Which phase did that schedule enter?

A customer says the plan changed, or says it did not. Tonight, three proofs, in a console you administer. Name the phase that schedule entered: the schedule id, the status, and the start and the end of the phase. If the schedule has not started, write that, including a null current phase and a null subscription, and do not invent an active phase to fill the hole. Show a different schedule whose price, or whose status, is not this one. Show another account's list that does not contain this schedule. If any of the three needs the person who wired billing, you have not named the phase. You have a roadmap and a guess.

This is the phase on one schedule. It is not the product. Which product did that price sell? is the product id and the unit amount on a price. A price can be right while no schedule stores it, and a schedule can name a price the catalog no longer sells. It is not the status of a subscription that already exists. Which status is that subscription in? is the status on a subscription id. A schedule can sit at not_started while the subscription field is null, and a subscription can be active while no schedule points at it. It is not the trial. Which trial did that customer start? is the trial on a customer. A trial end on a phase can be null while a trial on a different object is set. It is not the app an install granted, and it is not the voice a speech request used. Those are different objects, and they are not this id.

None of what follows is a rate for this product, and none of it is legal advice. The fields below are one processor's current samples of a subscription schedule, and the samples that share an id do not share a created time. The figures below that are a 2026 paper about a music-streaming churn set. A day count on someone else's contracts is not the phase on the schedule you opened tonight. A paper that says its own magnitudes may not generalize cannot name the phase.

Stripe's create-schedule response and Stripe's update-schedule response print the same id and do not print the same created time. The id is sub_sched_1Mr3YdLkdIwHu7ixjop3qtff. 34 characters. The customer on both is cus_NcI8FsMbh0OeFs. 18 characters. The price on both is price_1Mr3YcLkdIwHu7ixYCFhXHNb. 30 characters. object is subscription_schedule. 21 characters. livemode is false on both, so the objects are examples, not a live phase on your account. status is not_started on both. 11 characters. end_behavior is release on both. 7 characters. current_phase is null on both. subscription is null on both. quantity is 1 on both. The create response sets created to 1724058651. The update response sets created to 1680113835. The difference is 43944816 seconds. 508 times 86400 is 43891200, and the remainder is 53616. 53616 is 14 hours, 53 minutes, and 36 seconds. The gap is 508 days, 14 hours, 53 minutes, and 36 seconds. Same id, two created times. Do not paste the update page's created time onto the create response because the id matches (Stripe, Create a schedule, and The Subscription Schedule object).

The phase length is the next disagreement, and the create request is a third number that matches neither. The create response's phase starts at 1787130418 and ends at 1818666418. The difference is 31536000 seconds, which is 365 days exactly. The update response's phase, on the same schedule id, starts at 1680716828 and ends at 1712339228. The difference is 31622400 seconds, which is 366 days exactly. 366 is not 365. The create request on that page sends start_date=1787130418, which matches the create response's phase start, and end_behavior=release, which matches the field, and quantity=1, which matches the item. It also sends duration interval month and interval_count 1. A phase of 365 days is not one month. 30 times 86400 is 2592000. 31536000 divided by 2592000 is 12.166, not 1. 12 times 30 is 360. 365 minus 360 is 5. The page does not print 5, and it does not print 12.166. A request that asked for one month is not a phase you may describe as one month because the id came back.

The start is also not the created time, and the two samples do not share that gap. On the create response, start 1787130418 minus created 1724058651 is 63071767 seconds. 729 times 86400 is 62985600, and the remainder is 86167. 23 times 3600 is 82800. The remainder is 3367 seconds, which is 56 minutes and 7 seconds. The gap is 729 days, 23 hours, 56 minutes, and 7 seconds. On the update response, start 1680716828 minus created 1680113835 is 602993 seconds. 6 times 86400 is 518400. The remainder is 84593. 23 times 3600 is 82800. The remainder is 1793. 29 times 60 is 1740, and the leftover is 53 seconds. The gap is 6 days, 23 hours, 29 minutes, and 53 seconds. 729 days is not 6 days. A future start with status not_started is what the create sample stores. It is not a phase the customer has already entered.

