Sean Kenney: Hi. I'm Sean Kenney and welcome to the MFS All Angles podcast. Each episode, I'll sit down with analysts and portfolio managers to dive deep into industries, explore key themes for investors, and uncover how they approach investing from all angles. Our goal is to make the complex world of investing a bit simpler and provide you with practical insights that you can put into action. Today, we're focused on one of the most dynamic industries in the world, software, and I'm joined by Technology Sector Team Leader and MFS Software Analyst Matt Doherty. So Matt, thanks for joining the podcast.

Matthew Doherty: Thanks for having me.

Sean Kenney: Well, I'm excited for this conversation because software, as I mentioned, is one of the most dynamic and interesting categories and industries in the world. But something else that's interesting is your background before you became an investor, you were a competitive boxer. And my understanding is you had dozens of Golden Glove bouts, you had a championship Golden Glove bout, and you then turned in the ring for the pit, became an equity trader in 2010, evolved your career to be a long-term investor, and joined MFS in 2019. Is that right?

Matthew Doherty: That's right.

Sean Kenney: So tell me. I mean, your investment background is fascinating and interesting, but talk to me about boxing because I have to imagine that you took some lessons away from the ring. And as an investor, I'm sure you're taking it on the chin pretty regularly. I'm sure boxing prepared you for that. But what did you learn in the ring?

Matthew Doherty: Yeah, I'd say boxing was a big part of my life growing up and I think if there's one thing ... There's an old quote in boxing. It's, "Everybody has a plan until they get punched in the mouth." In investing, that happens too. When I used to think about how to prepare for it in boxing, it was always about playing the long game, not coming out too hard and fast, and always focusing on the long-term plan and making sure you're prepared, and I think those are kind of good corollaries to how I think about investing. You're not going to always know every answer, but if you prepare and you focus on the long term, I think good things can happen.

Sean Kenney: Yeah, focus on the process.

Matthew Doherty: Exactly.

Sean Kenney: Yeah. Yeah. Well, remind me not to get on your bad side, but we'll keep this right up the middle when we talk here. So the first thing I wanted to start with, let's talk about the industry because software is dynamic and it's ever-changing and it's been one of the most powerful engines of growth in the market. You look over the past 20 years, software is increasingly a bigger part of the equity indices, and I've heard you say that the game has changed. What's changed over the past 12 to 18 months for software companies?

Matthew Doherty: Sure. And even if we go back a little bit further, software is no stranger to migrations and changes in the industry. When we first started decades ago, you'd be selling CD-ROMs and discs and have software installed on your mainframe computer. Then we went through a shift of cloud computing and SaaS, software as a service. Between 2000 to 2020, there was a pretty strong secular move in SaaS. And then as we hit 2020 and COVID, there was a big pull forward of software demand. Companies were buying multiple different point solutions. When rates rose, you saw a pullback into the software cycle of buying patterns. You saw a lot of companies trying to absorb what they had already bought. That was at the same time around where ChatGPT came out in November 2022.

And it's interesting. A lot of technological evolutions can often happen at the same time as downturns. You saw that with the mobile era. And so originally gen AI and ChatGPT was thought about with software in terms of co-pilots and helping people that use software become more efficient. And then I think the biggest change, to your point, over the last 12 to 18 months has been the emergence of agentic AI and what that's going to mean for the way in which we do work in the future and the way in which we sell software in the future.

Sean Kenney: I think there's a bit of a misconception between the difference between gen AI and agentic. Are these the same things? Are they different? How would you describe that?

Matthew Doherty: Yeah, I guess I'd start by saying companies have always tried to automate workflows. And historically before gen AI, it was deterministic systems. So examples of that would be something called RPA, robotic process automation. It was essentially, to simplify it, it would be, "If you do this, if you click on this button in a software system, then do this workflow," and it was very rules-based and regimented. Gen AI came out with the concept of probabilistic systems, which is essentially using AI to come up with answers, novel answers, that are leveraging data that is coming up with an answer that maybe you didn't know it was even going to say. That was ChatGPT and it was very prompt based.

