Last-Mile Delivery: Guide | Lily Transport

Judged from the receiving end, good results in last-mile delivery depend on protected cargo condition, an agreed arrival window, usable records and controlled change without unsafe improvisation.

This article follows an illustrative movement of e-commerce orders from Sha Tin to Wan Chai and examines it from the perspective of practical guide. Quantities, timings and sites are hypothetical teaching examples, not a tariff or a claim about an existing customer.

The exact product requirement is defined by the cargo label and shipper for moisture, compression, seal or other handling controls. Vehicle, packaging and labour still need job-specific confirmation. Service scope, vehicle availability and additional arrangements should be verified in a written quotation. Review the distribution and sorting services alongside this article before turning the points below into a booking instruction. (Working case: 262 cartons, 6 stops.)

Plan Backwards From the Receiving Deadline

Consider a midweek replenishment request from an online shop: 262 cartons of e-commerce orders must leave Sha Tin, reach 6 receiving points and make the first post-evening-service appointment. This is a deliberately hypothetical situation used to expose the decisions behind last-mile delivery; it is not a Lily Transport package or a customer case study.

Identify constraints before drawing the route

Dispatch should mark the non-negotiable windows, booked lifts or bays, count-check time at each stop, whether a medium-duty tail-lift truck can enter and how long the product may be exposed during door openings. If the final stop is in Wan Chai, ten minutes lost at every early handover can remove all recovery time later in the day.

The most common last-mile waste comes from bad data, unprepared receivers and small waiting periods accumulating at every stop. Name one person who may authorise a sequence change, and place the route quotation details in the job record so that the driver is not trying to obtain conflicting instructions through several chat groups. (Baseline: 262 cartons, 6 stops.)

Case Walkthrough: Multi-Drop Trade-Offs

This is an illustrative case, not an actual client account. An online shop moves 262 cartons of e-commerce orders from Sha Tin to 6 locations three times a week. Telephone order sequence has been used as route sequence, so the last appointments are regularly missed and nobody can distinguish driving time from receiving delay.

Correct the data before redesigning the route

The team records entrances, height limits, contacts and average service time, then places hard post-evening-service appointments first and clusters the remainder. Loading follows reverse drop order. Arrival, start of unloading and signed completion become three timestamps. The first run tests whether the medium-duty tail-lift truck and workflow fit; it does not guarantee long-term performance. (Working case: 262 cartons, 6 stops.)

Suppose the evidence shows that travel to Wan Chai is stable but the previous site's goods lift adds twenty-five minutes. The response is a new booking or sequence, not pressure on a driver to recover time unsafely. This is why last-mile delivery needs traceable records. The most common last-mile waste comes from bad data, unprepared receivers and small waiting periods accumulating at every stop.

Use the distribution scenarios to re-brief the trial and define punctuality, POD return and exception response as success measures. (Scale: 262 cartons, 6 stops.)

Sequence Matters More Than It Appears

A last-mile delivery job can be managed through seven gates: booking, dispatch confirmation, arrival check, loading, in-transit reporting, acceptance and evidence return. Every gate needs an input, an owner and a completion signal; ambiguity left at one gate becomes delay at the next.

Convert the order into a traceable task

Lock the collection time in Sha Tin and verify that all 262 cartons of e-commerce orders are ready.
Confirm the medium-duty tail-lift truck, driver contact, vehicle entrance and handling equipment.
Record external condition, count and required packaging, seal and cargo condition; load in reverse delivery order.
Use agreed reporting milestones on the way to Wan Chai, reserving alerts for actionable exceptions.
The receiver checks count and condition, recording any variance with photographs at handover.
Return POD, rejected stock and open actions separately so that the job can be closed that day.

Distribution performance depends on address quality, route density, first-attempt success and proof-of-delivery speed, not dispatch speed alone. If no common format exists, adapt the route quotation details into a route SOP, trial it on one repeatable lane and revise it using actual waiting and reporting times. (Scenario: 262 cartons, 6 stops.)

The Logic Behind a Multi-Drop Sequence

Straight-line distance from Sha Tin to Wan Chai is a poor proxy for the working time of last-mile delivery. Tunnel choices, peak traffic, goods lifts in industrial buildings, booked retail bays, loading restrictions and the receiver's checking speed can all reorder a Hong Kong multi-drop route.

