean, that's that's the, I'll say, the unsexy Part A lot of this plumbing that you know, quite often and I mean, look at how many how many companies can we talk about, you know, we could go through a litany of of different examples where they, they went to market with one thing and that thing actually was the thing that led to the thing, right? Yeah. Or often they built you know, was, you know, cloud was one of my favorite, you know, things like cloud is trying to build this incredible pass to battle the cloud foundries in the Heroku is of the world and they had to build this container construct while they were at it and they called it Docker and well, that's the ceiling no one remembers cloud
classic, what slack was a video game or something? Yeah,
yeah, like audio. podcasts and transformation. And, and but those, again, like it's, you've solved, you knew this problem needed to be solved and bringing that through to the next thing. It's a beautiful opportunity to come together. So I'm excited for what's ahead for you, Kirk.
I appreciate Yeah, I think it was also when I had my company, it was right at the cusp of cloud native software. And I really wanted to do something cloud native. And so I started to rebuild everything that I had done on prem, in a cloud native world. And I learned a ton. And I think it was part of it was just for my own, like knowledge. And part of it was like, wow, there's actually interesting needs here. I mean, to bounce out. So yeah, I'm excited about I mean, I think we're got a great team. I mean, it's people I most of my noon before, and we're all pulling in same direction, which is really important. And it I mean, it's Yeah, even, I mean, we're gonna get into customers hands. So essentially, within six months of funding, which I think is pretty rare, and I mean, as long as, as long as we're not wrong in in that there's a need here, which, I mean, it's funny, I mean, from an investor side, it's been super positive talking to investors. But now my biggest thing is we got to make sure that customers love it as well. And so we're not just selling this thing to investors at this point. So we will prove it out soon.
We are in an opportune time for a lot of the things when you've got this sort of readiness of product that you've got access to funding despite really challenging time in the world we've had there is a surprising amount of capital that's available but looking for stuff that can get to market faster so again, I'm I'm I'm bullish on on where we're unstruck is going to be in the coming months. And on top of this, you've got a career as an advisor, and you got a book in you, if you can take all these lessons, and you forget about the Andreessen Horowitz folks, you, you've done this for real as well, and you've done it in a, you've done it with humility, and it's again,
could get it we're gonna break for long enough, I mean, that that wouldn't be a bad thing. So because
I know you're like you, I can't imagine that the moment you have an hour of free time you find four hours of work to cram into it.
It's mostly it's mostly like that, but it's got the downtime to kind of Germany to new ideas. And, and it also just, I mean, just the, I can't do everything. So it's like I'm having to sort of take a breath, like the team, like, finish their work. And then I mean, my parts, I mean, the majority of it's done for what we need to do. And so now I just gotta like, focus on all the other stuff, marketing and sales and all that because there's a lot of other pieces to the equation to not just writing codes.
Amen. Well, I tell you, I could do a whole we could do a whole podcast on that alone. I enjoy your time, Kurt, thank you very much for folks that did want to get a hold of you and get in contact. Of course, I'll have links to unstruck data. And I imagine that people will be seeing press releases coming out before too long for a variety of reasons, which is great. But what's the best way if people wanted to get connected? Yeah. I mean,
LinkedIn is probably placed up on the most so just, I think I'm on the only Kirk Marple there and it's also on Twitter, just at Kirk marble, and then at unstruck on Twitter, those are the main the main places you can find us.
And of course, unstruck a un str UK so for folks that are accidentally typing unstrapped with a C you'll you'll find that you're you're heading to the wrong place, but now very, very cool. I'm excited. So there you go. Kurt, Marple unstruck data. Fantastic. Thanks very much.
