In this episode of Beyond SaaS, Jason Niedle interviews Jim Kernan, Chief Revenue Officer of Luxonis, discussing the intersection of technology growth and robotics. Jim shares insights from his experience at the NYSE during a significant market downturn, emphasizing the importance of feedback loops in business operations. He explains Luxonis’s innovative approach to computer vision and the challenges of long sales cycles in the tech industry. The conversation also covers inbound marketing strategies, the difficulties of scaling for enterprise deals, and future trends in automation and robotics.
Takeaways
- Jim Kernan emphasizes the importance of feedback loops in business.
- The Fifth Discipline by Peter Senge is recommended for understanding systems thinking.
- Luxonis specializes in edge inference computer vision hardware.
- Sales cycles for Luxonis can range from 6 to 18 months.
- The company has a robust pipeline that helps predict revenue despite long sales cycles.
- Inbound marketing and SEO play a crucial role in Luxonis’s growth.
- The transition from B2C to B2B was a natural evolution for Luxonis.
- Challenges in scaling include servicing enterprise deals effectively.
- Automation and robotics are set to become more prevalent in various industries.
- The agricultural sector is seeing significant advancements in automation.
Sound Bites
“We’re constantly trying to get feedback from the market.”
“How do we put vision into our products?”
“It’s really wild, the things that people are doing.”
BeyondSaaS Transcript
Jason Niedle (00:00)
Today we’re talking with Jim Kernan, Chief Revenue Officer of Luxonis about tech growth, of course, and maybe we’ll also hear a story about his recent interview on the NYSE’s floor as the market was crashing.
Welcome to Beyond SaaS. I’m Jason Niedle, founder of Tethos. We solve growth problems for tech companies through strategy and design, cold email, educational campaigns, and much more. We’ve taken what we’ve learned from all these amazing leaders, and turned that into a hypergrowth playbook. You can grab that at tethos.com/podcast or drop the word growth, and I’ll DM it to you.
Today I’m really excited to talk to Jim. Jim is the Chief Revenue Officer at Luxonis. And they make robotic vision easy and accessible. And they do this through their Oak cameras. And they also do it through their cloud solutions like fleet management and Vision AI platform. And I’m curious to hear about how both of those are working. Jim is an experienced entrepreneur and growth catalyst with over 20 years driving nine figure revenue growth.
And I do want to hear that story about your recent interview on the stock exchange floor during a massive 6% sell-off. So welcome.
Jim (01:05)
Thank you. Appreciate it. Yeah, it was an interesting day. were part of a robotics conference that Citi was hosting. And part of it, they got us on a podcast with NYSE Wired, which is a great experience. I live right down the street and I literally walk by there probably once a week, but I’d never been inside the exchange. And so it was a cool experience, a lot of history in the building.
the trading floor, the options trading floor is still active and live. And it’s obviously there’s still people sitting at terminals and computer screens, but there was a lot going on that day, was the day after “Liberation Day”. So the markets were down quite heavily and the studio itself is a perch up over the top, so it was just a lot of noise, it was interesting trying to keep yourself contained and focused on the person who’s in front of you when there’s so much clamor behind you.
Jason Niedle (01:52)
What did it feel like? What was the vibe in there?
Jim (01:54)
You know, back in a former life, I traded structured derivatives at Barclays. And so I’ve been on trading floors before, it’s not a brand new experience to me by any means. But you know, just when you can hear people yelling and
and screaming and barking out orders to each other. You know that there’s something active because that only happens usually at the most dramatic moments. yeah, it just added to the ambiance. It was kind of like a stadium seating while you’re giving a talk or presentation. So it was interesting.
Jason Niedle (02:29)
It’s like saying you’ve never been to a football game and then they throw you in the middle of the world of the Super Bowl or something crazy and you’re like, what just happened? Yeah.
Jim (02:35)
Yeah, exactly. mean, yeah,
you don’t get you don’t get many days. I think that that day actually was the largest nominal point drop in the history of the stock exchange just because of the levels that we’re at, there have been bigger percentage drops, but nominally that was the largest one. So there aren’t too many tastes that you’re going to get where there’s much more activity.
