Developing Strategic Roadmaps for Technology Products and Services

Wade Erickson (00:00):

Welcome to another episode of Tech Leaders Unplugged. I'm your host today, Wade Erickson again, and I have, Don Gaspar of Gigantor Technologies. And, we'll be talking today about developing strategic roadmaps for technology, products, and services. Utopic near and dear to me, as a product developer and product manager, built a lot of these over the years and, there is definitely an art to it. And we're going to get to get unplugged here with Don about, product roadmaps. So, Don, thank you for joining. Appreciate you spending time with us in our community. And, yeah. So tell me a little bit about yourself and then let's talk about Gigantor Technologies. You're the CEO and the founder, so tell us a little bit about what, how you came up with the name, and what prompted you to start this company.

Don Gaspar (01:08):

Great. well, thank you for having me first. It's, it's great to be here. Anyway, Gigantor is a, a company, and it's been around for about four years now. And what really prompted us to form the company was we noticed when we looked at neural networks, they were moving too slow out there. And so we saw computer vision, especially deep neural networks, were all operating very slowly. 30 frames per second is impossible with HD for anything out there. And so we looked at the solution, figured out it was architecture, and everyone was using a 70-year-old Von Neumann architecture. And so what we did was my, my business partner, Mark Matthews, Mark is actually Microsoft's 17th employee, and Mark actually came up with a math proof. We looked at every research paper out there, every patent filing every patent. We went back and couldn't figure out why no one else was doing this. And we formed the company. The name was actually a domain name that I had for about 20 years or longer, maybe close to 30. There was an old anime show in the sixties called Gigantor, and it was a Japanese show called tsuji 20 something. It's about a giant robot that had an AI and I had the domain name and we've had it for all these years, and it seemed like a perfect fit for an AI acceleration company. And so we set up the company to solve several problems. The big one was latency, which is the response time for an AI that can save lives. We did massive acceleration of the process. We're input bound. We don't use any memory, which is a big trick. People are going to ask questions about that. And we're hitting the maximum possible throughput if we're input bound. And our power reduction is about 90%. So we've got a new architecture that's doing pretty well. Before that, I've been an executive for about 20 years. I've been part of, and, and fortunate to be part of two IPOs in there. And so I've been a CTO or executive vice president, typically running engineering and product organizations, both. And education wise, I, I've got a master's in distributed systems, which is computer science, but focuses on servers and scaling systems. I have an MBA from Pepperdine University, and I have an undergraduate degree in solid state physics. If you went way back in, in time, I was fortunate enough when I had hair to work at a little company called Apple Computer in Silicon Valley, and I learned most of what I know and still carry with me today. Back during that time period. It was just a big influence on my, my life, my family and my career. And I was really blessed to be part of that.

Wade Erickson (03:47):

Did you know John Couch while you were there?

Don Gaspar (03:50):

I did not.

Wade Erickson (03:51):

Okay. Yeah, I, I just read a book of his rewiring education this weekend, and he was employee number I think 54. And he's, he was the vice president of education there and stuff like that. Really, really good book. Anybody that's into modern ideas about education, that's the book to read. It's awesome.

Don Gaspar (04:11):

He's before my time. Okay. So I worked on the Finder, the desktop A UX, Apple's UX. Okay. And then I, I got fortunate enough to work when we moved over to Power PC to work on the proof of concept before it was a chip and give a demo at the San Jose Convention Center, which was just an absolute blast. So Labs and IBM and Motorola.

Wade Erickson (04:35):

Awesome. Awesome. Alright, so our topic today, developing strategic roadmaps for technology, products and services. What interested you about sharing this topic today? And tell me a little bit about some of your insights on building product roadmaps and some of the challenges and some of the, you know, little nuggets that you might have for that.

