Thank you for coming out. My name is Ethan Chong. I'm an associate under Ken Newman covering the industrial supply chain. With me today is Dennis Fehr, the CFO of Cognex Corporation. Dennis, thank you for coming out today. To kick things off, why don't you give a little bit of background about yourself and then an overview of who Cognex Corporation is?
Yeah, sure. Happy to, and good afternoon, everyone. Thanks for taking the time and your interest in Cognex. A few words about myself: Chief Financial Officer of Cognex, about a year in the role. It's my third CFO job. I had been at both private and public companies prior to that as a CFO. Who's Cognex? For those who don't know us, we are the Cognition Experts. That's what the name stands for. I like to say we bring the real world into the machine world. We take the photons in the real world, we convert them into images, and then we put our software and vision tools on top of it and really look at what's in this picture. That means, in fact, real automation.
For inspection, for example, or for guiding of robotic arms, can be identification of products, very good old barcode reading, if you want to say, and gauging as well. That's kind of what we do. I would say as a company, we are regarded as typically as a tech leader in the space, both in the pre-AI area and the pre-AI era. Now also, I would say since 2022, when we launched our first AI product, also in that now new AI area that we have really embraced AI as kind of the next level of machine vision and to drive basically additional use cases and penetration.
Overall, we are, I would say, in an attractive market, which has the potential for attractive growth rates driven by continuous cost out and quality optimization desired by our customers, reshoring and maybe regionalization of global manufacturing, as well as labor shortages in some instances. At the same time, while we operate as a kind of software on device model, that means we sell the software and hardware, and go out with that, we command kind of software-like margins. At least traditionally, have been like a historic average of 28% adjusted EBITDA margin that has compressed a little bit over the last couple of years. That's what we are right now addressing heads on as maybe as a new leadership team, if you want to say so. Matt Moschner, our CEO, came into the role just six weeks ago. As I mentioned before, I'm about a year in the role.
We're driving back to where we think we belong in terms of bottom line profitability. In that regard, looking optimistic into the future.
That's great. Maybe just to kick things off on my end, you guys just reported earnings a few weeks ago, and commentary around positive improving trends within factory automation and e-commerce logistics looked pretty good. Investors clearly seem to like it with the stock up 20% on the day. Can you talk a little bit more about what you're seeing in your end markets and what's driving visibility into the third quarter and second half of this year?
Sure, absolutely. Maybe before I go into a particular end market, maybe first big picture, right? On the one side, we are cyclical in nature, as very often industrial automation CapEx-based businesses are. We saw the peak of the last cycle in 2021 and then saw a more pronounced, I would say, down cycle. '22 flat, '23 strongly down, and then starting only to flatten out in 2024. During this time, the company also did some investments into future growth, which on the one side, as a new leadership team from a strategic angle, we like, but at the same time, it really compressed bottom line a lot.
I would say, I think what the market reacted a bit to, at least what I hope the market did in the last earnings call, was to say this refocus on bottom line profitability and at the same time showing also really first-up progress there. Maybe to the questions on the end markets more specifically, what we see there. First, which markets are we serving? Our largest end market is logistics and call it warehouse automation. That's about 24% of our business. The second is automotive. Third is, we call it packaging. That's fast-moving consumer goods and healthcare. It's just like the application is very similar on this one. That's why I put that into one. Consumer electronics the fourth and semiconductors the fifth. I think over the last 12 months, or probably now 18 months, we've seen logistics growing very nicely.
We see strong momentum there, as it's a market which is not very deeply penetrated or is actually the lowest penetrated market with machine vision. Therefore, we see really a long-term growth potential there. It's also the market where we typically have the most visibility. That means for us nine months, right? In all the other end markets, which we sometimes combine to call it factory automation, we just have like three months visibility, so much shorter. I would say maybe other highlights we saw in the quarter, we highlighted positively packaging, that we've made good progress there, especially with a dedicated sales focus. For the first time in many years, we highlighted consumer electronics as kind of an early sign of positivity. We have seen consumer electronics 2021 a great year. Since then, it was almost like kind of a refresh cycle, year after year.
It was the first time that we said, hey, in this market, we are now seeing the opportunities from the reshoring, seeing the first moving of supply chains from China into other Asian countries. In general, a more positive sentiment in that industry. That's maybe a few words up front on the vertical markets.