Jason Niedle (02:55)
Amazing. Do have a quick tip for listeners? A golden nugget?
Jim (02:58)
⁓
There’s a book that I love. It’s called The Fifth Discipline. The author is Peter Senge. I might be pronouncing that incorrectly, but I found it to be really, really helpful in just thinking about organizations and how the different components work together. It really teaches you about systems thinking.
And how everything is interconnected and there are multiple different systems like anything that we’re working on. You know, we kind of go through life in this linear fashion. Or we tend to think of things as individual tasks with a set of steps. But realistically, there’s a feedback loop that is being given at all points in time in even the most mundane tasks. One of the early examples that they use is just filling up a glass of water. You’d say, okay, what do you do to fill up a glass of water? You put it under the faucet, you turn on the faucet, and then you
you turn it off, right? That’s it, those are the steps. But there’s this constant feedback loop where you’re evaluating the level of water that is in the glass versus where you want it to be. And you’re evaluating that at constant times. if, for example, the faucet were to explode and have a lot more throughput all of a sudden, then you’d do something different. But it’s a good way to be thinking about your organization or any individual task is that
you’re going through, you’re constantly getting feedback and constantly learning and changing and evolving. It talks about how you can build organizations that think through problems in that fashion and help the teams that are cross-disciplined be able to work together.
Jason Niedle (04:19)
The feedback loop has been a theme for some reason on the last three podcasts that I’ve recorded because we kept talking about like ready, aim, fire, and a lot of people aim, aim, and you never get the feedback loop and exactly what you’re talking about with the water. Like even simple things. You know, in my house, it’s a really old house, so when you turn the hot water on, you listen and the sound changes when the water gets hot. It takes three minutes, right? And you’re like, oh, now I hear that it’s hot. I’m not even, you know, over there.
Jim (04:22)
Ha ha ha ha.
Hahaha
Jason Niedle (04:44)
But these little subtle cues, I think it’s such an important lesson for business of are we paying attention to those cues and are we changing? Because a lot of business are just ramrodding the same thing and letting the cup overflow if we’re using the metaphor.
Jim (04:56)
Yeah, absolutely. And from my perspective, where I sit within the organization trying to drive revenue forward, we’re constantly trying to get feedback from the market and understanding what they need and why they are coming to us to help them solve their problems and communicating that back with our engineering and product teams who are developing the new things that the companies are going to be able to use. And how do we get all of that to work together
to be able to produce something that is useful to the world.
Jason Niedle (05:24)
My last podcast guest was Robin Bonduelle of a company named Claap, CLAAP. And they have an AI that reduces all the bureaucracy and all the paperwork and all the stuff that happens in these sales calls through a number of different agents that are all coordinated.
And one of the things that they’re doing that’s really relevant to what you just said, and I think it’s so interesting, is in recording these calls, it comes back and it starts to give you a through line of like, hey, on the last 12 calls, you’ve been getting feedback that your pricing’s too expensive, or, and I’m like, that we’re just inches away from that being really, really relevant and available to all of us in a way that we never had those signals before.
Jim (05:58)
Yeah, I mean, that’s something in a bit of a parallel or a tangent, however you want to take it, I think about a lot is… the sales cycles for us are extremely long. Typically have them to where we get to meaty parts of the deal are six to 18 months. So because we’re selling hardware and people are typically in R&D when they’re buying our stuff. And then they’ll go through a proof of concept, maybe a pilot. then after some evaluation, they may choose us, they may not.
Jason Niedle (06:13)
Mm.
Jim (06:27)
And then they’ll go into full scale production. And so there are a couple of things that I’m trying to pull forward with those calls. Like one, how do you, as we have thousands of different customers, how do you keep track of all of that information across all of those different companies to try to have some prediction into what your revenue is going to be? Yeah, exactly. Over time. And then like their underlying themes, as you just mentioned.
Jason Niedle (06:38)
Mm-hmm.
And over time, right?
