Your Payment Deadline Is a Capacity Decision

Airlines worked this out in the 1980s, and the reasoning transfers almost unchanged.

When American Airlines built the system that became yield management — documented by Smith, Leimkuhler and Darrow in Yield Management at American Airlines, the Interfaces paper that won the 1991 Edelman Award — it rested on one premise: inventory was perishable, and not all customers were the same. A seat on Tuesday’s flight stops existing on Tuesday. Everything else followed from taking that seriously.

A seat in a class behaves the same way. Week three of a term happens once. If nobody is sitting in that chair, the chair was not stored for later; it was spent.

Most activity businesses know this instinctively about empty seats. Far fewer apply it to held seats — and that is where the money quietly goes.

A held seat is withdrawn stock

When a family registers and has not yet paid, the seat is usually theirs. It is off the market. It is also earning nothing.

That is a perfectly reasonable arrangement, and the alternative — selling the same seat twice — is worse. The part worth noticing is what the arrangement actually is. The number of days you let an unpaid registration sit is a decision about how much of your inventory you take off sale, and for how long.

Very few people have made that decision on purpose. It usually arrives as a habit — a fortnight, because a fortnight sounds fair — and gets applied to every course, every term, forever.

The same seat does not cost the same in week one and week eight

Here is the asymmetry that makes a flat deadline the wrong instrument.

In the opening weeks of a term, enquiries are live. Trial requests are arriving. Families who have just decided to start something are looking for a place, and they are deciding largely on venue and time rather than on you specifically. A seat held by an unpaid registration in that window is a seat you could genuinely have sold, to somebody who was in front of you, asking.

Eight weeks in, almost nobody is asking. The same held seat costs close to nothing, because there is no queue for it. Holding it for a month harms no one.

So a single company-wide deadline is wrong in both directions at once. Too generous in the window where capacity is genuinely scarce. Needlessly sharp in the window where it is not.

Which is uncomfortable, because the instinct runs the other way: people tend to be most forgiving at the start of term, when everything is busy and chasing feels unkind, and least forgiving later when the term’s finances come into focus.

Working out your own number

Three figures, all of which already exist in your own records.

1. Seat-days withheld. For the opening two or three weeks of your last term, how many registrations were awaiting payment, and how many days did each wait? Multiply them out. The total is the inventory that was off the market.

2. Occupancy over the same window — per venue. This part matters more than it looks. A free seat on the other side of the city does not help a family who chose you because you are near the nursery. Counting capacity company-wide produces a comfortable number that no individual family could have used.

3. Demand over the same window. Trial requests and enquiries during those same two or three weeks.

Put them beside each other. If you had demand while a meaningful block of seat-days was held, the deadline cost you places rather than days. If you had plenty of room and no queue, it cost you nothing at all — and tightening the rule would be solving a problem you do not have.

Both outcomes are useful. The point is to stop guessing.

Ask your own system for it

If your Zooza account is connected to an AI assistant over MCP, it can pull those figures directly rather than making you export anything. A prompt along these lines works:

For the first three weeks of my most recent term, pull three things and put them side by side per venue:

1. Registrations that were awaiting payment, and how many days each waited.
2. Occupancy over that same period — which classes had empty seats, and how many.
3. Trial requests and enquiries over that same period.

Then tell me, per venue: roughly how many seat-days were held by unpaid registrations, and how much demand I had at the same time. Use only figures you can actually retrieve — if something is not available, say so instead of estimating it.

That last line is worth keeping. The numbers are only useful if they are real.

What to change, if anything

If the exercise says your deadline is costing you places, the fix is not a blanket crackdown.

In Zooza the unpaid-cancellation rule is set per course, with its own grace window, and the cancellation is revoked automatically if the family settles inside it. That structure is the useful part here: courses opening into a busy term can run a tighter rule, courses mid-run or habitually half-full can run a looser one. The decision is made per course because the economics differ per course.

A grace window does real work in both directions, too. It gives a family who was simply slow a clear, consistent chance to keep their place — which is a kinder arrangement than an office that cancels when somebody gets round to looking, and more defensible than one that never cancels at all and quietly runs out of seats.

The honest version

None of this is an argument for being hard on families. Plenty of operators look at these three numbers and conclude their current terms are fine, which is a perfectly good outcome and worth knowing.

It is an argument against one specific thing: setting a payment deadline as though it were purely a question of money, when it is also, and sometimes mostly, a question of how much of your term you have taken off sale during the only weeks anyone was trying to buy it.

The broader version of this — how money reaches a class business, and the manual work that builds up around it — is in how many hands touch one payment. For how the pieces are wired, how Zooza wires up payment reconciliation walks the same path as an operator runs it.