The emergence to agentic AI is allowing these systems to act autonomously and perform workflows without the prompt. So it can actually go into a system and update a record. It can actually take a call from in a contact center and give an answer and then update something else. It allows the system to work a lot more autonomously, which then it starts to get into the conversation of what does that mean for labor? Because historically in just the gen AI copilot era, it was an assistant to a human. Now this is arguably potentially taking the place of a human.

Sean Kenney: So I would imagine this is necessarily positive for the software industry because you're tapping into not only the technology budgets of companies, but the labor budgets. You mentioned software as a service and the SaaS explosion that happened over the past decade-plus. The narrative now is SaaS is dead. What's driving that?

Matthew Doherty: There's three main concerns about software right now. The first one would be that because of autonomous agents and the ability to just spin up applications ... Because one of the use cases of agents, there's customer service use cases, but another one would be automated code generation, allowing an agent to just write an application for you, kind of explaining what you might want the application to look like, but then it just spins it up. Historically, if you were to start a software company and you needed to raise venture funding, you would want to go hire an army of developers, an army of salespeople, it would take a lot of money and time. I think the barrier of starting a software company has dropped pretty dramatically. And so that just equals competition is rising in the software industry so the moats become very important. That's number one.

The second concern would just be the pricing models. The pricing models need to evolve because historically, software sold licenses to people. They sold seats. If you had 100 people and you wanted 100 seats, that's how much you would pay the software company. In a world where you're having agents do work and they're acting autonomously, there's no people to sell that software to so the software companies need to start pivoting their business models. There's another reason why that would make sense too because if you think about a software company that might sell more of a consumption-based or an outcome-based pricing model, their costs are going to the large language models. Their costs are the compute that the large language models are spinning up. And so if your costs are variable but your revenue is not, then you have a mismatch and so they're incentivized to do that.

There's one more, the data moats. Historically, systems of record collect a lot of data underlying them and that has been a pretty strong moat around incumbent systems of operating. There's new technologies called model context protocol and different ways to seamlessly transmit data that is starting to build into question whether or not that data is going to have gravity with the incumbent system of record.

Sean Kenney: So if you're an investor and you see all this change happening and the unpredictability of the future, could you not just invest passively in a basket of software stocks and just let the winners rip and the losers will fail and you're going to be okay?

Matthew Doherty: Yeah, I don't think that's the right play mostly because the industry is at a turning point and I think dispersion in the future is going to be extremely high. I think the technological evolution that we're seeing within agentic AI is creating a pretty big question mark about what your moat truly is. And I think there will be incumbents that will be able to bridge that gap and make that pivot and adapt to this new way of working, but then also there's going to be significant disruption. And so making sure you're thinking about who you're investing behind, I think, has only become more important.

Sean Kenney: So selectivity matters in this space.

Matthew Doherty: Yeah.

Sean Kenney: So let's transition from an industry conversation to more of the businesses within software. When you're hunting for investments and looking at companies in the software industry, in your mind, what makes a great business in software?

Matthew Doherty: Some of it is very relatable to other industries. You always want a business with pricing power, you want to find businesses with large and expanding markets, and you want to have good competitive environment and good industry structure and good management teams. I think within software, I always like to look for a strong history of organic product development. Finding durable moats in software is very difficult and I think right now is probably the most important time to be thinking about that because of the rising competition that's coming. I've never really viewed fast mover or ease of use, which often historically had been quoted as advantages. When there's so much capital going into this industry and the barriers of starting new software company have dropped, I think that's when the moats start becoming the most important. I think there are still some durable and enduring moats, which would be switching costs or your brand and distribution, data gravity as we talked about earlier. And so really trying to find and identify those is important.

I'd say the last thing would just be a margin profile, so growth plus unit economics and making sure that that company is growing profitably. There's a couple of different types of software. There's horizontal software that sells sales or back office software to a bunch of different verticals and then there's vertical software that sells a type of software into a specific industry vertical, say it's the insurance industry or the banking industry. What we find is often those vertical software companies might have lower gross margins because they have more implementation services, but they have higher operating margins because they can be more efficient about how they think about R&D and going to market and selling to the customers because it's very targeted.