Jim (06:48)
you know, pricing is one impact, or is one thing, but like the feature requests that people have and the use cases that people have. And we have this blank canvas for how people can use our hardware, where literally we’re in dozens of different industries and there are dozens of different use cases where people are trying to solve with our product. And so there are core features that we want to be able to build to allow for as many of those things to work as best as they possibly can.
And how do we pull those threads from all of the different calls to surface relevant information to say, hey, product, actually really have a good idea. We have all of this data that suggests we should be building this feature to support our customers even better.
Jason Niedle (07:30)
Maybe you and Robin need to talk. So tell our listeners, because I read up on your company, but they probably haven’t. So tell our listeners about Luxonis.
Jim (07:32)
Potentially, yeah.
We make hardware. it’s edge inference computer vision hardware. So it’s AI enabled cameras is really what we do. So the eyes, ears and brains of robotics and automated systems is how we describe it very succinctly. And really what we do is we give machines the ability to perceive the world. We’re kind of an interface between the physical and the digital world. And so we work with robotics and automated systems, as I mentioned, and
again, allow them to process the world the same way that humans do process that that visual information and it’s starting to get a little bit easier to describe this stuff as people are aware of what LLMs are now and AI is permeated everything. Like ultimately in LLM is taking your prompt your input and saying what is the most likely response that this person wants based off of what I’ve fed into it and computer vision is basically.
Looking at the world and saying, what is this information that I’m looking at? And it’s looking at stuff in both 2D and 3D and all of the colors, pixels, the hues that are in that image to be able to draw some inference of what is in the scene. And so the models that are trained in LLMs are just like, give me a language response to what this is where
you models are trained to detect things in the real world. We’ve gone through our lives have processed billions of different images, like the rough numbers is we’ve processed about half a billion images every single year.
You have a lot of context and understanding of what a phone is or what headphones are, what a plant is, any of those things. You’ve had a feedback loop provide that to you. So like, if you have that model, you can put that directly on our camera and basically give a machine sight.
Jason Niedle (09:16)
It’s amazing and I think when you really think through how difficult that problem is or has been in the past, I’m looking at your screen, but a computer sees gray pixel, gray pixel, gray pixel, darker gray pixel, right? And it just sees all these little data points and then I look and I’m like, well, there’s Jim. And it’s such a fascinating problem to solve. What are you guys looking at? how, with sales cycles this long, how do you look at and predict growth?
Jim (09:26)
Yeah.
Yeah, exactly.
Yeah, it’s a difficult thing is all I can say. We’re fortunate with where we sit that we have a lot of people coming to us. We do next to no outbound efforts. I wish we had more time to be able to do that as part of why I’m trying to be more active and out in the world is allow people to find us. But we’re fortunate that we have a really great.
client base, like we truly found a need in the market. People have vision related problems where they have some automated system that they’re trying to or some robotic system that needs to be able to interact with the world. And so as they start to do research into how to solve that vision problem, like they come across Luxonis for the uniqueness is we’re really the only company in the world who’s able to process that visual
imagery directly on the device. And so, in going back to the question, how do we actually, you know, project growth? Like we really do a decent job like early on in our phone calls trying to extract like what does the R&D cycle, what does the proof of concept cycle, like what does production look like from the very, very beginning? And so, we break it down into these little wins along the way of, can we get that initial sale? Can we get
like the sale that is the initial proof of concept or the pilot phase, whatever that is, break it up into those small parts and try to do a good job of forecasting of when this stuff is going to happen. Again, the reason I mentioned that we’re fortunate to have people coming to us is we do have a robust pipeline. And so that allows us to withstand
the uncertainty that comes with the longevity of our sales cycle is, if you put enough stuff through it, then you’ll have stuff that comes out more regularly, even though it’s gonna take you 18 months. You if we started from scratch right now, it would be really, tough, because there’d be a dry period for some moments.
Jason Niedle (11:23)
Mm-hmm.
