Why Universities Move Slowly

SPEAKER_00 0:00

Most colleges and universities aren't slow to act because leaders don't care about innovation. They're slow because they were designed that way. A professional bureaucracy. In a professional bureaucracy, authority doesn't flow through a hierarchy. It flows through expertise. Doctors run hospitals, professors run universities, accountants run accounting firms. Work doesn't move because a manager gives instructions. It moves because trained experts use their own judgment. That design is incredibly good at protecting quality and expertise. But Mintzberg also points out the trade-off. Professional bureaucracies are excellent at improving what they already know how to do. They are less equipped when the world around them starts changing quickly. And right now the environment around higher education is changing very quickly. I think the core issue is that universities are being asked to adapt faster on cost, on relevance, on skills, on technology. Yet they are built in a way that makes fast change unusually hard, even when everyone agrees change is needed. And the cleanest explanation I've found for that comes from the organizational theorist Henry Minzberg. His point is simple. Organizations don't just have org charts, they have a way of working, a built-in pattern for how decisions get made, how work gets coordinated, and how things actually move forward. He calls these patterns configurations. Our universities follow one very clearly: the professional bureaucracy. Work gets coordinated through trained experts who have a lot of autonomy. Our universities aren't struggling with change by accident. We are built to do something else well. We are built for reliability, quality, and expert judgment. So the real question becomes: how do we keep those strengths while also getting better at adapting? Because Minzberg is also clear about something else. When an organization is pushed to operate in a way it wasn't built for, friction shows up. You're going to get dysfunction. So here's the tension for today's episode. Universities are built for a world where expertise is stable and change is slow. But I would argue that the price of slow keeps increasing, especially in higher ed. Now look at the startup world. It developed a different way of working, not because it's better, but because it had to survive. These organizations are built for learning speed, for testing ideas quickly, for making decisions based on what they learn. In fact, Harvard Business Review recently ran a piece arguing that venture capitalists were responsible for launching one-fifth of the 300 largest U.S. public companies, and that the authors estimate three-quarters of the largest U.S. companies founded in the past 50 years wouldn't exist or have achieved their scale without VC support. You're probably wondering what venture capital could possibly have to do with higher education. I think it's useful here because it spotlights something important. A huge amount of modern innovation is happening outside of large established organizations. It's not because our colleges and universities don't care, and I think this is a sore spot for a lot of those in higher ed. We do care, regardless of what the outside world thinks. But no matter how much caring those of us on the inside do, our institutions are simply harder to change. So today I want to ask a very specific question. What can our colleges and universities learn about how to adapt without losing what makes them valuable? And to answer it, we need to get concrete about what fast learning organizations actually

What Fast-Learning Organizations Do

SPEAKER_00 4:06

do day to day. Which brings us to our field notes for today. When you read across the startup and tech experimentation literature, the pattern that keeps jumping out is this. They don't treat experimentation as a side activity. They treat it as part of how the organization runs. They don't just brainstorm ideas. They create pathways for ideas to get tested quickly, cheaply, and in ways that actually change what happens next. Let me give you a vivid example. In 2012, a Microsoft employee working on Bing had an idea. Change the way ad headlines displayed. It would take just a few days of engineering time, basically nothing in the grand scheme. But it got deprioritized and sat there for more than six months, because it was one idea among hundreds. And then someone did the most startup thing imaginable. They said, the code is cheap, let's just test it. They launched a simple A B test, and within hours it triggered a too good to be true alert, because the new headline variation increased revenue by 12%. That change alone would have been worth over $100 million a year. It became the best revenue generating idea in Bing's history, something that had been invisible inside the way they normally prioritized ideas. That story is so important for universities because it exposes a hidden truth. I believe that we have a lot of hidden value that is discoverable through experimentation. Now, if that were just a cute anecdote, it would stay in the tech folklore bucket. But there's empirical work that says experimentation is a measurable performance lever. A study posted through Harvard University's Dash Repository looks at the adoption of A B testing in about 35,000 global startups with weekly performance measures between 2015 and 2019. Their headline result is pretty direct. After adoption, A B testing leads to about a 10% increase in visits in the first few months. And after a year of experimentation, the gains range from 30% to 100%. That is enormous.