Sean Kenney: Yeah. I mean, software hasn't necessarily been known for having these incredible moats around their businesses, but your view is moats matter now maybe more than ever?

Matthew Doherty: Yeah.

Sean Kenney: So if you think about investing, it's a lot more complicated unfortunately than just picking great companies. There's also the whole thing of picking a great stock. So when you think about picking a great software stock, what do you think about?

Matthew Doherty: I think you want to start with the moat. We want to find businesses that are winning, they have competitive advantages, and they're selling into a market that's growing and is secularly expanding. I think the other thing to think about is finding inflections in growth, periods where this software company is changing for the better, and that could be through a new product cycle. It could be through competition changing. It could be through management changes. I think we want to try to identify those names where there's an underappreciated earnings profile associated with a company because it's one thing to find a company that's going to inflect growth, but we want to also make sure that's not already well appreciated by the market and that it's already in estimates or it's already in its multiple. And so that's what we spend a lot of time thinking about. Is it a good business, is it in an attractive market, and is our earnings power underappreciated that we have with a long-term view?

Sean Kenney: In the software industry, there seems to be a lot of hype and a hyperbole and everything's the next big thing, particularly since a lot of the stocks are priced on the potential of future earnings and things like that. How do you gain confidence in identifying an inflection point and making an actual investment decision behind it? Do you have an example that you've had recently or over the years?

Matthew Doherty: So an example of this would be a portfolio manager came to me a couple of years ago, Eric Fishman, who manages our growth funds. And this is one of the benefits MFS too is it's not always a one-way street where the analyst is going and pitching a portfolio manager. Often, portfolio managers will come to the analyst and ask them, "Hey, what do you think about this stock?" And this was a vertical software company that sold ERP, enterprise resource planning software, to the insurance industry. This was a huge laggard software company because it went through what I would call innovators dilemma. Two decades ago, they used to sell on-premise software to this large insurance industry, but then the world started going to cloud. But they didn't really have incentive because their customers weren't really asking for the cloud so they didn't really have an incentive to rebuild the product.

2019 comes, new management comes in, changes and rewrites the cloud product. And at the time of 2020 and 2021 when a lot of software stocks were doing really well, this stock was not doing well. Their gross margins were declining. It was a really big underperformer. What we ended up seeing when we dug in a couple of years later was that the incremental gross profit was actually over 100%, meaning they were adding more gross profit than they were adding revenue and that was because they were cutting costs in their cost of sales because they were through their product cycle. We met with a management team. They confirmed that they were through the product cycle. They're on the other side of it.

One of the nice parts about working on the global central research platform is the property and casualty analyst that covers essentially their customers is two doors down for me. And so I walked over to her desk, her name's Molly Frew, and I asked her about the PNC industry and she said, "Well, rates are rising. This is actually good for this vertical. They have an investment book that's actually doing really well and they're talking about using this capital to modernize." And so that said, "Okay. New management, new product cycle, industry vertical getting better." And so then we started talking to the customers in the ecosystem, partners in the ecosystem and trying to dig in, "Are they going to be a winner if this whole industry starts moving to the cloud?" and we more and more heard yes. Actually, it's extremely difficult for us to use anybody else because the switching costs, back to our original comment with the moats.

So working through and collaborating with other analysts on the team and then using kind of the management access that we have just allowed us to gain confidence and thinking through that earnings power at the end of the day. And then the last part would be thinking through what the right terminal multiple would be and how to think about valuation.

Sean Kenney: You identify the inflection point. You have confidence to make an investment decision. Then the question becomes terminal value, valuation, how do you get there? So many industries are more profitability-focused, PE, price to free cash flow. Software is somewhat famously anchored to sales and the potential for profitability into the future, which can lead to hype cycles and a lot of the things you see in this space. So how do you think about valuation?

Matthew Doherty: Software sometimes gets lumped in with very volatile multiples. I guess at the end of the day, every software company is worth the present value of their future cash flows. When you're a software company that doesn't have upfront cash flows and more of the cash flow and the value is in the future and it's in your terminal value ... Well, when there starts to be questions about your existential risk and technical obsolescence and things like that, then there's questions about that terminal value and your multiple starts moving around a lot mostly because you have such long-dated cash flows.

