A failed delivery attempt occurs when the carrier reaches or tries to complete the receiving handoff but the shipment cannot be delivered under the required service conditions.
In one line: the route spent time, vehicle capacity and labour, but the receiver promise stayed open.
This is Article 50 in eduKateSG’s 100-article logistics authority build. The canonical parent remains How Logistics Works. Article 49 explained route sequencing. This article explains what happens when the stop is reached but the handoff still fails.
Reader Status and Scope
- Reader job: understand why a failed attempt is not a small local inconvenience but a multiplier of downstream logistics work.
- Mechanism owner: receiver unavailability, access failure, bad address data, failed authentication, unsafe drop, refusal, recovery and redelivery.
- Boundary: this article owns failure of the final handoff. It does not own route design in general or parcel-locker network design.
- Evidence anchor: current last-mile research continues to show that failed delivery attempts increase routing cost and that better customer-availability information, alternative delivery points and recovery options can reduce those failures.
Why a Failed Stop Costs More Than Zero
The vehicle still travelled to the area.
The driver still searched for parking, walked to the address, attempted contact, recorded the exception and carried the parcel away again.
The delivery outcome is still incomplete.
A failed attempt consumes the cost of service without earning the closure of service.
The Failure Multiplier
Attempt → failure → exception record → parcel retention / return → customer communication → re-planning → second route insertion → second handling → second handoff attempt.
One missed handoff can therefore create several new process steps.
Receiver Absence Is Only One Failure Type
- No authorised person is available.
- Address is incomplete or wrong.
- Building access is blocked.
- Security refuses entry.
- Recipient cannot be contacted.
- Payment or verification condition is not satisfied.
- Delivery window is missed.
- Safe-place conditions are not met.
- Receiver refuses damaged or incorrect goods.
- Loading bay or dock cannot accept the shipment.
These causes need different repairs. Treating all of them as “customer not home” hides the real failure pattern.
Bad Address Data Is an Upstream Cause of a Last-Mile Failure
The driver experiences the problem at the endpoint, but the cause may have formed much earlier.
Unit number missing. Postal code incorrect. Building name ambiguous. Contact number stale. Delivery instruction copied incorrectly.
This is why receiver information quality belongs inside logistics reliability rather than being treated as a customer-service detail.
Customer Availability Is a Probability, Not a Binary Permanent Fact
A person may be available at 08:00 and absent at 14:00.
Recent research on customer habit knowledge shows that delivery planning can improve when availability is represented as a time-varying profile rather than a crude morning-or-afternoon choice.
That turns failed delivery from a random surprise into a risk that can sometimes be reduced during route planning.
Time Windows Reduce Failure Only When They Are Useful
A four-hour window can still be too broad for a customer who cannot remain available that long.
A one-hour window can improve synchronization but makes routing harder because the vehicle has less flexibility.
The logistics system therefore trades receiver certainty against route freedom.
Narrower Windows Can Increase Route Cost
If several customers demand narrow overlapping windows, the vehicle may need more travel, waiting or extra routes to satisfy them.
The correct objective is not the narrowest possible window. It is the window narrow enough to improve first-attempt success without making the route infeasible or excessively expensive.
A Failed Attempt Can Damage Every Later Stop
Suppose the driver spends twelve unexpected minutes calling a receiver and trying several entrances.
Those twelve minutes do not disappear. Every downstream stop is now twelve minutes closer to lateness.
A final-mile failure can therefore propagate through the same route even before the redelivery cost is counted.
The Parcel Stays on the Vehicle
A failed parcel continues occupying cube and handling attention after the stop.
It can interfere with the loading sequence, require segregation from completed deliveries and later need return scanning or depot processing.
One failed handoff therefore alters the physical state of the remaining route.
Redelivery Creates a Second Route Demand
The next day already has its own orders.
A failed parcel now enters that future route as inherited work. If failure rates are high, tomorrow’s capacity is partly consumed by yesterday’s unresolved deliveries.
This can create a feedback loop in peaks: more failed attempts → more inherited workload → denser future routes → less slack → more failure.
Recovery Options Change the Cost of Failure
Research on attended home delivery models several recovery paths after synchronization failure:
- Leave at a predefined safe location.
- Redirect to a collection point.
- Attempt a second visit.
- Return to depot for future delivery.
Each option creates a different combination of cost, customer convenience and risk.
Safe-Place Delivery Trades Attendance for Security Risk
Leaving a parcel at an agreed safe place can close the route without waiting for the recipient.
But not every parcel or location is suitable. High-value, age-restricted, temperature-sensitive or signature-required goods may need an attended handoff.
The system should therefore encode when unattended delivery is permissible rather than leaving the decision entirely to driver improvisation.
Collection Points Convert Failure Into Receiver Travel
Redirecting a failed parcel to a nearby pickup point can avoid another home-delivery attempt.
The network saves vehicle work while the receiver assumes a collection journey.