Don Gaspar (04:59):

Sure. Well, the big thing is you need to build something that's not out there. I mean, why do another me too? I mean, does the world need a duplicate product from someone else? And you hear this all the time and VCs get inundated with this. They get people with the same ideas coming in and somehow they're going to do it differently and it's a mobile app or it's another SaaS, but deep down they're the same thing. And so typically what, what I've done, at least part of my groups is we do a competitive analysis of what's out there already and we compare what we have to it and we list all the major features out. And so you do a matrix with this and you look to see where you have parity and where you're going to have advantage. You have to have differentiation and it has to be of value to a customer or they're not going to buy it. And so you find the features that you can put out there and you build those, and you have to look at what the ROI is going to be of course. But if you build something that people value, they will come and they will take it and they will use it. And it has to be different enough. The other part of it's, and I found this extremely valuable, when you work on something, you have to make barriers of entry for other folks. You can't do something that someone else can just copy right out. And so, you know, in the case of a mobile app, if you're doing a new mobile app, it has to be something different and unique. Something you can patent is even better if you can protect it. So what we've done here, you know, for example, we do massive convolutional neural network acceleration or deep neural network acceleration. And what we've done is we're doing it orthogonal to everyone else. So we're not a TPU, we're not Von Neumann, we're very different. But to get this and to put some entry barriers, we have multiple patents. So we have six patents so far issued on it. This is just one example. Some are 80 pages with 30 plus claims on each one. And so they're very broad and they're set up as landmines for someone that tries to duplicate it or get into this area. When you're doing your own product, you want to do the same thing. And so you want to put barriers of entry for competitors. They may not be patents and some cases might be a difficult challenge that you're solving and you're doing it with some trade secrets or in a unique way. And so you definitely want to have that differentiation and that entry barrier. Another one, if you can, while you're building your roadmap, think of the exit barrier. So how do you get a customer? I mean, there's two exit barriers. How do you get them off a competing product if you're trying to solicit their business to come to yours? So there's an exit. You want the exit from your competitors to be very easy to your product. Likewise, you want to create an exit barrier for your own product to make it difficult. You don't want it impossible for someone to leave, but you want like fly paper. You want them stuck to you because your product is so good, it's so unique, it's so differentiated, and you're treating the customer really nice. You're getting their feedback and you incorporate into the product you incorporate in the roadmap where it makes sense. You know, in our case, we talked to the Pentagon and folks in the Pentagon and the military. They've had a lot of valuable input, they've had some insight to our product that we hadn't even thought of. And it's been extremely priceless really. And it's leading us to other contracts now with some of the prime defense contractors. But very, very fortunate if you listen to your good customers like that and keep them happy,

Wade Erickson (08:20):

I like that point where you talked about patents and trade secrets, is that I think probably best is to have a hybrid, right? Where you patent some of the core features that are maybe architecture in nature. But when it comes down to really execution, you keep those as trade secrets to, to really make your product be something that is not easily copied because you have not revealed all the, the good stuff in the patent. And so you can still use a build a moat around the castle, but don't let anybody in the castle.

Don Gaspar (08:52):

Absolutely. You're, you're right on Wade. And you know, like the patents, the patents will tell what we do, but it doesn't tell how we do it. And so there'll be some math and a smart person will figure it out when they go through 80 pages for each one, they'll be able to figure it out, but to duplicate it and it's a bigger effort because those are trade secrets. Sure. And so it makes it hard. But we had to have the patents because we are so different that no one would talk to us at all. They would say, well why didn't Nvidia do this? And I don't know why they didn't do it, but, and I don't know why anyone else didn't do it. The fact is we were the ones that came up with it and did it, but it does protect you this way. And you couldn't talk about your technology and how you do it without having the patents in place to protect you. cause Otherwise people will take your idea. They'll go out and patent it really quick and file it. So it is just the nature of it. I remember being at a Pentagon forum last year on talking and right when we published our numbers, we had someone from one of the largest AI companies in the, in the country that was part of the panel. They all left the room. I'm sure they were looking, our panels are looking up our patents. I'm pretty positive and trying to figure out why our numbers were much, much better than theirs. So it happens. You've got to protect yourself.

Wade Erickson (10:10):

Alright, I wanted to jump into a few questions if that Okay, sure. So, so as we talk about roadmaps how do you prioritize the feature development of the product roadmaps and, you know, what factors play a role in the decision making? You know, how should the why fit in there, you know the vision, the missions, the values, all of those kind of things I've kind of, you know, seen out there that need to feed into the product roadmap. Can you tell me a little bit about, you know, how you build out your feature sets and your product roadmaps to drive your development?

