Is Vertical Integration Really Always Best? Nope.

Hey, it's Michael Horn here on the future of Education. Wanted to follow up on the podcast I did on YouTube and other channels last week with the CEO Dan Evita of Engagely. Super interesting active learning platform that he talked about. But we spent a little bit of time talking about the importance of vertical integration. And Dan essentially made the observation that the most valuable companies are those that vertically integrate and control the value chain. Now the example he cited was Apple compared to Dell. But what's really interesting is that if you went back just, you know, 30 years ago, you would have seen that being horizontally integrated like Dell, so a modular structure instead of an interdependent one was way more valuable than Apple that was trying to be vertically integrated. So which one is better? I thought I'd just do a deep dive to explain that it depends on the stage of the industry that you're in, the maturity of the technology, and that there's not one right answer to this question. I'll also say the points that Dan and I made, that a lot of VCs have historically been wary of funding fully integrated, fully vertically integrated offerings, in my mind is a limitation that's something that has often constrained the ability of entrepreneurs to do what an industry needs at a given time. In the name of being capital efficient, they're essentially being shortsighted. This most perniciously hits in developing countries where there's an immature value chain and distribution network and to reach the customers themselves for any given innovation. With that said, let me do a little bit though on the theory of interdependence modularity. Show how it played out in the computing industry and then we can talk about what this might mean in the education context because I think it's a great example where the notion of something is always this way or best practices can really falls by the wayside. And it's really much more important to understand the circumstance you're in to understand the right scope and structure of your venture. So I'm going to go ahead and share my screen and bring, bring up this diagram here which you can see. And so on this diagram we have performance on the Y axis, time on the X axis. For those that know disruptive innovation theory, this is sort of your disruptive innovation diagram where those red dotted lines represent customers. And what you see is that they're relatively flat, slightly upward facing or sloping because customers needs or jobs to be done don't change that much over time. They're relatively static, slightly become more demanding over time. There are of course tiers in the marketplace, people at the high end who are satisfied, who are hard to satisfy, and there are people at the low end who are satisfied with very little. And what you see is that when an innovation comes into the marketplace, starting with that blue line, it starts as not good enough for the majority of users in a field. But the fundamental tenet of disruptive innovation is that technology improves faster than our lives change. So what at one point is not good enough for the majority of users actually, over time, outstrips the ability of users to absorb or want to pay for all these performance improvements. When that happens, that's when a disruptive innovation, represented by the green line comes in at the bottom of the market and does its own upward market slope. Disruptive innovation start is not as good as the leading innovations, but they get better and better over time and come to replace the dominant. As people say, hey, that's good enough for my needs. I don't want to pay for these improvements that I don't need. In the early innings of any innovation, you compete by improving the basic functionality and reliability of your innovation. And to do that, you need these vertically integrated architectures. Vertical integration is absolutely paramount, proprietary interdependent architectures. And the reason is that the way one part functions depends upon the way another part works and functions and vice versa. So if you make any change in the architecture that's going to unpredictably interdependently impact an adjacent function. Okay, so if you want to do one thing, you got to do everything. In essence. The classic examples of this are, yes, Apple Computer is a great example, but so too would be the IBM mainframes. In the beginning of the computing industry, IBM, if they had just wanted to exist as a manufacturer or something seller of mainframe computers, they couldn't have, because there were no reliable folks that made logic, circuitry, core memory, operating systems. And so IBM had to integrate across all parts of the value chain to build those component parts, design them holistically with the computer, manufacture and assemble the computer themselves, and then sell it itself because there was no resellers either. Microsoft Windows also intricately interdependent within the operating system. Historically speaking, if you changed any 10 lines of code within that operating system because you wanted your customized version of it, well, it would screw up the way the entire operating system worked, because those 10 lines of code worked in very unpredictably interdependent ways with the rest of the code. Okay, so we're starting to get at the trade off here. Vertical integration optimizes performance, functionality, reliability, when your innovation is not good enough in the beginning of an industry, you need that architecture to get it adopted. But as you start to overserve people, people become more demanding, looking for more affordability but also more customization. They want it adapted to them increasingly. And that's where you see a new kind of architecture emerge. When you're starting to compete by improving speed, responsiveness, customization. And it needs to be affordable and the technology is matured so we understand the way these parts work and function. There are value chains that exist that make the different component parts and so forth. That's when you need these modular open architectures. You start to define standard interfaces between the different component parts. And people can work. One company can work on one part, another company can work on another, and