Experimentation That Changes Decisions

SPEAKER_00 6:12

Now I think experimentation is much bigger than what those of us in higher education typically think of. Generally speaking, experimentation is not simply testing how an ad headline is displayed or what button color to use on a website. Experimentation is a way to learn faster after making strategic decisions. It helps us adjust based on what's actually happening. So at this point you might still say, okay, fine, but universities aren't running online ad platforms. We can't A-B test everything. Yes, absolutely. But when I say that experimentation isn't a side activity, the real issue isn't whether we can test something, as if the experiment itself is the main activity. The issue is whether we know how to turn what we learn into decisions that actually change something. Right now we treat experimentation as the work. We run the test and collect the data. But none of that matters if it doesn't move a decision. MIT Sloan School of Management had a piece in 2025 that I think is one of the most useful bridges here. It argues that measurement isn't insight and that leaders should stop framing data as evidence. Your dashboard does not tell us that you're engaged in evidence-based decision making. Unlike raw data, evidence is actually contextual and directional. Evidence is what happens when your data is interpreted in context, tied to a specific problem, and used to point in a direction. The question isn't what can we measure? It actually isn't even what problem are we trying to solve. The question is what would we need to know to make a decision? Even with experiments, even with evidence, there's a deeper organizational problem. Do results actually change what happens next? And that's where another Harvard Business Review idea becomes painfully relevant. In a piece on building a culture of experimentation, the authors basically diagnose what every university committee recognizes in its bones. We're good at asking questions, but not at specifying what would count as a feasible answer. So tests happen, people learn something, and then the organization doesn't move. The results rarely promote organizational change. If you translate that to higher ed, it sounds like this. We're going to launch a pilot advising initiative. Great, what decision will change if it works? Oh, we'll wait and see what we learn. And that's exactly how institutions end up with pilot fatigue. Because the issue isn't that we're running pilots. It's that the pilot isn't tied to a decision. No one has said, if this works, we will scale it. If this works, we will fund it. If this works, we will stop doing something else. So even if the results are strong, nothing actually changes. The pilot doesn't get expanded. It doesn't replace an existing model. It doesn't shift resources or structure. People do the work. They generate the data, and then it just sits there. At that point, pilots stop signaling innovation. They start signaling work that won't go anywhere. And that's when fatigue sets in. Because real experimentation is supposed to create speed. Test, decide, scale, or stop, reallocate, move. But in most cases it becomes test, observe, discuss, delay, move on. So instead of building momentum, the organization builds drag. And over time that has real consequences. Staff time gets consumed without payoff. Resources get spread across initiatives that never fully land. And leaders lose credibility when pilots don't lead to anything. So the institution doesn't develop learning speed, it develops pilot fatigue.

