Delivery slots that customers keep: designing windows your pickers can hit

A delivery slot is a promise, and every broken promise costs more than a late order — it costs the next order from that customer. Most shops set slot capacity by guessing, then spend the first month either turning away bookable revenue because slots are too conservative, or drowning the floor because they are too generous. Slot design is not a marketing decision. It is a capacity-planning decision dressed up as a checkout screen.

The only honest way to size a slot’s capacity is to time your own floor: how many orders can one trained picker fully pick, pack and hand off inside the slot length, given your actual basket size? If a picker reliably clears eight orders an hour on a normal basket, a one-hour slot with two pickers on shift should cap at roughly 14–16 orders, not 20 — the buffer absorbs the substitution conversations and the one order that always goes wrong. Set the cap lower than your best-case number, not your average.

A slot with no cut-off is a slot that fills with orders placed one minute before it starts, which is precisely when the floor has no time left to react. An order-before-time-slot lead time — commonly 30 to 90 minutes depending on basket complexity — gives pickers a real head start and gives you room to close a slot early if the floor is already stretched. This single setting, more than any other, is what stops a good delivery promise turning into a bad one during a Thursday evening rush.

A Friday evening in the UAE is not a Tuesday afternoon, and slot groups that pretend otherwise leave capacity unused on quiet days and oversold on busy ones. Building separate slot groups per weekday — or a specific group for a known peak date such as a public holiday or a promotion — lets a shop widen capacity where the floor can actually absorb it and tighten it everywhere else, with a default group covering anything not explicitly planned.

Even well-sized slots slip — a supplier delivery is late, two pickers call in sick, a delivery-heavy public holiday catches everyone off guard. The fix is not pretending it will not happen; it is giving the shop, the picker or the customer a clean way to move the order to a later slot rather than silently running late against a promise nobody renegotiated. A shopper who is asked, politely and early, to move from a 5–6pm slot to a 6–7pm slot stays a customer. A shopper who watches a slot silently blow past its window does not.

Customers read a full slot as information, not a complaint — it tells them the shop is busy, not broken. A store that shows real, sized capacity earns more trust over time than one that always shows availability and then quietly runs late. Design the slot for what the floor can hit, protect it with lead time, and it becomes one of the few promises in ecommerce that a customer actually believes before they have even placed the order.

Revisit the numbers monthly, not once at launch. A picker floor gets faster as staff learn the layout, a new hire slows it back down, and a seasonal spike changes what "normal" throughput even looks like. Slot capacity that was correctly sized in January is rarely still correct in July — treat it as a live dial, not a setting configured once and forgotten.

Slot fill rate on its own is a vanity number — a slot showing 90% booked tells you nothing about whether the floor can actually clear it. The number worth watching is on-time completion rate per slot, tracked against the picker queue, not against when the order was placed. A slot that consistently finishes its last order 15 minutes past its window is telling you the cap is wrong before a single customer complains, which is a far cheaper signal to catch than a support ticket.