Bubble tea fresh-made topping shelf-life and waste SOP - labels carry a "use-by time" not a "made-at time," expired batches quarantined on the spot, logged the same day, and re-labeling to extend life is explicitly banned

You are an operations supervisor for a bubble tea shop.
Write a shelf-life and waste SOP for fresh-made toppings -
one that can be filed and used to train new hires.

Scope (bound it first, so it does not creep):
Fresh-made toppings: pearls/boba, taro balls, pudding,
grass jelly, brewed tea bases, milk foam, cut fresh
fruit. Not covered: unopened dry goods and bottled
syrups (packaging dates govern those; they stay out of
this SOP).

Cover four steps, each stating who, when, the action,
and the record:

Step 1 - label at the moment a batch is done:
- Three lines on the label: item / completion time /
  use-by time
- Use-by time = completion time + the item's hold time,
  taken from our brand manual [blank space for me to
  fill the manual reference]. If the manual does not
  list an item, ask me - do not copy a number from
  another brand or "industry practice."
- Label goes on the front of the storage container
  where it is visible, not on the lid (lids get
  swapped and the label is gone)

Step 2 - expiry checks (two fixed times daily, plus
glances during the rush):
- Fixed times: before the afternoon peak and before
  the evening peak, checking every container's use-by
  time
- Past its time but unfinished: move it into the waste
  quarantine area (physically separate from items in
  use) immediately - it does not stay on the bar to
  "finish this box first"
- Banned behaviors, written as their own bolded lines:
  no pouring off half and re-labeling the rest, no
  merging an expired batch into a fresh one, no
  peeling labels early

Step 3 - waste logging (as it happens, never batched to
month-end):
- Four fields: item / rough quantity (by scoop or box,
  gram precision not required) / reason (expired,
  off appearance, failed batch) / time
- Totaled once at close and reported to the manager
  the same day

Step 4 - opening prep sequence (the hidden driver of
shelf-life waste):
- Principle: longest lead time first (anything cooked
  and rested); short-life items in split batches sized
  to the expected peak, never a full day's worth in
  one go
- For each topping, "how many batches per day, first
  batch at what time" [I fill this in]

Format:
- Two A4 pages max, step headings bold, banned
  behaviors as their own bolded block
- A sample label drawn in a box that can be copied
  as-is
- Ends with "three things a day-one hire must be able
  to do"

Constraints:
- Every hold-time number traces to our manual. The SOP
  must not contain "typically 4 hours" style industry
  defaults - copying another brand's hold time is a
  food-safety hazard, not experience.
- No "raise shelf-life awareness" - every requirement
  lands on an action and a record.
- No new forms or systems: label stickers and the
  existing waste notebook are enough. This SOP must be
  runnable at tomorrow's opening.

Why it is written this way

Printing a use-by time instead of a made-at time (and asking staff to do arithmetic at speed) is the most practical single change here. Nobody does addition mid-rush: looking at "14:20," the worker needs to know whether the bin can still be served, not when it was made minus what plus how long. The completion time is still printed, but only for tracing; the judgment call always reads the use-by time directly.

Hold times must come from this store's own manual, and gaps must be asked about - this is the one place the SOP could cause an incident. Hold times for fresh toppings vary enormously by brand (the same pearls, under different recipes and storage conditions, can carry manual values that differ by a factor of two), and both models and internet articles love to supply "typically 4 hours." Copying another brand's hold time does not save effort; it transfers food-safety liability from the brand manual to a piece of paper.

Re-labeling gets its own bolded ban because it is the most common collapse point of shelf-life systems, and it covers its own tracks: one re-label without consequence and no label in the store is trustworthy anymore - the system exists in name only. Merging an expired batch into a fresh one is the same crime: after the merge, the fresh batch's label is a lie. These bans must live in a forbidden-behaviors block, not folded into positive phrasing where they evaporate.

Waste logged "as it happens" defends against month-end backfilling. Backfilled quantities never reconcile (nobody remembers how much they poured three weeks ago), and that is exactly where count variances leak from. The same-day total to the manager is what lets an anomaly (one item scrapping heavily every day) surface before it becomes a big number. The closing section on opening-prep sequence matters because it controls the source of "expired waste": batching a full day of a short-life topping at once is simply pre-booking half of tomorrow's waste.

Variables