Pilot Fatigue And Lost Credibility

SPEAKER_00 9:50

And if your instinct is, sure, but universities are too complex, too public-facing to work like that, it's worth broadening the lens. This shift toward faster learning isn't just happening in tech. It's happening across sectors. Even government is starting to change how it works. Deloitte has been tracking how even government, arguably the most rule-bound, risk-constrained environment we have, is starting to rethink how it works. And the shift is very specific. Traditionally, government operates the same way many universities do. Long planning cycles, large coordinated projects, decisions made up front, success evaluated at the very end. In some cases, it can take years just to define requirements, and years more to actually procure a solution. That model should sound familiar. It's the same structure that produces pilot fatigue in higher ed. Activity happens up front, learning happens late, and decisions don't move until it's already too late to adapt. What's changing is not that government suddenly values innovation. It's that they're trying to operationalize learning differently. They're experimenting with what Deloitte calls agile procurement, testing ideas through smaller prototypes, iterating in real time, reallocating resources based on what actually works. In other words, they are changing how learning connects to decisions. Which brings us back to higher education. If learning speed is becoming a competitive advantage across the economy, what's the headline we're missing in our industry? Here's the headline. I think higher ed underweights. Speed of organizational learning is becoming a competitive factor, not in some abstract way. CrunchBase reports that in 2025 AI captured close to 50% of all global funding, up from 34% in 2024. And they put a number on it. $202.3 billion invested in AI in 2025. They also report AI funding grew more than 75% year over year versus 2024. Now you don't have to be bullish or bearish on AI to see what this implies. Will the AI bubble burst? I don't know. But what does an investment in AI actually mean? It means an investment in speed. The world outside of higher ed is investing in speed. Technology is changing the speed at which organizations can operate. What used to slow us down, things like coordination, communication, iteration, is getting dramatically easier. It's faster to test something,

Speed Becomes A Competitive Advantage

SPEAKER_00 12:23

faster to adjust it, faster to align people around it. In other words, the cost of learning is dropping, and the pace of learning is increasing. You can see the scale of that shift in where capital is flowing. CrunchBase reports that foundation model companies raised $80 billion in 2025 alone, about 40% of all global AI funding. And just two companies, OpenAI and Anthropic, account for 14% of total venture investment. That tells us something important. A huge amount of money is being concentrated into a small set of technologies. When that happens, those technologies improve very quickly. And they don't stay contained. They show up everywhere, in tools, in workflows, and in how work gets done. You can already see it in small ways. Not long ago, most of our conversations about workplace tech sounded like I hate Microsoft Teams. Now it's Claude Code this, Claude Code that. That shift happened fast. The tools are getting better. The workflows are getting faster. The expectations are changing, and everything downstream moves with it. And you know who's downstream? Us. We're downstream. Deloitte's 2026 Global Human Capital Trends Survey found that seven in ten business leaders say their primary competitive strategy over the next three years is to be fast and nimble, to adapt quickly to changing customer and market needs. That's based on more than 9,000 leaders across 89 countries. This isn't a narrow view. It's the dominant strategy. Organizations are aligning themselves around speed. Now look at what the labor market is signaling to learners. LinkedIn's 2025 WorkCange report leads with a stat that should make our entire industry pause. By 2030, 70% of the skills used in most jobs will change. And people are already responding to that reality. Professionals entering the workforce today are on pace to hold twice as many jobs over their careers compared to just 15 years ago. Since 2022, the rate at which LinkedIn members add new skills has increased by 140%. In other words, learners are not waiting. They are already adapting. They are already behaving as if the environment is shifting under their feet. So now you have both sides of the system moving. Organizations are prioritizing speed. Workers are adapting in real time. And that raises a much bigger question for higher ed. Colleges and universities are built around multi-year cycles. Think of curriculum approvals, program reviews, accreditation windows, the structure of our academic calendars. But the market is moving toward a world where skills, roles, and tools change continuously. So the question becomes: what happens when our pace and the world's pace drift too far apart? If the people doing the work can move faster, then how do we make the systems around them more nimble? That matters for higher education. Because universities are, at their core, coordination machines. We coordinate expertise across departments. We coordinate credentials across disciplines. We coordinate decision making across shared governance. That design made sense for a world where coordination was expensive. But as coordination gets cheaper everywhere else, universities don't just look structured. We start to look slow, not because we're incapable, but because we're optimized for a different set of constraints. Now you might be thinking higher education is a public good. We're in a completely different lane. Yes, and that's exactly why this matters. Because being a public good doesn't insulate you from change. I believe it actually increases our responsibility to respond to it. If skills are changing faster, if careers are becoming less linear, if learners are adapting in real time, then the role of higher education becomes more, not less, critical. This isn't about becoming a startup. It's about whether a public institution can update what it offers, how it supports learners, and how it allocates resources at a pace that actually matches the world those learners are entering. Because if we can't, something else will fill that gap. And we're already seeing what that something else looks like. Employers, alternative learning providers, and even individual content creators are experimenting. They're moving faster. They're building new ways for people to learn and demonstrate skills. Not because they care more than we do or they're more innovative than we are, but because they're able to respond faster to what learners need. And the answer can't be we'll just double down on the traditional model and focus on the students who still want traditional learning. Because there are fewer and fewer of those students. All of us aren't going to eat, because the pie is shrinking. So the question isn't whether higher ed is different. It is. The question is whether being different means being disconnected from the world, or whether it means taking the responsibility of adaptation more seriously than anyone else.

