Which channels produced paying students?

A student who enrolls for one term at $600 and stays for five is worth $3,000, and a sibling frequently follows. Judged on the first payment, every channel looks barely viable.

Because the match is on the parent, every later term credits the channel that produced them — and the channels producing families who stay are rarely the ones producing the cheapest inquiries.

Enrollment spikes before exams and collapses in summer. Comparing consecutive months tells you about the calendar, not about the marketing.

With no attribution window, the same term can be compared against the same term last year, which is the only comparison that isolates what changed in the marketing.

Almost every tutoring center puts a free diagnostic assessment between the inquiry and the enrollment. A parent calls, a time is booked, the child is tested, and two weeks later a program is recommended and a price is given. That assessment is where the sale is actually made, and it is also the point at which most centers stop measuring, because the marketing report ends at the call and the enrollment report starts at the payment. Nothing in either file describes what happened in between them.

Using the assessment bookings as the lead file is what closes the gap. Each marketing source then carries two numbers: how many assessments it produced and what those assessments became in tuition. One channel may book assessments cheaply and convert them poorly while another books fewer at a higher cost and converts most of them, which would reverse the ranking a center has been managing to for years.

The no-show rate belongs in the same reading. An assessment booked and not attended costs a tutor's time and produces nothing, and it is not evenly distributed across sources. A channel producing parents who were half-interested at eleven at night will fill the diary and empty it again, and the center experiences that as a scheduling problem rather than a marketing one until both files are put side by side. Once they are, it becomes a decision about where the inquiries come from.

Tutoring is usually sold in blocks. A parent buys twenty hours, or a term, or a package running to an examination date, and the money arrives at the start while the teaching happens across the following months. Which date the export carries decides whether the report describes when families committed or when the center delivered, and those two versions of a year can look quite different in a business with a strong seasonal shape. Neither is wrong; using both inside one history is.

For attribution the payment date is generally the more useful one, because it sits closest to the decision the advertising produced. Where a center recognizes revenue as hours are delivered, both can be exported: the payment row credits the channel that produced the family, and a delivery row carrying the same contact detail credits it again as the hours are used. The report answers either question without anybody re-running an export. What it cannot do is reconcile a file that switched convention partway through.

Unused hours are worth tracking too. Packages expire, families move, a child's needs change, and the difference between hours bought and hours taught is a real number in most centers. A channel producing families who buy large packages and use half of them is producing revenue and also producing a refund conversation and a review risk. The report counts the amount you export, so export what was billed and keep delivered hours in your own records beside it; it does not hold the two side by side.

A tutoring center's best sources are frequently people rather than platforms. A head of year who mentions you, a special educational needs coordinator, a pediatrician, the parent committee at a local school, another center that does not teach what you teach. None of these appears in any advertising account, and all of them produce enrolled students who pay tuition for terms at a time. Most centers maintain these relationships attentively and measure them not at all, which is an odd combination to live with.

Each of them is a list, even when it is an informal one. A referral log, the contacts collected at a school open evening, the families from a workshop the center ran at the local library. Added to the lead file, each tagged with its source, they rank on tuition beside the paid campaigns, which does two useful things at once: it shows which relationships genuinely produce, and it stops paid media being credited for students who arrived because a teacher said your name in a meeting.

That second effect matters more than it sounds. Where relationship-sourced students are invisible they fall into the unattributed bucket, and a center concludes that a third of its business comes from nowhere in particular. Where they are uploaded, that share resolves into named sources somebody can actually invest in. The remainder stays in Direct / Unknown, honestly, and it is a smaller and far more interesting number once the relationships have been lifted out of it. That can be the single most useful thing a first report produces.