Context
Over nine years I held marketing leadership roles at three unrelated educational organizations: two universities and a multi-brand online education company. Different sizes, different students, different decades of technology. The centerpiece of this case is the university where I directed digital marketing and an annual lead generation budget exceeding $1M.
What Looked Broken
At each organization, the stated problem was some version of "we need to improve enrollment marketing." More leads. Better campaigns. Stronger recruitment.
What I Noticed
All three organizations knew exactly who enrolled. Enrollment was the finish line, and the finish line was well photographed.
What nobody tracked was the middle: what happened to everyone who inquired and didn't enroll. Close-lost prospects weren't analyzed. They weren't re-engaged. Nobody asked why they left. They just became extra rows in the CRM.
The same blindness, three times, in three organizations that had never met each other.
What the Evidence Showed
At the university, the only tangible marketing metrics were web traffic and button clicks. Lead data lived in at least five disconnected systems: the admissions CRM, two outsourced program-management vendors with their own CRMs and microsites, an athletics CRM, and the student information system, which didn't sync enrollment status back to the CRM marketing could see. Vendors ran ads to the same prospects without knowing what the others were doing, which meant the university was sometimes bidding against itself for the same student.
The data to answer the funnel questions existed. It lived in CSV exports that had never been compared.
The Actual Constraint
The institution could see the finish line but not the journey to it. This wasn't an absence of data. It was an absence of connected data, comparable data, and trusted definitions. You cannot optimize acquisition when you cannot observe what happens after acquisition.
My Hypothesis
If I could reconstruct even an approximate funnel from the available exports, we could work backward from the enrollment goal to the required inquiries per program, set defensible cost-per-lead targets, and stop treating marketing spend as an act of faith.
What I Architected
- A common language: written definitions for inquiry, stealth applicant, applicant, completed file, admit, confirmed, matriculant, so five departments could stop meaning five different things by "lead."
- A seven-stage funnel model with per-channel conversion rates, cost per lead, cost per enrollment, and potential revenue per marketing effort. [ILLUSTRATIVE: the published funnel matrix was designed as the report the university should have been able to run; figures in the example matrix are illustrative. See verification notes.]
- A lead generation calculator and budget calculator that turned the enrollment goal into required inquiries per program per term (roughly 100 per program per term) and reset the cost-per-action benchmark from the ~$250 the institution had been paying toward the $50–100 industry median.
- A written diagnosis-and-recommendations strategy for university leadership: centralized digital oversight, CRM/SIS synchronization, and a marketing automation platform. I wrote the strategy; my Director of Digital Services contributed.
- The FY media architecture built on the model: brand awareness and lead generation as distinct objectives with distinct funnel effects, instead of one undifferentiated ad budget.
What I Personally Built / Did
I exported the CSVs from Salesforce myself, joined and compared them myself, and built the model, the calculators, and the strategy document myself. I presented the work to admissions leadership, academic deans, and my VP.
What Happened
The people closest to enrollment recognized it immediately. Undergraduate admissions used it. The deans engaged with it. My VP backed it.
Then it got political. My direct manager removed me from the funnel work and reassigned me. I believe the university's enrollment marketing agency worked from the model afterward, but I can't verify that, so I won't claim it.
The structural recommendations survived the politics: the university purchased and onboarded the marketing automation platform I proposed, and we began building program-specific landing pages matched to genuinely different audiences (an MBA prospect and an EdD prospect are not the same person, and for the first time the infrastructure could treat them differently).
What Became Possible
For the first time, someone could ask "how many inquiries does this program need to hit its enrollment goal, and what should each one cost?" and get an answer derived from evidence instead of tradition.
What I Learned
Visibility work is threatening in a way dashboards never warn you about. When nobody can see the funnel, everybody gets to have their own version of what's working. A model that connects spend to enrollment takes those versions away, and not everyone experiences that as a gift.
Evidence / Artifacts
Sanitized funnel matrix, lead generation calculator, budget calculator, and strategy excerpts. All redacted to remove institution, vendor, and personnel names. Illustrative figures labeled as such.
Related Work
The same blindness appears at the other two organizations in this case: at one, I answered it by refusing to buy leads (Built, Not Bought); at the other, by building the company's first CRM (The Company That Had Never Had a CRM). Eventually the pattern got a name: Visibility Debt, and a diagnostic (the Revenue Health Matrix).