Don Gaspar (10:45):

Yeah, absolutely. So I think the first one is you get defects and that happens and that needs to be factored in and you need a percentage of your time to look at that. And so if you do have a defect that is bothering a client or it's interfering with any of your processes, you need to fix that and get that out quickly. Stop what you're doing, get that out, get it released, keep your customers happy. The other one is you get your customer feedback, but also your roadmap is going to follow what your vision is. So ours is, we do, you know, to do massive AI acceleration, you know, for neural networks, reducing latency and power and trying to remove barriers of entry for people doing networks and deploying them. And so for us it has to fit in with that mission and that's what we're doing. And so some of it is it comes from clients, others, it comes from us doing research and looking at it and talking to folks and seeing where it fits in. Quite frankly, the military has given us some, some good input on it and we prioritize sometimes based on ROI. If we have something that's new and we know it's going to help sell new products or new services, then we'll prioritize that. And, but not all the time is it ROI, you got to keep your clients happy, which will come back to benefit you I guess in long term ROI. But you know, you have to balance the long and short term. Anyway, it's very fun to do this and just very fortunate to, to have been doing this for so long.

Wade Erickson (12:11):

Great. Thanks for that. There was a question that came in real quick. We'll, we'll take that before we jump in. They were talking about the patents being valid outside the US I know good patent attorneys submit 'em both. They submit 'em to the US patent office and the global major western countries at the same time. And there's been times, and I have a patent that it got more traction outside of the country than it did in the us So sometimes you can get better protection outside of the US than the US So, so the question the person posed Yeah, if you could have got a good patent attorney, they'll submit 'em in parallel. You don't have to wait for one versus the other. So

Don Gaspar (12:47):

Yeah. And typically what we've done is we wait until it gets approval in the US and we still have a timeline for protection and then we file it. 'cause You have a certain amount of time before you can file it and you certainly want protection. And Korea is the big one. Japan, all that area. Europe is the other area. And a lot of the other countries will follow these along with the us. China won't follow anything. So it doesn't matter to file there. They're not going to file it. They're going to copy it if they want it regardless and not pay royalties. But you need to do that. You have about a year, I think from the time your patent is issued to go and do that. And it's not that expensive. If you know how to do most of the paperwork yourself, you can get it out for about $40,000. And cover most of the world.

Wade Erickson (13:32):

Okay. next one ahead. What strategies do you employ on gathering and prioritizing the customer feedback? We talked about that a little bit, you know, on during the iteration, whether it's defects or ideas and then, you know you know, who is the audience for the product roadmap and the stakeholders providing inputs, whether it's outside, inside and who, who really is the audience for the product roadmap? 'cause They're difficult to build, sometimes people will follow them.

Don Gaspar (14:00):

Wow. Well that can really vary. It really depends on the company and what you're building. But, but yeah. Well we look at barriers. I mean, that's one, we want to know why someone else isn't doing this. You can run into things known as tar pit technologies or tar pit features. People aren't doing it because, well, they've probably done it before and it led them into a tar pit that costs a lot of money and it took a lot of money to get out of it and they ended up abandoning the effort. And so you've got to watch for technologies like that. And so you look at what you're doing, you try to estimate it, scope it, see where it's valuable, look at the potential revenue or new clients. We take customer feedback if it's a large client that may have more importance than someone else. So we waited a little heavier. So if the army had asked for something and they're working with us, we would bend over backwards to get what they need immediately and, and help assuage any concerns there. So in terms of who the audience is, certainly parts of it would be to your clients. Not all of it's to your clients. You need to keep something secret while you're doing it until you release it so someone else doesn't go and copy it. But certainly internally you have product managers. Engineering is always a big part of this. They have to estimate it and scope this and look at the efforts and also minimize the defects and the probability of defects as it's coming out. And the complexity of maintaining the code or whatever you write. In our case, we do a lot of hardware. So we take neural networks, we make them into circuits that run on FPGAs or we make a bespoke asic. And so a mistake in one of these could cost millions of dollars for us. And so we have to be very careful the CFO or whoever your main finance person is, should be involved in larger projects. So the budget is very important. You've got to track every penny that the investors and your customers are paying, you know, into the company to keep it running. And so you have to factor all of that in. Product certainly executive leadership at times. Sometimes sales will have some input. You need to be very careful there. You can't get in the trap where sales guys say that they can't sell because they need feature X and, and feature X isn't going to sell feature X is not going to do anything. It can be, maybe it's something to give you parody with a minor feature that a competitor has. You can put that down the road. And so the important ones are going to help the bottom line long term and short term. You have to balance 'em. It's it's more of a art, I think than than a science in some cases.

