Which channels produced paid enrollments?

Channels that produce free signups cheaply are frequently the worst at producing paid enrollments. Optimizing to signups therefore actively moves budget toward the weakest channels.

Ranking on paid revenue is the correction. In this category the two rankings commonly disagree completely, which is exactly why the exercise is worth running.

Affiliate and partner networks self-report conversions, and those figures rarely reconcile with the revenue in your own system.

Adding the affiliate lead export to the lead file, each row tagged with the affiliate, and matching it against your own paid enrollments gives an independent count — possibly the first one the business has ever had.

Most online course businesses offer a money-back period, and a meaningful share of enrollments are reversed inside it. That makes an enrollment count a forward-looking estimate rather than a record, and it makes the estimate worst for exactly the channels that sell hardest. A campaign built on urgency converts well and refunds more; a campaign built on a detailed curriculum page converts more slowly and refunds less. Both look identical on the day the enrollment lands, and only one produces revenue the business keeps.

Exporting net revenue after the refund period corrects this without any extra work; the report counts each row's amount as it stands and does not subtract a refund column. The practical consequence is that a channel report should be run on a period that has fully closed rather than on the current month, which is the opposite of how most teams read a dashboard. A month old enough to have settled tells you considerably more than a live one reporting things that may yet be undone.

Chargebacks sit behind the same door and behave the same way. They arrive later, they cluster around certain kinds of acquisition, and they cost the business in operational time as well as in reversed revenue. Where the payment system can export them they belong in the file, for the same reason returns belong in a retailer's: attributing revenue that came back is the fastest way to make a report untrustworthy, and a report nobody trusts stops being used long before anybody admits it.

A great deal of online tuition is sold on installments. Twelve monthly payments, a three-part plan, a subscription with a minimum term. The headline enrollment value assumes every one of them arrives, and some of them do not: cards expire, plans lapse, learners disappear after the second module. A channel ranked on the contracted value of what it sold is being credited for money that has not turned up yet, and the overstatement is largest for channels selling to people least able to commit.

Exporting payments as they are taken, rather than contracts as they are signed, fixes this cleanly. Each installment is a row carrying the learner's contact detail and its own date, and every one of them credits the channel that produced the learner originally, because the match is on the person and nothing expires in between. The ranking then reflects money received and updates itself as the plans run their course, so a channel that sold well and collected badly can no longer hide behind a signature date.

It also gives a straightforward answer to a question these businesses argue about constantly: whether a longer plan at a higher total beats a shorter one at a lower price. That depends entirely on completion of payment, which varies by acquisition source, and it is answerable from the same two files rather than from a spreadsheet somebody maintains by hand. The answer tends to differ by course as well as by channel, which is exactly what a product column is there for.

Online education runs on lists. A webinar registration, a free lesson sequence, a downloadable workbook, a newsletter somebody else owns, an affiliate's audience, a podcast appearance that ends with a link. Each of those is a group of people with an email address attached, which is the only requirement for a lead source, and combined in the lead file with each row tagged by source, each can be ranked on paid tuition beside the advertising accounts on precisely the same terms. Very few of these businesses have ever seen all of them on one page.

The comparison is usually where the strategy changes. A webinar list that converts a small audience at a high rate can out-earn a campaign with ten times the reach and a fraction of the intent, and the ranking says so in revenue rather than in conversion rate, which is the unit that decides where next month's effort goes. It also shows which lists have aged: an audience that produced enrollments two years ago and produces none now is a specific problem with a specific fix.

Affiliate and partner reporting is where this bites hardest, because those partners report their own results and are paid on them. An independent count from your own enrollment file is not an accusation; it is the same question answered from the other side of the table. Where the numbers agree, a partnership becomes easier to defend internally. Where they do not, the conversation is at least about two specific lists of people rather than about methodology, which is a better argument to have.