To your question about sales multiples, we don't really use that as a starting point in our construct. We think about underappreciated earnings power, whether that's through a migration, through a product cycle adoption in an S curve, and think about what could that free cash flow per share be multiple years out? What's the right multiple to put on that free cash flow per share based on the duration of growth and the rate of growth going forward from there? And at the end of the day, that does back up into an implied EV to sales, but it's not our starting point generally. We try to anchor to free cash flow dynamics.

And then the last thing I'd say is software, you might sometimes hear about the rule of 40. It's a heavily used construct in software where it's your growth rate plus your margin profile. So if you grow 20% and you have a 20% free cash flow margin, that's 40%. And then what does that mean for what your multiple should be? I think that can get investors tripped up because it's a point in time and it doesn't really take into account duration. We've never really used that rule of 40 framework very much because that can lead you down the wrong path of buying fast-growing commodities. Because what's really important is not that you're growing 30% this year, but what is it over the next three to four years? And if that goes from 30% to 10% to 5%, it's going to be a really bad stock, but it might've looked good on that chart at that moment.

Sean Kenney: So valuation alone isn't a great screen for this space? There's a lot more to it?

Matthew Doherty: Valuation is rarely a catalyst in software. I think where you want to start is where fundamentals are changing, inflecting, and getting better and better than where expectations currently are.

Sean Kenney: So let's talk about a potential catalyst, and that's AI. And we started a little bit of that conversation earlier on, but if you think about AI, gen AI, and agentic in the future, do you see AI being a positive disruptor or is it a threat to the industry?

Matthew Doherty: It's evolving quickly. We're going to have a lot more information over the coming years. My thought is where we stand today is it's expansionary to the tangible addressable market because it's inherently a new budget to go after, which is labor. Who's going to win in there is going to have distinction based on what they're doing today and different characteristics of their existing business model. But what's happening today I think is you're seeing a big usage in the consumer space, ChatGPT. Unlike other technical shifts like the cloud, the cloud when it first came out was met with a lot of reluctance. People didn't know whether this is going to be secure in the cloud, "I'm not sure if I want to do that," whereas AI is in every boardroom and every household being talked about right now. And so enterprises are trying to think about how they're going to leverage AI, but they're concerned about their data readiness and data governance and how much data is going to be released into an agent that they have really no control over the outcome of.

And so what I think is happening with IT budgets in the enterprise is AI is kind of sucking away from what they may have historically spent on, whether that be business services or professional services or that next subscription from a software company. I think boards and management teams are just thinking a little bit harder about, "How am I going to work in the future? Should I sign this three year deal with the software company or are we going to actually change the way we're doing this?" So right now, you're seeing a little bit of an air pocket in the enterprise and I think it's going to take a long time from an enterprise standpoint. Consumer is different, but I think like many other technical shifts, you're going to over-hype it in the very beginning. But the long term, it's probably underappreciated how much of a change this can make to organizations operating leverage.

Sean Kenney: So last episode, we had Matt Scholder on, who covers the pharma space, and we talked about the role AI will play in drug discovery in the pharma space. And what he talked about was the winners are really going to be those who can apply AI to unique data sets. I would imagine that incumbents would have an advantage and they might have access to these data sets. Is that a fair assumption or is that false?

Matthew Doherty: I definitely agree that domain expertise is incredibly important because if you think about it, data is the gold of AI. You need context and you need data to be able to utilize AI tools in a way that's going to be productive for your business. From an incumbent standpoint, they have distribution. They have the install base. They have the data. But what they also have is a bit of innovator's dilemma because they have existing business models that are arguably done in a different way than the future of work could be potentially done. And the AI native companies, the startups that are going to run fast at this, they're not going to be burdened with that. And so I think what we need to do is make sure we're constantly thinking about what incumbent vendors have the architecture and modern tech stacks and management foresight and strategy to be able to bridge that gap such that the AI startups that are probably sprinting a bit faster, they can withhold that competitive concern.

