The one-page version: 35 days out, 12 of them in transit, 7 in sourcing
Start at the date and work backwards, because every figure in this plan is measured from the warehouse rather than from the calendar. Our recorded median from release to delivery is 12 days, with a worst case of 27 (Recorded, 24 parcels, window 2026-08-01 to 2026-09-30).
Add the sourcing week and the answer is 35 days before the date for a comfortable run, and 27 days before it for a run that accepts the worst case we have recorded.
1. Write the date the parcel must arrive. 2. Subtract 12 days for the median journey. 3. Subtract a further 7 days for sourcing. 4. Release the order on or before that day.
Two dates matter rather than one: the date the parcel has to arrive, and the date the recipient stops being at that address. A present that lands two days late at a family home is a shrug, while the same parcel arriving after a house move is a return and a second delivery, which is a far worse outcome than a late one.
Block one, sourcing: the seven days you can still move
Sourcing is the stretch where a seller controls the clock. Domestic freight to the warehouse takes two to five days depending on volume, and a relist request can add another three without anyone calling it a delay.
Order the slow lines first. A seller who needs five days should be paid on day one of the window, while a seller who ships next day can wait until day four without moving the arrival date.
Watch the domestic leg for a repeat. Where the same seller took four days on a previous order, plan four again and keep the tracking number in the note.
Sourcing has a weekend shape that a day count hides. A seller who confirms on a Friday and posts on the Monday has spent three calendar days without anything moving, and three of those weekends inside a 35-day plan is nine days gone before the warehouse has seen the parcel.
A seller who confirms late is not always a seller who ships late. Where a confirmation arrives on day five and dispatch follows the same afternoon, the plan has lost two days and the transit has lost none, so the figure to correct is the sourcing allowance rather than the delivery estimate you started with.
Block two, warehouse and QC: the ten days that absorb a retake
The warehouse leg carries a median of 2 days in our record, which covers shelving, QC photography and the packing queue in one figure. Ten days at this stage is generous, and generous is the point of a buffer.
Review the QC photographs inside this stretch rather than after it. A retake request needs a day to be actioned, and a request sent on the day before the parcel is packed becomes a cancellation question instead.
1. Read the QC photographs for each item. 2. Request a retake on anything unclear. 3. Confirm the retake before the packing queue starts. 4. Note the date the parcel was marked outbound.
QC photography is the one stretch where a delay is worth accepting. A retake adds a day to the warehouse leg and removes the risk of shipping a wrong size, while a wrong size returned after a birthday costs a second parcel, a second opening rung and another week rather than another day.
Block three, the trunk leg: nine days nobody can influence
The trunk leg carries a median of 9 days from the outbound mark to the first UK scan, and it is the longest and least controllable stretch in the whole plan. Nothing you do in this window changes it.
Use the time for the parts that are still yours: the address, the alerts and the payment method. An address corrected now costs nothing; the same correction after packing becomes a redelivery of one to three days.
Expect the tracking page to lag here. Two of our 24 parcels showed no event for three days and then jumped three stages at once (Recorded, 24 parcels), and that pattern is normal rather than alarming.
Trunk days move with the line rather than with the parcel. Our recorded median of 9 days covers the four services in the set, and a slower service can add three days without anything going wrong, so the line chosen at the payment step is worth more attention than the tracking page checked afterwards.
Block four, the last mile: one to three days and a fallback
The last mile runs one to three days in our record, with three days typical when a first attempt fails and a redelivery is booked. A signature requirement is the most common reason a first attempt fails at a weekday address.
1. Confirm someone is at the address on the expected day. 2. Give the recipient the tracking number. 3. Agree a neighbour plan in advance. 4. Keep the phone reachable on the delivery day.
Days five to zero are also the window where a gift decision can still change. Where the parcel is clearly going to miss, a small item bought locally covers the date and the parcel becomes a second gift rather than a failure.
Agree the fallback before the delivery day rather than after it. A neighbour arrangement, a safe place or a collection point each removes the three-day redelivery penalty, and every one of those arrangements has to exist before the first attempt fails, because the courier will not create it at the door.
When the one-page version is enough, and when to send a card instead
Split the gift when the date is the point rather than the parcel. A card sent on the day plus a parcel arriving a week later satisfies a birthday in a way that a single late parcel does not, and the choice costs one card against the risk of a present that turns up after the occasion has passed with nothing to mark it.
A parcel is the wrong answer when the date is inside 21 days, when the item is still with a seller, and when the recipient would rather choose. Three conditions in a row is a strong signal to stop planning around freight.
The arithmetic supports that. A 12-day median plus a sourcing week leaves nothing for a queue window, and the worst case in our record, 27 days, would miss the date on its own (Recorded, 24 parcels).
Decide early rather than late. A decision taken at T-30 leaves every option open, while the same decision at T-7 leaves only the express version of a plan that was never going to fit.
Where a card is the answer, say so early rather than late. A card chosen at T-30 can be written, posted or handed over on the day itself, while the same decision taken at T-3 leaves a queue and a hope, and the parcel turns up afterwards to no particular occasion at all.
Choose the fallback at the same moment you place the order rather than at T-3. A card bought on day one and held back covers a missed date without a second parcel, and a fallback decided in advance is a decision, while the same fallback chosen on the day is a scramble in a queue with the date already gone.
The data point behind this note
Counting backwards from a fixed date, our recorded median of 12 days plus a 7-day sourcing week puts the latest comfortable release at 35 days before the date, against a worst recorded run of 27 days (Recorded, 24 parcels).
Rates last checked 2026-10-02. Where a figure is community-reported we say so; where we could not verify it, we write Not verified instead of estimating.