But obviously over time, you can start to measure, OK, good, I know I have X number of pilots going right now, so 30% of those are going to come back. And we roughly know what each of them is looking at in terms of numbers. And therefore, I have some predictability.
Jim (11:43)
Yeah, and there definitely is predictability in the sense people are constantly buying cameras for evaluation purposes. And that stuff is pretty steady. And we understand what the growth looks like when they’re going through each one of those phases. The stuff that we want to get better at predicting and forecasting are the big chunky deals, right? Like those that are nine figures and above.
when those are going to land. And ideally, we have those that are spread out evenly so we have consistent growth throughout periods of time. But yeah, mean, for sure, it’s a very real problem for us. If you think in the traditional SaaS sense, as soon as you have a large sale, everyone’s hitting the gong, ringing the bell, really excited. Hey, we just made a whole bunch of money. And that’s because you have
Jason Niedle (12:14)
Right, because then production becomes a problem, right? At some level.
Jim (12:32)
such high margins, there’s next to no cost to onboard an additional client outside of what you’re already spending. So you’re like, great, we have this influx of cashflow. Yeah.
Jason Niedle (12:38)
and very little throughput problems, right? You can add on 100
or 100,000 users and what do you care most of the time?
Jim (12:44)
Yeah,
who cares? Just turn on another server, right? Like that’s the most you’re going to have for us. When that happens, we’re like, we need to pay attention to our cashflow all of a sudden, because we’re going to be outputting huge amounts of cash just to be able to produce the units for somebody and then, you know, there’s timelines down the road where, they want net payments. so like, we basically have to front load all of the cash versus having everything on the backend. So
Yeah, like there is a very real problem. we were to get to a place where we had a large number of very, very big deals all hit simultaneously, it’d be a great problem to have, right? We’d be able to figure it out, but it actually does cause like a cash crunch problem in those scenarios.
Jason Niedle (13:30)
Tell me, so you’re not doing much outbound, that’s actually really interesting, and you get inbound. Are you guys running educational campaigns? Like how are you getting that inbound, or is it just great SEO and great reputation?
Jim (13:40)
Yeah, I, you know, there is some SEO to it. Realistically, it’s somewhat of a niche field. And once people start to do research, our name does come up. Like we have tons of documentation. you know, we’ve done enough of the groundwork to be able to have a foundation for people to be able to find us. And we were very fortunate in the early goings of the company. We just, we timed it very, very well and not that it was intentional, but had a couple of really successful kick-starters.
That were really how we got going. So sold over a million dollars worth of products and two different Kickstarters. And that really like laid the foundation for the company going forward. So there was this one point where we’re selling 90% of our products to hobbyists and individuals, basically a direct to consumer. And that’s completely flipped on its head. And it wasn’t something that we actively made the decision to do. not like we said, hey, we need to pivot from B to C to B to B.
It just kind of naturally happened as all of those people who are buying those cameras were working in R and D labs and they’re working at startups. And then like when they started to have vision related problems as like robotics, it started to come more into the mainstream. then after chat GPT like in, 2022, like that release, all of a sudden you start to have downward pressure from the C suite saying, how do we get AI into all of these, into these systems? And so the people had already had experience working with us and then
Jason Niedle (14:53)
Hmm.
Jim (15:00)
we end up being a choice in the, or one of the options that are evaluated when people are pursuing how do we put vision into our products so, yeah, it’s a little bit SEO, like a lot of documentation, people knowing us a bit realistically. I think it’s just a signal that this is where the world is going into a sensor driven environment where robots and machines are going to be interacting with people.
And we just happened to be placed in the right position.
Jason Niedle (15:29)
Well, and you have the right reputation and the right brand. That’s key. Yeah. So what’s holding you back from faster sales? Internally, what’s your constraint and how are you working to solve it if you are?
Jim (15:31)
Yeah, that’s tough too, but you know, I’d rather it come out of your mouth than mine.