A Dual Operating Model For Higher Ed

SPEAKER_00 17:28

Here's the idea create protected pilot units that can move fast, while the core institution continues to do what it does best. Build cross-functional teams that can run short cycle experiments with clear decision rules, while shared governance remains in place for high stakes, irreversible decisions. In other words, build organizational flexibility. And this works because it respects a core truth. As Minzberg argues, universities are professional bureaucracies for a reason. Expert autonomy and legitimate process are part of the product. You're not changing the mission. You're changing how the institution learns. There's a phrase from Stanford Social Innovation Review that captures the tension perfectly. Tech meets the medieval university. The point isn't that universities are outdated. It's that they were built for a world where authority, expertise, and coordination work differently, and those design choices still shape how decisions get made today. What SSIR proposes is a dual transformation, improve the core, while building new models that can respond to emerging needs. There are already institutions doing versions of this well. You see it in online divisions that were given room to operate differently. You see it in continuing education units that can move faster, because they sit slightly outside traditional structures. You see it in workforce partnerships where institutions had to adapt quickly to employer needs. You see it in small cross-functional teams formed to tackle specific student or program challenges, bringing together advising, institutional research, instruction, and operations to run focused experiments and adjust in real time. Run six to ten week experiments on things that can be tested without compromising academic integrity, like student success interventions, advising workflows, short credentials, and stackable pathways, employer co-designed learning, curriculum sequencing, frictions that students are experiencing. The issue isn't that higher ed can't do this. It's that these models are often isolated. They're treated as exceptions, not as part of a deliberate operating design. What if we stood up a small, cross-functional pilot unit? Faculty, instructional design, institutional research, student services, employer partnerships, whomever, working together in the same system. We talk a lot about innovation in higher education, and on the surface, that sounds exactly right. Of course we should be innovating. Of course we should be trying new things. But if you look more closely at how that actually plays out, what we call innovation is often something much looser. It tends to signal idea generation. Come up with something new, be creative, launch an initiative, see what happens. And in practice that often turns into a familiar pattern. A group comes together, generates a promising idea, and then steps back and says, let's see what we learn. But there's no clear question driving the work, no defined decision on the other side, and no commitment to what will actually change if the results are positive. So the work continues, but nothing really moves. Innovation becomes activity without commitment. What I'm describing instead is a shift in posture from innovation to experimentation, not because innovation is wrong, but because it's incomplete on its own. When the goal is to learn quickly and deliberately, the structure of the work changes. You start with a specific question, not what should we try, but what do we need to know in order to make a decision? You define that decision up front. You make the work time bound, and you design the effort so that the outcome tells you what to do next. So instead of saying let's try a new advising model, you're asking, if we change X in advising, does outcome Y improve enough to justify scaling? That's a different kind of work. It's more focused, more disciplined, and it carries consequence because something has to happen with what you learn. And that's the part that colleges and universities almost always skip. You pre-commit to the decision. You don't just say, let's pilot. You say, if we see X change in outcome Y, we scale. If not, we stop or redesign. If you take that shift seriously from innovation to experimentation, then the next question is obvious. What actually makes

