Michael Horn argues workforce development programs need bundled services, not modular ones. When performance is immature, tight integration works better: IBM's mainframe ecosystem succeeded through bundling, while fragmented programs like Job Corps underdeliver. Horn proposes integrating training with temp-agency placements so employers can trial graduates with low risk.
Good evening, everyone. Michael Horn here. Excited to talk to you about my latest piece for Forbes, in which I talk about why workforce development solutions must actually bundle up. And my friend Ryan Craig, investor in the education and workforce training space and a noted author, he recently penned a piece where he told a number of pretty funny stories about his kids to explain why even as higher education it's time for it to unbundle, workforce development solutions actually have to do the opposite. They have to bundle across all the interfaces that matter for reducing education friction and for reducing hiring friction for both students and employers. And the funny thing is that Ryan didn't have to rely on his amusing stories of his kids. There's actually great theory of innovation in the name of interdependence and modularity that Clay Christensen coined over two decades ago to explain why Ryan's intuition around this is absolutely correct. This is the theory that Ben Thompson at Strategy uses all the time to help explain the successes or failures of companies like Apple or Samsung and folks like that in the shifting domain, in the microprocessor space and so forth. But it's a theory we can use in higher education as well. And so I'm going to illustrate the basic tenets of this theory over here on my whiteboard. And we're going to put a basic graph that explains any sector. And on the Y axis we'll chart performance. Hopefully you can read that. And on the X axis we'll chart time. And what we basically see is that in any sector, the customers or people that use a certain service, they have this. Their sort of trajectory in life is relatively flat, slightly sloping upwards, but relatively flat. And it's because customers lives don't actually change that much from day to day. So what they're demanding one day, the next day it's going to be something similar and so forth. Now, of course, there's a range. There's some people at the high end who are never satisfied with no matter what you give them. And there's some people at the low end who are satisfied with very little. But basically we have these trajectories of improvement that customers can utilize or absorb over time. Now we have a second line or trajectory that's really interesting, which is the pace of technology improvement over time. And what it says is that any new service or product, when it gets its start in a new sector, it tends to start at the bottom. It's not good enough for the majority of users in the field. And over time it gets better and better. As people understand the service more, they understand how to perfect it, they start to put in more features and functionality and so forth that appeals to more and more people up that chain. As you go up this performance axis, sometimes that marches small year to year incremental improvements, sometimes it's giant breakthrough leaps forward. But the point is that over time, services get better and better and better. Now, the funny thing is that at the beginning of any of these sectors, because the new thing is so immature, the technologies and the various features of it are so poorly understood. But if you hope to make something that's just good enough that any customers will use it, then you actually have to make it with an interdependent architecture, meaning the way one part of the service works and functions will depend upon the way another part works and functions, and vice versa. And so for any organization that's getting its start in an immature market, or introducing something that's not good enough for the majority of customers, in a poorly understood service, if you want to make it good enough, you got to do it interdependent. You got to bundle up all the parts of the value chain and do everything. So IBM, for example, in the early years of the mainframe computers in the 1940s and 50s, they actually had to build the logic circuitry, the core memory, the operating system. They had to have the salesforce, they did the design, assembly, everything, everything. Because no one knew how a personal computer should work. Higher education looked like this for a long time. And we've bundled up more and more and more. As colleges and universities offer you all sorts of majors, they bundle research with teaching and learning. You have sports teams that you pay for, dorm service, food on and on and on, deeply bundled service that's expensive and overshoots what a lot of people need. And so Ryan's point often has been that it's time for this, these expensive universities to start unbundling. Aptly true. It's something that Andrew Kelly and I have written about in some policy briefs and written about it elsewhere as well. But the interesting thing that Ryan observed in his piece, and then I talked about in my Forbes column, was that workforce development providers, there's this whole Alphabet soup of folks like the job corps and so forth that basically are underperforming in terms of giving students what they need to retrain in their lives and get them to good jobs. They're underperforming what students need, they're underperforming what employers need, and they're very unbundled, actually very modular offerings. And modularity makes a ton of sense when you start to overshoot because you want to start to be able to respond with more affordable, customized, convenient, rapidly responsive services. So for example, in the personal computer world, Dell personal computers in the 1990s was a totally modular or unbundled offering where they literally didn't make any of the parts inside of the computer great for allowing for affordable customization, not great for maximizing performance. And so if you're in this side of the world where the products and services are underperforming, you need a bundled architecture. If you're on this side, you need a modular architecture and workforce development because it's underperforming the market. What the customers need, students and employers, we need a bundled offering, not an unbundled offering. So workforce development, that has been totally just one off solutions point solutions totally ineffective. They need to start integrating across the interfaces that matter, meaning things that reduce education friction and hiring friction. Often that'll mean integrating services that actually hire, for example, students after they finish a program into sort of a temp agency type arrangement where they place you into a job, that employer gets experience with you with no risk, and then can hire you if you actually prove to be a good fit. And the program is consistently delivering graduates that are great for those jobs. Just one example of how you can start to imagine that integration across the value chain. So that's the bottom line. There's a good body of innovation theory, interdependence and modularity that shows why it's time for workforce development solutions to bundle even as higher education is unbundling. Have a great night.