That's great. You touched on it a little bit, but maybe on the margin side of things, it was pretty good to see you guys have EBITDA start trending towards that 20% you guys had outlined earlier. Can you help me understand what sort of initiatives are in place right now and how much of it is more structural as opposed to just flexing with the company's needs at the moment?
Yeah, no, thanks. Thanks on that. On the one side, great progress, right? For four quarters, with year over year EPS increase, first quarters since Q2 2023, above 20% adjusted EBITDA. Nice progress. If you think about it, we're looking at the entire P&L, right? In our mind, Matt's and mine, we're thinking about like, let's drive what we can influence and what we can control. What can we control? On the one side, we can control on the top line. We can control a bit like where do we focus our sales efforts, like less automotive, which is not a good market right now, and more on like packaging, healthcare, fast-moving consumer goods. That gives us at least a few percentage points of growth, even if we would leave market out for the moment. We can control our cost base.
At our recent investor day, we talked about that we believe our biggest opportunity in driving bottom line profitability is through OpEx efficiency. That's what we have seen now in these quarterly results that we saw kind of year over year in a constant currency basis, OpEx coming down 3%. In that regard, we think we're really driving structural cost optimization. It's there to last. It's not something driven by anything external. It's really something we control and we are focused on.
Yeah, that's great. You said you just had your company's first analyst day, which is very exciting. You and Matt talked a little more constructively about M&A as a pillar of the long-term growth strategy. I guess maybe can you touch a little bit more about what sort of targets you guys are looking at? Maybe say higher margin software as opposed to hardware like Vortex or what that might look like. Just overall, what's the shift behind the strategy?
Yeah, no, thanks. Thanks for that. Yeah, we talked about investor day about our kind of new growth algorithm. In general, first of all, we brought it down a bit to where it was before. We kind of sliced it in almost like you can say three components. First, we said like, what's the growth? Oh, I think my mic dropped. Sorry for that, guys. Three components. The first one is, is it still on? Is it good? Great. Okay. Underlying market growth means like how much is the market of our customers growing. We said like through cycle, it's 4%. That's really the biggest block of growth. It's like 6% - 7% penetration. That means we see the opportunity to penetrate our customers from two directions more. One thing is more applications enabled by the adoption of AI.
The second one is about going to more customers, which we have traditionally not served. That's kind of the two vectors there. The third one is M&A. That's really the smallest in the three growth components with 3%. We brought that in as we believe Cognex has a very strong balance sheet, no debt, very strong cash generation. Could we add additional shareholder value creation through a synergistic M&A planning or execution? That's kind of what we wanted to convey with that. What are we looking for? We look for highly synergistic businesses with the conviction that they can meet our financial framework criteria, especially being in that 20% - 30% adjusted EBITDA, either being within or with a clear path to within. You know, in the past for Cognex, synergistic M&A really meant focusing on tech.
Is that a tech which we like, which is maybe a niche which we need? That still exists for us today. As we have built out a really strong direct sales force, there's also a great opportunity for us to add and create synergies by driving more products through our sales force. That could be through additional tech, of course. It can be through consolidation in the market, or it can be by entering adjacency. That's kind of what is driving us there. I really want to clarify that we don't think M&A is the key piece to the growth algorithm. It's actually the smallest.
Got it. That's helpful. On the first two pillars that you kind of mentioned on the organic side of things, you guys recently announced OneVision, which is a very exciting product. Maybe for people who don't know, could you explain what that is? Who you're kind of seeing the most traction with in terms of what types of customers are maybe adopting OneVision?
Yeah, I think OneVision is one of these great examples to say, like, let's drive market penetration. Maybe let's talk about what it is, but probably don't know the knowledge here in the room. There might be different levels of information. Let's maybe take a step back first. What do we do mostly? Right? We sell, and I mentioned it before, we sell software on device. Right? That means, think about, like, you put a machine vision system on a production line, and it really executes and does all the computational task on the device in the line. That's what customers like because for two reasons: latency and kind of speed going into a cloud just doesn't work. Second, cybersecurity. The customers, they really like to have devices on the line, and they don't like very much to be in the cloud.
