A few weeks ago, a senior leader at one of our R1 partners—a dean of professional education at a flagship public university—told me she gets the same question in every internal meeting.
"Why don’t we just build this ourselves?"
She's not annoyed by the question. She thinks it's the right one to ask. But she also thinks most of her colleagues are answering the wrong version of it.
Here's how she puts it: The question almost everyone reaches for is, “Can we build a course?” And the honest answer at most flagship R1s is yes, obviously, we can. We have the faculty, we have the LMS, we have the brand.
But that's not the question being asked. The actual question is, “Can we operate a micro-credential program at the cadence the market demands?” That's a much harder yes.
The data here is uncomfortable. UPCEA and Modern Campus's 2026 Institutional Perspectives on Microcredentials found that institutional adoption of credential innovation has essentially flattened, as53% of institutions say they've embraced credentialing initiatives in 2025, nearly identical to the 54% who said the same in 2021. Five years of strategic plans, task forces, advisory committees, and RFPs, yet the needle barely moved. The analysis attributes the stall to “limited resources and legacy systems,” which is institutional shorthand for we know what to build, but we can't figure out how to operate it.
Then the dean did something more useful. She walked me through the framework she actually uses internally: 12 questions she runs her team through, organized into three buckets. With her permission, I want to share her framework. You can run this diagnostic yourself, before your next steering committee.
Bucket A—Operational infrastructure: Who does the work?
The first six questions in this framework are the unglamorous ones. They're also where most internal builds quietly stall.
1. Who recruits and hires the instructors? Adjunct pools are a different animal from industry practitioners who can teach a current-state AI course on Wednesday evenings. The recruiting motion, the comp band, the speed of hiring—none of it looks like academic recruiting.
2. Who pays them, and where does that cost sit on the org chart? This is the question that ends more in-house builds than any other. The UPCEA 2024 Staffing & Structure Survey found that 44% of online and PCE units are housed within the provost's office, with the rest scattered across colleges, standalones, or hybrid configurations. Such revenue-sharing arrangements with academic units are a patchwork. Continuing ed budgets aren't sized for industry-practitioner pay. The general fund won't absorb it. This math is a problem most institutions haven't solved.
3. Who manages HR, performance, and offboarding for non-faculty instructors? Faculty governance doesn't cover them. HR isn't sure they're employees. When a student complains, who handles it, and on what timeline?
4. Who does QA against the certification or industry standard? PMP, Six Sigma, AWS, HRCI—each one has standards that shift. Who owns mapping your curriculum to them, and more importantly, re-mapping when the standards shift? This is also a regulatory question now. The Higher Learning Commission and NECHE are both launching microcredential endorsement processes, partly because, as HLC's president put it, accreditors can't yet confirm that the operating shops behind many micro-credentials have the financial data or outcomes metrics to back them up.
5. Who owns registration, refunds, and ADA compliance for non-matriculated learners? The systems your registrar runs weren't built for someone who isn't pursuing a degree. Refund policy, drop windows, accommodations—every one of them needs a parallel track.
6. Who handles complaints about instructor quality or course content? Not “if.” When. And the answer can't be “the dean.”
Notice the pattern here: Each question is solvable, just as each role can be identified and filled. Combined, these roles form a cross-departmental operations team that functions as a small company within the university.
Bucket B—Instructional design: Can we design for a different learner?
The next four questions move from operations to design. They're the ones faculty colleagues are most likely to push back on, because they sound like an attack on the academic core. They're not. They're about a different product.
7. Can we design for online delivery, not just port in-person teaching? Recorded lectures aren't online learning. They're recorded lectures. Designing for the medium is a different craft, and it requires instructional designers who've done it at scale.
8. Can we build in shorter learning units instead of semester-length courses? A working adult isn't going to give you 15 weeks. They'll give you six to 10, in stackable pieces, with clear outcomes at each step. The fastest-growing segment of U.S. higher education right now is the undergraduate certificate, up 4.8% in spring 2025 alone and 20% above pre-pandemic levels. That's not a niche. That's the market telling us what shape it wants. Clearly, they want to start exactly at the time they want to.
9. Can we serve a non-matriculated, non-degree-seeking learner? Everything in the university's machinery is built around the matriculated student: the advising model, the calendar, the financial aid plumbing, the success metrics. A non-degree learner sits outside almost every system you have, and that learner is showing up. The National Student Clearinghouse counts 41.9 million Americans with Some College, No Credential, and the share of older students (25+) in fall 2024 enrollment grew nearly 20% year over year—almost six times the rate of traditional-age growth.
