Which channels produced enrolled, tuition-paying students?

An inquiry in autumn becomes an application in winter, an offer in spring, a matriculation in the following autumn and tuition for three more years. No platform claims any part of that.

Because the match is on the student and there is no attribution window, the campaign that produced the inquiry is credited for the tuition whenever it is paid — which is the figure a provost or a board is actually spending against.

Recruitment fairs, campus visits and school outreach absorb enormous institutional budgets and are reported as attendance figures.

Registration lists go into the lead file, tagged with the fair or open day. Ranked on matriculated tuition, they can finally be compared against digital spend — and the ranking that results may not be the one the budget currently reflects.

A student who has accepted an offer and paid a deposit is not yet a student. Between that point and the census date there is an appeal to another institution, a visa decision, a change in family circumstances, a gap year decided in August, and a number of people who simply never arrive. Institutions measure this carefully and give it a name, and they almost never connect it back to which channel produced the inquiry, so the recruitment report and the enrollment report describe different cohorts.

Using matriculated, fee-liable students as the revenue file is what puts the two back together. With deposits or acceptances as the lead file, each tagged with the channel that recruited the student, the report shows each channel's count beside the students who matriculated, which supplies the second reading: how many of each channel's applicants actually turn up. Those rates can differ, sometimes substantially, and the differences tend to have causes an admissions office can address — how accurately the channel described the course, how far away its audience lives, how much of its volume was speculative in the first place.

None of this requires a change in how the institution records anything. Student records exist because tuition has to be billed, and inquiry records exist because somebody was meant to follow them up. Both were kept for reasons unconnected with measurement, which is precisely why a cycle that closed two years ago can be analyzed this week. For an institution whose planning horizon is a full academic year, being able to look backwards rather than forwards is not a minor convenience.

International recruitment frequently runs through agents who are paid a commission when a student enrolls. That arrangement makes an agent's own reporting a poor basis for evaluating that agent, not because anybody is dishonest but because the incentive and the measurement sit on the same side of the table. An institution paying commission across dozens of agents in a dozen markets has a strong interest in an independent count, and rarely has a practical way of producing one. The two files it already holds are that way.

Each agent's submitted applicant list goes into the lead file with the agent named as the source, and matriculated tuition matches against it on the student's own contact detail. The result is a ranking of agents on enrolled tuition rather than on applications submitted, which is a materially different list wherever application volume has been the thing rewarded. It also makes a specific commission claim checkable: the student either appears on that agent's list or the revenue stays unattributed, and a partial link waits for a person rather than counting itself.

The same treatment applies to the rest of international recruitment, which is rarely only agents. Overseas fairs, partner institutions, articulation agreements with colleges abroad, alumni networks in a particular city. Each produces a list and each can be ranked on the tuition it eventually generated, with currency and phone country set per dataset so that a mixed international file normalizes correctly rather than quietly failing to match half of it. That failure mode is the one nobody notices, because it looks like a poor result.

A university's marketing budget serves markets with almost nothing in common. Undergraduate recruitment runs on a national admissions timetable and speaks to seventeen-year-olds and their parents. Postgraduate taught recruitment runs on its own cycle and speaks to people with careers. Executive and continuing education is effectively a business-to-business sale. Research degrees are a supervision relationship. Ranking channels once across all of that produces a table describing undergraduate recruitment which is then used to fund everything else. The bias is structural rather than occasional, and it grows quietly every planning round.

Adding a program or level column, mapped as Product / service, separates them: the report shows each program's tuition and how much of it traces to a channel, rather than one institutional figure. It does not rank channels within each program, but a program whose tuition sits mostly in Direct / Unknown is one the current channels are not reaching, whatever the institutional ranking says. Seeing them apart lets a faculty argue for spend on evidence about its own market rather than on a share of an institutional figure that was never about it, and it gives the annual argument about who subsidizes whom something to stand on.

Tuition across the whole degree is the other axis worth keeping in view. A student recruited once pays for three or four years, and a channel producing students who persist is worth more than one producing an equal number who leave after the first. Because later years carry the same student contact detail, they credit the original channel automatically, and an institution can compare recruitment spend against the revenue it genuinely generated rather than against a single year's fee. For a four-year degree that is a fourfold difference in what the channel appears to be worth.