Yeah, I mean, like it is a really, really difficult thing. I just mentioned how we like started with B2C and are now in this B2B state. It is a really, really difficult thing to be able to have the level of consistency that is required to service these enterprise deals. So it’s just demanding on the resources and that happens in a couple of ways. But I think that the key one is that
like the computer vision isn’t really a totally commercially solved problem yet. And that’s partially why we’re able to have the opportunity. It’s not like, you you look at a two-year-old, they can pick up an iPad and they can flip through it and it’s really, really easy. It’s simple for them to use. You still need some engineering effort to have this stuff actually be able to be useful into the world, even though we make this great device, it still takes effort. And even for people who are very, very experienced in it.
they still need our help to work through some of those issues. so, it’s just gonna take time for us to continue to work through these and for us to continue to build our platform to a place where we can support more core vision functions, ⁓ like directly out of the box. And we do a pretty good job of it right now. We have hundreds of different applications where people can get get started right away, but those are sitting in that proof of concept phase or just the light bulb phase.
Jason Niedle (16:42)
Mmm.
Jim (16:54)
And getting from that phase to working reliably like 99.99% of the time, right? That’s really like the answer to that question is the same reason why all cars aren’t self-driving at this point. Like there’s enough risk in that fully production state that it’s just gonna take the world a little bit of time to get there.
Jason Niedle (17:12)
I know I only have you for a couple more minutes. The one minute quick answer, what are some trends that you see in the next year coming up?
Jim (17:18)
From our perspective, I think that you’re going to start to see in the coming years, a lot more, like subtle automation than you have seen in the past. I think there are a lot of companies who are looking at how do they enhance the experience.
of customers, how do they enhance their data collection by having cameras take information from the world. And they exist, but all of the cameras that exist in the world right now that you recognize, like security cameras or, those are probably the
the most logical, cameras that are out in the world, they’re really just passive. They’re collecting data to be reviewed later. Yeah, exactly. Like dumb cameras versus a camera being able to be in a store recognizing when the shelves are empty and being able to have things restocked very, very quickly.
Jason Niedle (17:57)
dumb cameras versus smart cameras.
Hmm.
Jim (18:13)
One thing that I would love to see us, in inspections, things like that. So an area and an opportunity that I think that we have is like in the rental car space. So like it’s those types of jobs. And I think this will frame it a little bit. So like when you go and you return your rental car, you have somebody going around and like,
marking up is there a dent in, like where are all the dents, where are all the scratches, that sort of stuff. You’ll start to see cameras in place gathering that information and feeding it back. So ⁓ it’ll basically offer some seamlessness to the way that the world works. we see it with a lot of our customers who are exploring these types of things already and starting to get them to production ready states. So it’ll be interesting to see that stuff
Jason Niedle (18:38)
Mmm.
Jim (18:52)
actually come in and you probably won’t notice it right away, but there will be a lot of automation that starts to creep into the world.
Jason Niedle (18:58)
Really cool. Man, it’s mind-blowing what’s going to happen here.
Jim (19:00)
Yeah, I mean, if I were able to share like some of the stuff that it’s really, really crazy, the agricultural industry is really, really fascinating. You’d think it’s, you know, farming is kind of a boring and an old industry, but there’s so much automation and so much robotics that’s in in that space, from yield estimations from soil analysis.
Literally like robots with like nine different arms picking strawberries off of branches. It’s really wild. The things that people are doing and it’s only going to become more interesting.
Jason Niedle (19:27)
That’s amazing. Well, I could talk forever. I know you have a cut off. Where can our audience find you?
Jim (19:31)
Yeah, luxonis.com is where you can learn more about the company and there’s a link directly to our shop. can buy our cameras directly from the store and then any of the social handles, we’re just luxonis. L-U-X-O-N-I-S.
Jason Niedle (19:44)
Jim, thank you so much for being on Beyond SaaS. For tech leaders out there, we’re committed to exploring tech growth, so we drop episodes twice a week on Tuesdays and Thursdays. And you can find me, Jason Niedle, at tethos.com, t-e-t-h-o-s.com. And don’t forget to grab that hyperscale playbook at tethos.com/podcast. If you got some value today, please share, like, all that fun stuff. Until next time, this is Beyond SaaS