Wade Erickson (16:39):

Great. so next question is kind of like we just a little touched a little bit about that. You know, what, when you're building your product roadmap, how do you kind of balance and align both the short term business objectives and some of the long-term goals that you're really trying to achieve as an overall business in the product line you building?

Don Gaspar (16:58):

That really varies. So we have some long-term goals here, for example, where we're making a universal computer vision chip. So basically a chip that's going to do, you know, any convolutional neural network on a standard fab that's done in the us. And so that's a two to three year effort out. And so right now we've got another product and we put things in the circuit form and make bespoke asics, which we're focused on. And so we have to balance that need. So there is that short term need that brings revenue in. As a startup, that's very important to have, in fact, than any company. It's very important to have, but even more so in a startup because people are looking for validation when they invest. They're looking for references when they want to come and use your product and your services. And so it, it's a tough balance. So we typically will follow the revenue or a big client that gives us some name recognition. We'll typically follow that while we're simultaneously doing the long term, but it can't interrupt the long term manufacturing of the asic for example. We're still designing, laying things out. We just have to staff and be ready for that and balance it. There's no, there's no secret to it. It's, it's a tough, it's a very tough thing to do.

Wade Erickson (18:13):

Great. Great. so can you share a specific instance where your team, you know, had to pivot whether it was to adapt product strategy based on the market inputs, which you've talked about with the Army and other outside customers that really drove some of the, the features that, you know, you got to pay the bills, right. While you're working on the, the big things, you know, or any other changing industry dynamics. So maybe, maybe talk a little bit about, you know, we've talked about the military and DOD applications, but can you share a little bit in the, the private sector some areas that might benefit? Is it banking, is it finance trading? What, what other areas could benefit?

Don Gaspar (18:51):

Sure. I think probably a good example is a, a while back I was the CTO for a payments company and a great company that had a high growth trajectory. And I was just really, really honored to be a part of that for, for several years. We got a client that came in VRBO. So we had this client that was a hundred times the size of our company. We were really blessed. Their COO Brett Bellum, he runs a very large company, I think BigCommerce right now. But he liked what we had. He wanted some new things that we didn't have. And we followed because it was such a huge account, it went, we went from eight or 9 million in revenue to over a hundred million. But it required a pivot. So the pivot was instead of just doing payments and doing paper processing of merchants, we had to automate the process for the merchants and be able to approve thousands and tens of thousands small merchants, homeowners basically, that were renting their places out. And so we had to invent a system that would automate this and do it in real time. We had three seconds to do it basically. And then they also wanted an API, which was really helped us as a company. So we had an internal API and what we did was we exposed the API, but we adapted the API based on a lot of feedback they had for their needs. And this effectively weaponized an API for all payments out there. Open new markets for us based on the customer feedback. And it gave us this huge growth trajectory when I was there from, like I said, it was eight or 9 million a year. I, I think we went to well over a hundred while I was still there and it kept climbing when I left. And so it was just listening to the client, getting their feedback. And that pivot was from being just a payment service provider to being a technology company, developing instant underwriting, and also making a really high performance API that can handle thousands and thousands of transactions. Anyway, listen to the client. Some of them really know what they're doing. Some of 'em know your business better than you might know it yourself. In this case, Brett knew it much better than we did. Brett had been the president of PayPal Europe, so he's definitely someone to listen to.

