Why enrollment is hard to run today
1. Every call runs on a partial picture
Each system records transactions, so simple questions take a week of exports.
2. Students wait weeks while their files sit in a queue
Applications, transcripts, and transfer credit are keyed and checked by hand.
3. A drop in applications can be a gain
Without the market backdrop, teams chase the wrong problem.
You gained share in a market that shrank
Illustrative
2 · The ontology we build
The enrollment model
We bring your systems, outside data, and policies together, model the domain, and compute what it runs on.
A slice of the full model: the labeled objects are the ones this work runs on most.
Segment
One population in the entering class, such as first-year, transfer, adult, or international students, with its own target.
- population
- enum
- residency
- enum
- target
- integer
- makes upEntering classone
- includesProspectmany
Entering class
The class the institution is building for one term, with its goal and its date of record.
- term
- term
- goal
- integer
- census_date
- date
- class_forecast
- Census enrollment for each segment, with a range, refreshed daily. From prospect, application, and market.
- is divided intoSegmentmany
Market
A region the institution recruits from: its graduates, how many go to college, and where they enroll.
- region
- geo
- hs_graduates
- integer
- college_going_rate
- percent
- Æ’market_share
enrolled ÷ (hs_graduates × college_going_rate)
- is home toProspectmany
Campaign
A recruiting tactic aimed at an audience: travel, events, mail, messages, or calls.
- tactic
- enum
- audience
- enum
- cost
- money
- Æ’added_deposit_cost
cost ÷ (tested_by.measured_lift × prospects reached)
- deliversTouchpointmany
- is tested byExperimentmany
Touchpoint
One interaction with one prospect: a visit, a fair, a call, an event, or a message they opened.
- channel
- enum
- occurred_on
- date
- responded
- boolean
- is delivered byCampaignone
- reachesProspectone
Prospect
A person considering the institution, from first inquiry through census.
- stage
- enum
- first_source
- enum
- intended_major
- string
- likely_to_enroll
- Chance this prospect enrolls by census. From touchpoint, milestone, application, and market.
- next_best_step
- The touch most likely to work next. From experiment, touchpoint, and milestone.
- Æ’behind_pace
milestones completed < expected_by(stage, days_to_census)
- belongs toSegmentone
- lives inMarketone
- is reached byTouchpointmany
- is assigned toExperimentmany
- submitsApplicationmany
- completesMilestonemany
Application
One application for one entry term, with its requirements and its decision.
- entry_term
- term
- submitted_on
- date
- decision
- enum
- Æ’missing_items
required_items(program, segment) − received_items
- is submitted byProspectone
- includesTranscriptmany
Transcript
An academic record from a prior school or college, read line by line.
- institution
- string
- received_on
- date
- gpa
- decimal
- Æ’transfer_credits
sum(prior_courses where grade ≥ C → counts_as.credits)
- is part ofApplicationone
- listsPrior coursemany
Experiment
A test of a tactic against a holdout or matched group, followed through census.
- hypothesis
- text
- metric
- enum
- ends_on
- date
- Æ’measured_lift
deposit_rate(treated) − deposit_rate(holdout)
- testsCampaignone
- assignsProspectmany
Milestone
A step that predicts enrolling: a campus visit, a completed file, admitted student day, orientation sign-up.
- kind
- enum
- due_by
- date
- completed_on
- date
- is completed byProspectone
Course
A course in the institution's catalog that coursework from elsewhere can count as.
- subject
- string
- number
- string
- credits
- decimal
- is credited fromPrior coursemany
Prior course
A course the student took somewhere else, as it appears on a transcript.
- title
- string
- credits
- decimal
- grade
- string
- is listed onTranscriptone
- counts asCourseone
What it enables
3 · What we deploy on top
Applications, agents, and workflows on the model
Each one reads the model and changes it only through governed actions.
Enrollment HQ
Where the enrollment team runs the cycle: the class forecast against goal by segment, every market’s share, prospects behind pace, and live experiments.
Release decision: Director signs off on exceptions to policy
Recruiting agent
Works the prospects who are behind pace. It sends the next best step within approved messages and logs every touch back to the model.
Counselors approve the messages it can send
Application, transcript, and transfer credit processing
Reads every application and transcript, matches prior courses to equivalencies, and computes transfer credit. Exceptions go to an evaluator.
Approve equivalency: Department chair for any new equivalency; Release decision: Director signs off on exceptions to policy
4 · Outcomes
What the enrollment model lets you do
1. Forecast against the market backdrop
Share shows next to raw counts, so down 5% in a market that fell 12% reads as a gain
Enrollment leadershipWhat the model sees
Your applications and deposits set against each market’s graduates, college-goers, and where they enroll
2. Process every application, transcript, and transfer credit evaluation
Decisions go out faster, and every student gets an academic index that feeds D, F, and W forecasts downstream
Admissions and registrarWhat the model sees
Every application, transcript, and course, read and structured
3. See who is falling behind, and the best next step
Your team or agents get the next step that matters most for each student
Your team decidesWhat the model sees
Each student’s progress against the milestones that predict enrolling
4. Run experiments across the funnel and measure them
Every campaign, message, or outreach change shows its real impact
Enrollment leadershipWhat the model sees
Holdouts and matched groups, tracked through census
5. Learn which recruiting tactics work, and feed them back
What works becomes part of the model and next cycle’s plans
Enrollment leadershipWhat the model sees
The visits, fairs, messages, and campaigns that moved students, by segment
Projected value when a change goes live, and realized value against a holdout, a matched group, or the prior cycle, reported separately.
Net tuition per added student, every year they stay
Projected net tuition against goal
Likelihood to enroll when a file is complete within 14 days
Start with Enrollment
We scope the first domain with your team and agree on the outcomes to measure.