they can just snap together in very crisply codified, understood ways. Dell, personal computers, Linux, those are the classic examples. Dell, right? You went on their website in the 1990s and you could specify, I want this kind of disk drive from Seagate, I want this type of monitor, this type of memory, this type of ram, et cetera, et cetera. They would snap together all these parts, none of which they made, and ship you out an affordable customized computer within 24 to 48 hours. Linux operating system is the, is the flip side of Microsoft Windows. You can substitute different kernels to customize the operating system. Not good enough in terms of raw functionality perhaps, but far more customizable at an affordable price. Okay, so there's a trade off in these things. Now here's the question that Dan asked, which is which one is more valuable? And it depends, Right? Let me show you the evolution of the computing industry over time. Which is it started with IBM completely vertically integrated there, and then it's almost like this went through a cheese slicer, right? And we started to see horizontally integrated companies emerge that are easier to scale by the way, horizontal integration. The reason you can create value out of it and is it's more scalable, but the market has to be ready for that. And so you see Dell, Microsoft, intel made the majority of the profits in those late 80s or 90s when the computing industry modularized, in effect now, it then shifted back to a vertical integration again. Okay, so Apple's vertical integration, that was a benefit in the early years of the personal computer industry, as you can see here, over time became a liability and Apple suffered. It then became a benefit again when it came time to re architect and relaunch the computer interface, which they did in the form of course of the iPhone with the iOS operating system, the app store that they offer, and on and on. Okay. And they're doing more vertical integration right now because there's differentiation, there's a need to compete on functionality, reliability. There's more demand at the high end of the market for these sorts of. For these sorts of improvements still. Okay. And so right now, they're at the high point. Now, what's interesting, and I'll stop sharing my screen here, is that. Let's see if we can stop this. So is that what you see is that it's a point in time, right, when something is valuable or not. It's not like an absolute law. Engagely, interestingly enough, is not as vertically integrated as, say, Minerva, a point that Dan very clearly agreed with when you talked about how vertically you could. Integrated you could be. Minerva Project says, we want to control the pedagogy, the curriculum itself and how you structure courses. And, you know, Engagely says, no, we don't go that far to vertically integrate that way. The question is, what is the right scope? And a theory can help you see what that is and where there's modular interfaces that you can fit into. One of the problems, I suspect, from Minerva Project is selling into an established structure that has clear interfaces where faculty have control of how they do the teaching. That's a struggle to say, just reinvent your model. New educational programs, new universities, they probably can fit right, fit well with, because they can architect around the vertical integration of Minerva, as opposed to having a plug and play interface, which Engagely is going for. And I think by being relatively more horizontally integrated, less vertically integrated than Minerva, easier time to scale. Now, the question that we asked relative to Class, which is built on the Zoom platform, is perhaps that's too limited because customers right now want more functionality and reliability. And that's the crux of the question that essentially Dan is making a bet on and Engagely is making a bet on. And I think there's a lot of rationale behind what they're doing in this moment. But I didn't want to leave it the case that vertical integration is always better because that's simply not the case. It does allow you to innovate more, improve functionality and reliability. But there are times in an industry where a modular structure is where the lion's share of profits are made, as it's easier to scale. Depends on the circumstance. Having made that point, I'll make one other point as well, which is Engagely talked a lot, not just about the active learning, but Dan talked a lot also about the social environment, super clever things that they're doing create making a more social experience that is engaging, motivating. People want to be a part of it. Really intrigued by all of that. The only caveat I will say is this is also circumstance driven. Everyone says, oh, you must have learned better in study groups. That wasn't the case for me. Maybe it's my personality. But I think also when you're a novice learner, there is considerable evidence that learning in groups taxes your brain, taxes your working memory. At a time you're just trying to get the basic facts down and all that interaction overloads your working memory. As you move into an expertise, as you start to become less of a novice learner and you start to have some expertise in a given field, then engagement and interaction becomes much more both important from a motivational perspective, but also viable from a learning perspective. And the trick in a lot of these systems, I think in the longer run is getting that balance right of when do you want that personal, maybe more individual experience and when do you want that group collaboration and social experience? I think the latter, the group collaboration clearly important for motivation, but we have to make sure we don't overload novice learners in given subjects at the same time. So again, circumstance driven when you deploy what strategy and that's a better way to construct a theory based on circumstance as opposed to an absolute way something always works. Hope this has been interesting. Normally I don't jump on after these conversations, but I thought it would be illuminating to pull this apart a little bit and go deep in some theory work as you think about, as you think about your own innovation efforts in education and elsewhere. Thanks so much for joining me on the future of education. Be back next time.