10. Can we refresh the curriculum regularly, sometimes every few weeks, when the topic demands it? This is the question that broke the most builds in the last 18months. Lightcast's analysis of 1.3 billion job postings found that 32% of the skills required for the average job changed between 2021 and 2024. This assertion from World Economic Forum (2025) is one I keep coming back to: “The half-life of skills is around four years. In digital fields like AI, it’s probably even closer to two years.”
In AI specifically, that compresses further. Over six days in November 2025, four flagship models—GPT-5.1, Grok 4.1, Gemini 3 Pro, and Claude Opus 4.5—shipped within a week of each other, and material capability updates have landed every two to four months since. If your course- update process runs annually, you're already behind on day one.
Bucket C—Institutional readiness: Is the org actually built for this?
The last two questions are the strategic ones. They're also the ones that surface latest, usually after the budget is already spent.
11. Has leadership internalized that this learner isn’t the 18–22 residential student? Every operating decision flows from this. If the cabinet is still imagining the same student walking across the quad, the program will be designed for that student, and ignored by the actual adult learner. WICHE’s Knocking at the College Door report projects that the U.S. high school graduate population peaked at 3.9 million in 2025 and will decline through 2041. The strategic hedge against that decline? The adult learner. But adult learners, especially those who already have a degree, don’t buy what most universities are currently selling.
12. Does the faculty/governance model allow rapid iteration without a year-long approval cycle? Curriculum committees were designed to protect academic rigor. They were not designed for a product that needs to ship a new unit in two weeks because the standard shifted. Both things are true. Both have to coexist. And the published governance calendars at most R1 publics confirm what every dean already knows: A substantive curricular change usually takes 12–18 months from proposal to first enrollment.
Four shifts, one operating model
If you run these 12 questions, you’ll notice a pattern. What looks like one program is actually four shifts at once. Here’s the shape of it:
Modality shift: In-person to human-supported online. Not async, not self-paced, but a live-cohort experience with practitioners and dedicated support that happens to be digital. Most institutions treat this as a content-delivery problem. The ones that win build an actual learning product, one engineered around how adults retain, apply, and stick with new skills. That's not a distribution change. It's a product change.
Length shift: Semester to short-form, stackable
Learner shift: Degree-seeker to working adult
Cadence shift: Annual revision to continuous iteration
Frequency shift: From fixed semester starts to start when they need to learn or build skills, meeting working adults at the moment the need is real, not the moment the calendar allows.
Each shift is doable in isolation. Most institutions have already done one or two. Compounded, they're a new operating model. This is the part the "let's just build it" conversation usually misses.
What this isn’t an argument for
I want to be careful here, because the easy read of all of this is “you can’t do it, hire us.” That isn’t the argument.
Several universities have built micro-credential operations internally. Some of them are excellent. The honest version of this post is that most institutions underestimate what “building it” actually means once you’re in market and managing the rate of change we’re seeing in the workforce—and a few have built the operating muscle to absorb that change. Both can be true.
But there’s a number from this past summer I can’t get out of my head. Inside Higher Ed reported in August 2025, citing Validated Insights, that roughly 57 million Americans are interested in AI skills and 8.7 million are actively learning them—but only about 7,000 are doing so through a credit-bearing higher-ed program. Workforce aligned solutions are serving more than 99% of that demand. YouTube tutorials, LinkedIn Learning, EdTech companies like Ziplines, and millions of working professionals who are stitching it together on Reddit and Discord on Sunday nights. That's not a criticism of those channels. It's where the demand went when no one else built for it at the scale and cadence it needed. And it's the opportunity sitting in front of every flagship R1 right now — if the operating model can support it.
What I'd push back on, in any internal conversation, is the version of "build vs. partner" that gets answered in 20 minutes by a committee that's never run a non-credit P&L.
The reframe
Let's revisit that one question that routinely pops up during internal ProEd meetings: "Why not just build it ourselves?"
When I asked the dean to share her thoughts on this, her response perfectly reframed the question in a way that's applicable to every higher ed institution:
“The real question isn’t whether a university can build a micro-credential internally. It’s whether it can operate one well.”
If you’re running this calculus inside your own institution right now, two things:
One: Run the 12 questions above with your team before the next steering committee. Not as a vendor sales tool. As an internal diagnostic. Most of the value is in watching where your team gets quiet.
Two: If it’s useful, I’ve collected how a few of our R1 partners actually answered these questions, in their own words. Happy to share.
— Clayton