Sean Kenney: So how do you research these data startups? Where do you start on that?

Matthew Doherty: Yeah, this is another great part of the global central research platform is we're constantly going to conferences, constantly traveling with team trips. We just recently did a Silicon Valley trip with our entire tech team. We had analysts that cover semiconductors, internet, software all in the same meetings, and so we meet with privates. We're constantly tracking who's growing and also what buying centers they're disrupting from. So if XYZ private company is growing really fast, we want to understand who should have gotten that sale? Who is that taking net new ARR from? And often, it's at the very early, small, mid-market businesses, but if you start to hear some of the enterprises switching and adopting a new player, that's an important insight and data point for a lot of the incumbent public players that we're looking at.

Sean Kenney: Yeah. So I have to think about as we look forward, we talked a little bit about the past 12 to 18 months. As we look forward over the next 12 to 18 months, if you were to summarize a few key themes that you're focused on in this space, what would they be?

Matthew Doherty: Thinking through the entire ecosystem is going to be important and not just tracking what's happening in software. I think we want to understand what's happening at the hyperscaler level, hyperscalers and how much that they're spending on CapEx, and what's the ROI on that spend? What's happening in the semiconductor ecosystem and what does that mean for the uptake of AI? And is that because of the consumer uptake or are we actually starting to see some enterprises really start to ramp their inference tokens. Watching that and also watching the enterprise ROI, this is another great part about being able to collaborate with your teams in different sectors because some of these enterprise customers, they might be a consumer company or a healthcare company that I can talk to Matt about.

Understanding the ROI that the customer is seeing is going to better inform the potential value capture that the software company is going to be able to create as long as the competitive dynamics are such that this won't just be a pricing war to zero at the software layer. And so we're watching. I think the whole ecosystem is important to keep an eye on. And the last thing I'd say is keeping an eye on competitors that we might not see coming, not currently AI application software companies today which would be the frontier models, and if they're going to potentially move up the stack and start playing in the application layer, I think is something to keep a close eye on.

Sean Kenney: Okay. And so does that sort of capture what you would consider key risks in the space or is there more to think about?

Matthew Doherty: The key risks for the space are that every technical change in every technical architecture change historically in technology has created risk of insolvency and disruption. I mean, you saw this with mobile, you saw it with internet. You saw it with Blockbuster and Barnes and Noble and multiple companies that missed that tech shift. I think the biggest risk in the industry right now for some of these large incumbents is that they can't bridge that gap and they don't have the moat to withstand competition or they don't have the cultural or management strategic vision to actually pivot the business in the way that it needs to.

Sean Kenney: Okay. Well, we covered a lot. You packed the punch. So let me just try to recap a few key themes for our listeners. The first is that dispersion is likely to widen and this is a time to be selective in the space. The second is that moats matter and thinking about the rate and the duration of growth is going to be key for good businesses and good stocks in the future. And the third is that AI is likely to be a disruptor, but it's the companies that can use unique data sets and integrate AI into unique data sets, not necessarily just having a big incumbency advantage of those that are likely to win. Are those three key themes that our listeners can take away?

Matthew Doherty: Absolutely. Incumbency won't be enough, but if you end up being a winner in this space, I think the market is growing because the labor budgets is what you're really going after and that's TAM expansionary. And so I think it's both a big risk to the incumbents, but it's also a really big opportunity for the industry.

Sean Kenney: Yeah. Well, that's really helpful. Thank you for the time, for your insights into this space and ...

Matthew Doherty: Of course.

Sean Kenney: Appreciate you being here with us.

Matthew Doherty: Yeah. Thanks for having me.

Sean Kenney: And thank you for listening to All Angles. If you found this episode helpful, subscribe so you don't miss any future episodes. And until next time, consider your investment decisions from all angles.

The views expressed are those of the speaker and are subject to change at any time. These views are for informational purposes only and should not be relied on as a recommendation to purchase any security or as an offer of securities or investment advice. No forecast can be guaranteed. Past performance is no guarantee of future results.

Unless otherwise indicated, logos and product and service names are trademarks of MFS® and its affiliates and may be registered in certain countries.

66259.1