That means for you as a machine vision provider and a machine vision user, you can only use as much compute power as you have on the device. You can't use easily what the compute power is available on the cloud. That means as Cognex, I would almost argue one of our most secret sauce is to create machine vision models which are just optimized for the available compute power and make it work and cover as much as you can with it. That's probably where we are better than everyone else in the industry. Still, there can be use cases, especially very complex inspection tasks, where that compute power just is not enough. You would want to tap into the cloud in some way. The way how we found to do it is OneVision. Think about, like, OneVision maybe as a training room in the cloud.
That means a customer has a device, maybe has already used it for inspection tasks, and now he wants to drive it to the next level of inspection. Let's extract all the data from this, our Cognex device, bring it to the OneVision cloud, and then run a deep learning training model in the cloud using the compute power of the cloud and basically improving the model, not improving, let's say, the one inspection. It's improving the model in the training room. Then you bring it back from the training room back into the device. At this moment, you close the door to the training room, you cut your connection to the cloud, you're back in the line, you have the speed, and you have the cybersecurity aspect. That's basically what OneVision does. Why is that important? It's important from two aspects.
It makes it easier for, let's say, kind of the middle segment of the market to use deep learning type of applications, which in the past you would need to use complex PC-based kind of cloud-connected vision systems, very complex to set up, very complex to operate and to maintain. It adds basically or it gives access to a new customer group to that ability. For the more sophisticated customer, like the top of the market, they maybe have been using this more complex system, but it was still a hassle for them. It becomes much easier for them. For them, it drives efficiency. Therefore, we believe it's really a strong kind of differentiator and another addition to our Cognex ecosystem, which makes our customers being attracted to buy and stay with Cognex.
Got it. Maybe help me understand. Is it, at the moment, are you seeing more traction with those customers that maybe have it easier to scale, or is it the customers that have more complex tech needs?
Traditionally, Cognex as a company has always served kind of the most sophisticated customers with the most sophisticated problems. We're pretty good at that, and we serve some of the most iconic companies around the world or globally with that. In that regard, that's really where our stronghold is, and that's a good business. Especially in logistics, where I think it's well known, Amazon is our largest customer, goes very well. At the same time, we drove this sales initiative to go broader into the market. We like what we see at the moment in the packaging vertical, where we really see kind of yielding results. That kind of led to some of the more positive commentary in the last earnings call, where we said, okay, instead of seeing that more flattish this year, we see some growth happening for that.
In that regard, I would say there's traction in both, but very clearly, our original stronghold is the more sophisticated side, very clearly.
Got it. Maybe on the longer-term view of things, when we think about AI becoming more democratized and processing becoming more available, obviously, you guys have positioned yourself well within the machine vision space. Are you guys concerned that any smaller players may be adopting technology at a faster rate and just the accessibility to computing power, maybe that becomes a threat at some point?
Yeah, it's a great question. I think there are two sides how you think about it. There's a smaller player potential, startups entering the space. Sometimes you're getting asked like, do you think NVIDIA and Meta will be your future competitors with the large models which they train? I would say maybe the first part is very easy to answer because history already proved us that it's not a risk to startups, right? A lot of startups, like four years ago, collecting a lot of money, trying to get into the space, and they have all not been really successful in that regard. I think we can really say history has proven otherwise. Now, on the other side, with the hyperscaler, the large language model provider, obviously, that vote is still out in that sense. We believe that there are very good reasons why it would not.
First of all, I think one thing is you need to have very high accuracy, right? You can't have like a 99%. You need to more go towards five sigma, six sigma, that type of very industrial machine vision specific task. We actually ran a comparison of some of the large language models against our machine vision specific model, and we can see that we beat them on accuracy, on industrial machine vision tasks. That's one, accuracy. I think even more important is that what I said before, it just can't only run in the cloud. It needs to run on the device. That means you really need to optimize it for efficiency, right? You can't use this big language model which uses clusters and clusters of GPUs, but you don't have that on your device. In that regard, that is a big step.
I mentioned before, I think that's where really the secret sauce of Cognex is. The other topic, scalability. How easy is it to transfer that model from this specific task to another one? Now, if you run another product on this line, can you let a model still address it? If you want to add now add another production line, how easy is it to transfer? That's where we really consistently see our advantage and really also kind of our right to win, so to say. That's why we're not entirely concerned about kind of these large language models. At the same time, we feel like there's a lot of opportunity bringing AI into machine vision. Why? Because with AI, we are able to solve so much more applications, which we couldn't do in the past, that it's really a demand generator.