Wade Erickson (21:08):

Makes sense. Basically,

Don Gaspar (21:09):

He, he points in a direction we, we followed and it really paid off.

Wade Erickson (21:14):

Excellent. Well, this is the part of the show. I like to pivot a little bit and talk about you. You know, I noticed you know, like you said in your resume, you've had a lot of senior positions at big companies and, you know, iconic companies over the years. And you decided to go out on your own about a few, four years ago. And a lot of the folks watching right now are, you know, maybe in that position where they've been working 10, 15 or five years and they have something that they want to go do. Can you talk a little bit about what motivated you to jump outside? I know you said you had a couple IPOs, so maybe you have some money in the bank, I don't know. But you know, what was really, I think it boils down to courage and confidence, but I want to talk, have you tell what was it that really convinced you that you could make a run at this?

Don Gaspar (22:10):

Yeah. well, well thanks for asking that. I was getting, you know, our last company got bought by Warburg Pinkus, so it wasn't an IPO, but there, there was an exit and, you know, I've been really, really fortunate, but we had a great idea, you know, there was a big problem out there. The value of AI and deep learning in general is expected to be twice of what the internet value creation has been over the next 10 years. So the internet created about 18 trillion in value according to the ARC organization from Kathy, Kathy Woods report. And deep learning is expected to be more than 30 trillion in value creation over the next decade. And so we saw this and we saw that there were some major problems slowing it down and slowing the adoption. And the big one is it was just moving too slow. And so for me, I was looking at retiring or, you know, and I was interviewing with some executive retained search for a couple positions and Mark had reached out to me, Mark Matthews, my, my business partner. And we've known each other for about 25 years. We've worked together, we've coded together even for a place where to before Microsoft acquired, then we had to convert a product. And so we knew each other pretty well. Mark had some ideas and we went through it and we spent four months to research the whole market. And every research paper, every patent, every patent application, everyone was doing it differently than what market had proposed initially. The math all worked and we went and we built it and we tried it and we know there's a need for it. And so there's naysayers out there. You need to tune them out, you will get them. I think the more naysayers you get, the better of a product you probably have, especially if they're, you know, quote unquote investors. The smart ones will get it, they'll investigate it. The, the, the ones that aren't so bright will go, why didn't Nvidia do this? Or Why didn't someone else do this? And they don't understand it and they don't take the time to try to understand what you're doing and why it's different. So for me it was a no brainer. It was, this is an opportunity of a lifetime and we need to run with it as fast as we can and get it out there.

Wade Erickson (24:31):

Thanks. Thanks for that personal insight there on taking, taking on the, the competitors and non-com competitors and maybe even jumping out there with something that's new. Well we're kind of at the top of the hour with the show. Is there anything else you'd like to share before I kind of announce next week's guest and wrap it up for the day?

Wade Erickson (24:55):

Yeah, well thank you Wade. Yeah, have fun with what you're doing and don't get discouraged by people if they tell you no or nay. Go out there and research it yourself and look, and deep down, you'll, you'll see the answer. If there's a market for what you're doing, if it's differentiated, it's something that's difficult for other people to reproduce, you know, go pursue it and, and have fun. That's the important thing. I've been paid for 35 years or more, 40 years in a career. I've been paid to be in a toy store. Have fun with what you're doing, build, create, make a difference.

Wade Erickson (25:28):

Great advice, great advice. Appreciate that. Alright, well before we wrap, I want to bring up next week's guest is Kevin Fray Field, CTO of Mark 43. It is on Wednesday, February 14th. Valentines Day, times day. Go out and get those cards for those lovely partners of yours. We'll be talking about the 21st century public safety platform, empowering communities and their governments with new technologies that improve the safety and quality of life for all. So that's next week on Valentine's Day. Same time, of course. Nine 30 Pacific time. Alright, well thank you so much again, Don, for spending a little bit of time with us, sharing your background and of course your very exciting technology. I could probably talk to you for hours just about this stuff and but we don't have that time maybe after, after the show. But again, appreciate your time and we we'll see everybody next week and have a great day. Don stick around, we'll talk a little bit after the show.

Don Gaspar (26:38):

Great. Thank you.