From Innovation To Decision-Driven Experiments

SPEAKER_00 21:50

experimentation work? Because you can't just tell people to test more ideas and expect it to happen. The environment has to support it. And this is where the research gets very important. Very clear. McKinsey's agility survey with more than 2,500 respondents across industries shows one of the starkest differences between bureaucratic and agile environments is the use of minimum viable products. Only 19% of respondents in bureaucratic units said they use MVPs to quickly test ideas. In agile environments, that number jumps to 74%. If we take experimentation seriously, then we have to talk about one concept that tends to make people in higher education deeply uncomfortable. Minimum viable products. I use them regularly. In the work we do inside an online learning unit, we run multiple experiments at any given time. And many of those start as something intentionally minimal. And I'll be honest, this is where the tension shows up immediately. People worry. They worry that we're shortchanging students. They worry that we're putting something into the world that isn't fully built. They worry that we're diminishing a reputation that has been built on quality over decades. And those concerns are real. Because higher education is not designed to produce anything minimum. Everything about how we operate, curriculum approval, shared governance, accreditation, committee review, is built to ensure that what we offer is fully vetted, fully developed, and fully defensible before it reaches a student. That's a feature of the system. But it also means we've designed ourselves out of the ability to test. So when people hear minimum viable product, what they often hear is lower quality, less rigor, something incomplete. But that's not what an MVP is. An MVP is not a lower quality version of the final product. It's the smallest, most ethical, most responsible version of an idea that allows you to answer a specific question. It's scoped, it's intentional, and it's designed to protect students while still allowing the institution to learn. So instead of asking, can we build and approve an entirely new program? You ask, can we test one element of this idea with a small group of students under controlled conditions to see if it actually improves an outcome we care about? That might look like a redesigned advising interaction with a subset of students, a short form credential embedded into an existing structure, a different sequence of courses tested with one cohort, a new employer-informed module piloted in a single section. These are not shortcuts, they are focused tests, and in many ways they are more responsible than what we often do instead. Because right now we frequently make large, irreversible decisions, launching full programs, restructuring offerings, investing significant resources

Minimum Viable Products Without Lowering Standards

SPEAKER_00 24:44

without ever testing the underlying assumptions in a small, contained way. So the real risk isn't the MVP. The real risk is committing at full scale without learning first. And once you see it that way, the role of an MVP shifts. It's not about lowering the bar. It's about learning fast enough to protect both students and the institution from getting it wrong at scale. So when we ask what higher education can learn from the startup world, the answer is not that we should move fast and break things. That's not our role, and it's not our responsibility. What the startup world actually figured out, often out of necessity, is something much more specific. It built organizations that are designed to learn, not occasionally, not informally, but as a core part of how they operate. Ideas don't just get generated, they get tested. Results don't just get observed, they change decisions. Small bets are used to inform larger commitments. And that's the piece that matters for higher education. Because right now we are extraordinarily good at protecting quality once something is built. That is a real strength of the system. But we are much less equipped to learn before we commit at scale. We tend to make large, coordinated, resource-intensive decisions without having tested the underlying assumptions in a contained, deliberate way. In a more stable environment, that trade-off was manageable. But in a world where skills are changing, where careers are less linear, and where learners are already adapting in real time, the cost of getting it wrong and the cost of moving slowly has increased. So this is not a call to abandon shared governance, and it's not a call to lower standards. It's a call to add a missing capability. The ability to run disciplined, ethical, decision-driven experiments that allow the institution to learn before it commits at full scale. And when you think about it that way, this isn't about becoming more like a startup. In many ways, it's about becoming more responsible. It's about reducing the risk of large-scale missteps, surfacing what actually works in practice, and aligning decisions more closely with real outcomes rather than assumptions. The institutions that figure this out will not be the ones that move the fastest in a superficial sense. They will be the ones that can learn in a structured, continuous way while still maintaining the legitimacy, expertise, and public mission that define higher education. Because the question in front of us is not whether change is coming. It already is. The question is whether we build the capacity to engage with that change deliberately or whether we continue to react to it after the fact. And that distinction between reacting and learning is what will shape what higher education becomes next.