Got it. As you guys look to maybe fulfill this demand, you guys recently announced a new leadership team to help increase sales and growth and innovation. You're a part of that. I guess maybe from your perspective, what's kind of the most exciting thing that you want to see in the next few months from this group?
Right. First, I would say I would applaud Matt how fast he has moved on that, right? He's less than six weeks in the job and put together his leadership team. That's definitely part of it. He has been seven years already in Cognex before. I think if you look a little bit into what was changed, there's a very clear alignment of that team towards the strategic objectives which Matt outlined. First, being number one in AI, technology, and machine vision. You now have people from that leadership team who have not reported to a CEO in the past, they're now reporting to the CEO. That means they get more weight in the overall discussion, right? Second strategic priority, drive customer experience, be number one in customer experience in the industry.
Now, our Head of Customer Success, who runs our cross-functional customer experience push, hasn't reported to the CEO in the past, now reports to the CEO. Again, very clear tie to the strategic objectives. Last, to double the number of customers. Here we have a new Head of Sales, with Karl Gerst. He's not new to the company. He's 27 years in the company, but he never worked in sales. He worked on the product side. What he brings is process, process focus, very clearly driving a playbook, a very consistent sales playbook. I would say that's really a bit of a shift for Cognex as well. Again, here, very strong alignment to the strategic objectives. Maybe you ask what excites me the most about this team. I think for me, it's really focused execution.
I think that's really what we want to bring across to everyone on the street is that we put our targets at investor day. I think we are very focused and very clear about we will drive these targets. We think we can control at least a good portion of the destiny, right? We can't control generally the market, but we can control some part of that growth by our sales resource allocation. We can drive our bottom line performance by being disciplined on cost. Therefore, I think I'm pretty excited about that.
That's very exciting. It's great to hear you guys are all focused from the top of the management all the way down. Maybe touch on the emerging customer initiative. I know that's kind of a big point for you guys. I know that the second cohort is probably ramping about now. Maybe just how they're progressing and what makes you success or makes you confident in the success for the Cognoids?
Right. Maybe a different information level here in the room. What is it? What are we trying to do? If you go back to the last peak of the last cycle, 2021, Cognex, fantastic growth in the years before, but very high customer concentration with two very large customers, right? At the same time, you see some of our competitors, namely Keyence, being very successful in the broader market. In that regard, you would think like, hey, is there a great opportunity for Cognex to grow outside of the established customer base, going to more customers, with attractive margins, actually. That is kind of the strategic rationale behind it. I would say, the initiative launched in 2023 was hiring the first class. The first class of, maybe I should say, what are we trying to do?
Bringing in a different type of salespeople, which are more transactional focused and going broader into the market. We decided we are hiring a bunch of between college grads and kind of one, two, three years experience and bring them in, train them, and then deploy them into the market. We hired in 2023, released the first class in 2024, and then just released the second class beginning of this year. I think we went through a couple of learnings. I think we talked about that in prior earnings calls about like what products are they selling? What are really the right products? We adjusted a bit on that. We certainly also learned in terms of management. They are a bit different in terms of management. In the beginning, we kept sales management separated for that reason.
At the same time, we also saw it drove a lot of inefficiencies in terms of the coordination of the sales force. We combined the sales forces. We also realized like when we say about going to more customers, that does not mean necessarily two new logos. It sometimes can also mean that within one logo, you have different buyers, and we haven't visited all these buyers. We also let them go to existing logos, but trying to find new buyers who have maybe small buckets of budget, which we haven't tapped into in the past. We went through this learning curve, I would say. I think we made some positive comments about what we saw in the packaging market in the last earnings call. In that regard, I think we feel like we are moving into the right direction.
At the same time, I would say it's clearly a strategic shift for Cognex, right? It's really kind of almost repositioning Cognex from only focusing on top of the market into a broader play. That won't happen overnight. In that regard, I would say we're on the right track. There's still maybe I would call it opportunity because we expect that the sales productivity of each of these salespeople will gradually increase over three years' time. That means we'll still have a lot to get from this organization.
Got it. Yeah, that's all been very helpful. I know we're kind of up on time here. Dennis, I appreciate the time today. It's been great.