A debrief for the store team: the after-payment pages, the link contract, and what the store has to do

One address to hand the store's agent. The four post-sale pages exist and nothing points at them yet, because the pointing is the store's to do and nobody had written down what to point where. After payment↗ is that write-up, on the site rather than in a private note, under More and linked from Admin↗ and from Pricing, after you pay↗. It has five parts: the four pages and what each promises; the link contract, four success addresses and two query parameters, order and, at level 1, shape, with the fifteen slugs the store already uses; the five things the store has to do, in the order they unblock a sale, with the level-3 product-page text ready to paste; what this site guarantees, the stamped manifest, the hash check, no key on any page; and what neither site has settled. Its markdown twin, after-payment.md, is the version an agent reads.

The one that matters most is number two. The pages promise that a person follows up within 24 hours at levels 2, 3 and 4, and today nothing carries a sale from the store to that person: not the level, not the order reference, not the address paid with. The debrief says so plainly and proposes the channel that works today, an email per sale to the follow-up mailbox, and the one that is better once it exists, a vault the store writes into. Until one of them exists, the 24 hours starts without us.

The pages are public now. They were unlisted and marked not for indexing, on the reasoning that a page reached from a payment link is not a page to find. The lead's call is the opposite for now: the aim is a working flow and the first orders, so the four pages are indexable, reachable from Pricing and from the debrief, and easy to hand to anybody. They stay out of the header menu, which is full.

Not done. The store's side of every item in section three, which is the point of the page. The debrief names one mailbox for the follow-up, the same one the pages name; who owns it is an open question in section five, not a decision made here. The link test now ignores a query string when checking that a link's file exists, which it did not need to before.