This is not automatically worse or better. Article 52 will test when the exchange is efficient and convenient enough to justify moving the receiving boundary.
Same-Day Reattempt Can Be Cheaper Than Tomorrow
If the receiver becomes available later and the vehicle will pass nearby again, a second visit in the same route can sometimes recover the delivery with less incremental cost than returning the parcel to depot and rebuilding it into tomorrow’s network.
Research on stochastic customer availability has found that second same-day visits can reduce costs under some route structures when balanced carefully against extra travel.
Calling the Customer Has a Cost Too
Phone calls can rescue a delivery, but repeated manual calling adds service time and creates safety issues if drivers interact with devices while operating vehicles.
Pre-arrival notifications, live tracking and controlled communication can reduce the need for improvisation at the door.
ETA Accuracy Can Reduce Synchronization Failure
A receiver who sees a meaningful arrival window can prepare to be present.
But a highly volatile ETA can create the opposite effect: the customer cannot plan around it.
This is why ETA Prediction and failed delivery are connected through behaviour, not merely data display.
Proof of Attempt Must Be Distinct From Proof of Delivery
If the driver reached the correct address and nobody was available, the system should record an attempted delivery state rather than closing the shipment as delivered.
Proof of Delivery owns the successful final state. This article owns the failed boundary that leaves the promise open.
Failure Codes Should Be Diagnostic
“Failed delivery” is too broad for improvement.
- Recipient absent.
- Incorrect address.
- Access denied.
- No safe location.
- Receiver refused.
- Parcel damaged.
- Cash / verification condition failed.
- Driver ran out of route time.
- Building closed.
- Vehicle unable to access.
The categories should be few enough for reliable use and specific enough to point toward different repairs.
Do Not Let Drivers Game Failure Codes
If performance penalties are attached to some reasons but not others, operators may learn to choose the code that protects their metric rather than the code that describes reality.
Failure data is trustworthy only when incentives support truthful classification and evidence can test the event.
First-Attempt Success Is a Better Receiver Metric Than Raw Stops Attempted
A driver can attempt one hundred stops and complete eighty. Another can attempt ninety and complete eighty-eight.
Raw activity makes the first route look busier. Receiver closure makes the second route more effective.
First-attempt delivery success should therefore sit beside route productivity when the service depends on attended handoffs.
Failure Risk Is Not Uniform Across Stops
Historical data can reveal higher failure rates by building type, time of day, recipient profile, access condition or delivery service.
Planning can then use that information without discriminating unfairly: route timing, notification or alternative delivery option can be adjusted to improve the chance of successful receipt.
Failed Delivery Attempts at Three Zoom Levels
One failed stop
Why exactly did the handoff fail, and which recovery path closes it with the least additional system cost?
One route
How much schedule disruption, retained freight and downstream lateness does each failure create?
One network
Which address, availability, access and service-design patterns repeatedly create failure, and can lockers, better information or different windows remove the cause?
A Singapore Lens
Singapore’s dense residential environment can make physical travel short while receiver coordination remains difficult. Security desks, condominium access, HDB blocks, office towers and recipient schedules create different handoff conditions even when addresses sit minutes apart.
The route does not fail because the island is large. It fails when the final receiving condition is not aligned with the delivery attempt.
Hostile Test: “Our Failed Delivery Rate Fell”
Why?
Did more parcels get genuinely delivered first time? Did drivers leave more parcels at risky locations? Did the failure code change? Did customers travel farther to collect? Did the network simply extend promised windows?
A lower failure metric is useful only when the receiver outcome improved without hiding cost elsewhere.
Failed-Delivery Audit
- What exact condition prevented delivery?
- Was the address and contact data correct?
- Was the receiver expected to be present?
- Was the ETA useful enough for preparation?
- How much stop time was consumed before failure?
- What happened to the parcel afterward?
- Was same-day recovery possible?
- Could safe-place or pickup-point delivery have been used?
- Was proof of attempt captured separately from POD?
- Are failure codes truthful and diagnostic?
- What is first-attempt success by stop type and time window?
- How much inherited workload do failed attempts create for future routes?
Evidence and Further Reading
Recent research directly models failed home delivery as a synchronization and routing problem. See The impact of customer habit knowledge on last-mile logistics costs and satisfaction (2026), which examines how availability information can reduce failed home deliveries, and Bonomi et al., Optimizing Attended Home Delivery: Multiple recovery options and customer availability profiles to face synchronization failures (2025).
Return to the Logistics Hub
Failed delivery attempts explain why a missed final handoff multiplies work rather than merely postponing receipt. Return to How Logistics Works for the complete mechanism. Continue next to Delivery Density | Why Close Customers Change Final-Mile Economics.
Final compression: a failed attempt is expensive because the route paid to reach the receiver but did not close the promise. The parcel then becomes future work, and the real repair is to remove the cause of repeated failure rather than merely schedule another visit.