Two more samples use other ids, and neither is the create response. The cancel sample's id is sub_sched_1Mr3owLkdIwHu7ix38CXMudt. 34 characters again, and not the same string. The customer is cus_NcIPFRC981NmaY. 18 characters, not cus_NcI8FsMbh0OeFs. The price is price_1Mr3owLkdIwHu7ix0RyYpQzk. 30 characters, not the create price. status is canceled. created is 1680114846. canceled_at is 1680114847. The difference is 1 second. end_behavior is still release. canceled is not release, and a status of canceled is not an end_behavior of cancel. The phase start is 1680716828, the same start as the update sample, and the phase end is 1712339228, the same end. That phase starts after the cancel. 1680716828 minus 1680114847 is 601981 seconds. 6 times 86400 is 518400. The remainder is 83581. 23 times 3600 is 82800. The remainder is 781 seconds, which is 13 minutes and 1 second. The phase start is 6 days, 23 hours, 13 minutes, and 1 second after canceled_at. A phase that begins after the schedule was canceled is not a phase the customer entered. The release sample is a third id, sub_sched_1Mr3hWLkdIwHu7ixA5zxZvNI, customer cus_NcII9GZkTPAnor, price price_1Mr3hVLkdIwHu7ixWuJp9ew0. status is released. created is 1680114386, and released_at is the same 1680114386. The gap is 0. released_subscription is null. The attribute text says released_subscription is the id of the subscription once managed by the schedule if it is released. The released sample stores null in that field. A null is not a subscription id (Stripe, Cancel a schedule, and Release a schedule).

The enums are longer than the sentence that introduces one of them, and the samples do not walk the whole list. The object page says status values are not_started, active, completed, released, and canceled. That is 5 names. The three samples show not_started, canceled, and released. 3 is not 5. active is not in these samples. completed is not in these samples. The same page's end_behavior sentence says possible values are release or cancel, with the default being release. That sentence names 2. The enum under it lists cancel, none, release, and renew. That list is 4. 4 is not 2. none and renew have no sentence under them on that list. All three samples store release, which is the default, and none of them stores cancel, none, or renew. release as an end_behavior is not the status released. The create page says each customer can have up to 500 active or scheduled subscriptions. 500 is not quantity 1, and 500 is not 365 days. proration_behavior on the samples is create_prorations. 17 characters. The update page's enum for that field is always_invoice, create_prorations, and none. 3 names. The sample uses one of them. billing_cycle_anchor on default_settings is automatic. 9 characters. On the phase it is null. collection_method on default_settings is charge_automatically. 20 characters. On the phase it is null. A default is not the phase. The version header on the sample request is 2026-10-28.endive. 17 characters. A version string is not a schedule id.

The study people reach for when they want a day count of a subscription is a different document, and it is about a music service, not a schedule id. Xiao Han, Yao Xiao, Chenyu Wu, and Tongchen Zhang ask how many days of early behavior suffice for subscription churn prediction, on the public KKBox set. They print a table of PR-AUC, ROC-AUC, and percent of PR gain. The rows they show for the early window are 0, 7, 14, 21, 30, 45, 60, 90, and 120. That is 9 rows. A tenth row, 180, is marked near-decision and its percent cell is a dash. 9 is not 10. At 0 days the PR-AUC is 0.367 and the ROC-AUC is 0.878. At 120 the PR-AUC is 0.456 and the ROC-AUC is 0.901, and they mark the PR gain 100%. 0.456 minus 0.367 is 0.089. The page does not print 0.089. At 7 days the PR-AUC is 0.387. 0.387 minus 0.367 is 0.020. 0.020 divided by 0.089 is 0.2247, which is 22.47%. The cell they print for that row is 23%. 22.47 is not 23. At 14 days, 0.397 minus 0.367 is 0.030, and 0.030 divided by 0.089 is 33.71%. The cell is 33%. 33.71 is not 33. The 120 to 180 step is one they do print: PR 0.520 minus 0.456 is 0.064, and ROC 0.919 minus 0.901 is 0.018. Those two match the sentence. A cell that matches is not a cell that does not (Han, Xiao, Wu, and Zhang, How Early Is Early Enough?).