Sequence deadlines first, then cluster districts

If three of the 6 points accept goods only during the post-evening-service window, mark them as fixed nodes before grouping the remaining addresses by district and vehicle access. Load in reverse drop order, keeping possible rejects and returned totes where they will not block the next delivery. (Example: 262 cartons, 6 stops.)

Store an entrance image, parking note, floor, goods-lift detail and contact for each address;
measure driving and service time separately instead of hiding waiting inside one total;
trigger an alert before the window is lost so that the receiver hears early;
leave one adjustable stop near the end and state who may approve a change.

The most common last-mile waste comes from bad data, unprepared receivers and small waiting periods accumulating at every stop. The distribution scenarios explains coverage at a high level; daily sequence must still be recalculated from live orders and road conditions.

Total Cost Is More Useful Than the Lowest Price

A quotation for last-mile delivery is not a kilometre calculation between Sha Tin and Wan Chai. Vehicle and temperature band, 6 stops, the post-evening-service window, loading labour, tail-lift use, parking, tunnels, waiting, return cargo and late changes all affect the capacity a provider must reserve.

Break price into comparable components

Base movement: vehicle type, included hours, distance or geographic zone;
Handling: labour, counting, pallets, sorting, packing or additional crew;
Constraints: special equipment preparation, timed deliveries, multiple stops, waiting and overnight work;
Exceptions: extra stop, re-delivery, return, cancellation, overtime and emergency dispatch;
Information: GPS, temperature report, POD turnaround and monthly performance review.

If 262 cartons of e-commerce orders move every week, evaluate monthly route cost together with rejection and re-delivery, not the cheapest advertised trip multiplied by frequency. Store delivery emphasises loading bays and count checks, while residential e-commerce delivery emphasises contact and failed-delivery rules. Use the distribution scenarios to request written charge triggers.

Practices Worth Repeating

The last-mile delivery scenario shows that improvement does not have to start with a new fleet or platform. Better address and cargo data, fewer unauthorised changes and measured service time at each stop make it easier to decide whether the medium-duty tail-lift truck, sequence or provider genuinely needs to change.

Change the four highest-impact controls next

Measure whether the Sha Tin shipment is ready before arrival instead of when loading starts;
give e-commerce orders one version-controlled specification instead of verbal instructions;
separate a repeatedly slow site from occasional congestion and tackle the controllable cause;
assign every exception one owner and due date, then verify the fix before the next dispatch.

Store delivery emphasises loading bays and count checks, while residential e-commerce delivery emphasises contact and failed-delivery rules. Run the revised process on both an ordinary and a busy day. When expanding, use the distribution scenarios and add districts in controlled stages instead of changing every variable together. (Example: 262 cartons, 6 stops.)

Turn the Guide Into Eight Actions

The checklist below can become the internal order form for last-mile delivery. Give every field an owner; mark a genuinely inapplicable item with a reason so that blank space is not mistaken for confirmation.

Eight checks before dispatch

Product, 262 cartons or 2 pallets, dimensions, weight, packaging and stackability;
the product requirement, which is defined by the cargo label and shipper for moisture, compression, seal or other handling controls;
full access, height, bay, lift and contact details for Sha Tin, Wan Chai and intermediate stops;
sequence of 6 stops, the post-evening-service window and expected service time;
actual medium-duty tail-lift truck dimensions, payload, tail lift, pallet truck, restraint and cleanliness;
count, document, photograph, seal or cargo condition and POD return format;
authorised action for waiting, route change, rejection, return, re-delivery and emergency;
scope, charge triggers, cancellation, insurance and measurable trial success criteria.

Distribution performance depends on address quality, route density, first-attempt success and proof-of-delivery speed, not dispatch speed alone. Send the completed list with the distribution and sorting services to each shortlisted provider and require an item-by-item response. Unanswered points belong in the trial plan, not with a driver at the kerb.

Conclusion: Define the Requirement Before Comparing the Service

There is no single configuration of last-mile delivery that suits every business. For e-commerce orders moving from Sha Tin to Wan Chai, a dependable plan comes from accurate cargo data, a vehicle that can enter every site, a workable sequence, clear handover evidence and an exception process that can start immediately. The quotation translates those conditions into capacity and cost.

Review the distribution scenarios, then use the route quotation details to assemble product, quantity, addresses, windows, cargo condition and handling needs. Comparing providers against the same complete brief is a more consistent route to a suitable Hong Kong logistics arrangement. (Working case: 262 cartons, 6 stops.)