The same paper's other splits are a different scale from a phase. Manual-renewal users are 28% of users and 86.6% of churners, and the floor on that segment collapses to ROC 0.63. Auto-renewers are 72% of the cohort, with 1.55% churn and a floor ROC of 0.84. 28 plus 72 is 100. That addition matches. 0.63 is not 0.878, and 0.10 of PR lift on the manual segment by day 120 is not the table cell 0.456. The primary design excludes 47% of acquirees. 100 minus 47 is 53. The page does not print 53. They summarize an inflection band of about 45 to 90 days, and they gloss it as six weeks to three months. 6 times 7 is 42. 45 is not 42. The page does not print 42. The abstract says the magnitudes may not generalize past one music-streaming dataset. A day count on KKBox is not sub_sched_1Mr3YdLkdIwHu7ixjop3qtff, and 0.367 is not a phase start.

Do the three proofs on one real schedule, in the console that actually holds it. First, the binding. Copy the schedule id, the status, and the phase start and end. If current_phase is null and subscription is null and status is not_started, write that. Do not infer a 365-day phase because a sample prints it, and do not infer quantity 1 because a sample item prints it. If the request asked for one month and the stored end is 365 days later, write both. Do not repair the pair. Second, the other schedule. A different id, with a different price or a different status, is not this one. The cancel id and the release id are that shape on the sample pages, and they are still samples. A chart of 9 windows on a music service is not that second schedule. If you cannot show a schedule that is not this phase, you have a counter, not a record. Third, the other account. A list that belongs to someone else does not contain this schedule id. If the only list is a single shared table and you cannot say whose phase it is, write that. A tenant you cannot name is the finding. It is not a reason to tell the buyer the plan already changed.

Three moves feel like the proof and are a second problem. Copying the month off the pricing page because the phase length was embarrassing, and then describing the missing id as a cleanup. Quoting 23% because it looks precise, while the schedule you were asked about is one row, and the study that printed 23% does not equal 0.020 divided by 0.089. Asking a model which phase the customer entered. The model will pick a plausible date. It will not open the schedule. A product id you have not kept, while the customer is arguing about the plan name, is which product did that price sell?. Opening that screen does not name the phase.

The path is ordinary when an operator can open one merchant's wallet without opening the next merchant's wallet as the way to prove it. PayOrch is a payment orchestration dashboard: wallets, providers, merchants, fees, and fraud controls in one view. The shape that matters here is one merchant you can point at, in an account you administer, distinct from the next merchant. That case is not a schedule ledger. Nothing in the description is a headcount, a duration, or a result figure. If the only control on the screen shows a day's volume and will not name a schedule id, you do not have this proof. You have a chart.

There is a version of this afternoon where hiring the path is the wrong purchase. Nothing is for sale, because the product does not take money yet. You have no product. You are still choosing who will build, and the document in front of you is a term sheet. The criteria for that choice sit on the page written for a top SaaS development company. That page is criteria. It does not rank anyone, including us. The only fact in the room is a count of days on someone else's music subscriptions. 45 is not your phase, and 0.456 is not your phase, and 365 on a sample is not your phase. A customer has already seen another customer's catalog. Naming an id does not close a leak. Or you wanted a lawyer to say whether a phase is a contract. Hire the lawyer.

When the schedule is real and the console cannot name the phase, the missing object is specific. An admin action, in an account you administer, that shows one schedule id, one status, and one phase start and end or an explicit not-started line, shows a different schedule that is not that phase, and shows another account's list without it. A startup software development company that expects this will have put that action next to the rest of the billing screen, before a stranger's complaint is the way you discover the vendor owns the phase. One that needs a call in order for you to see which phase is stored is telling you who holds the dates. If the first slice is a fixed capability, an AI integration sprint is the shape where the acceptance sentence can include the three proofs. If you do not yet know which system stores the schedule, a two-week AI readiness assessment maps the systems. The findings are yours. They are not a promise that a paper's day count applies, and they are not 365.

The app an installation granted, and the voice a speech request used, are different objects: which app did that install grant? and which voice did that speech use?. Do not record either on this page. An installation id is not a schedule id. A voice name is not a phase end.

Before you tell them the plan already changed, do the three proofs and write the id. Name the schedule, the status, and the phase start and end, or write that it has not started. Show a different schedule that is not that phase. Show another account's list that does not hold it. Do not paste a sample month onto a 365-day span, and do not ask a model which of the dates it was. A paper can tell you what it counted on other people's contracts, including the percent that does not match its own subtraction. The binding is the phase you can defend.