Real-Time Visibility Ends at Handover: Tracking the COD Cash Chain in Southeast Asia

The delivery is complete and the money has not moved

It is 19:40 in a Jakarta control room. The dashboard shows 4,200 of 4,300 orders delivered, a 97.7% completion rate, and one hundred exceptions already triaged and scheduled for reattempt. By every measure the operation had a good day.

Sixty-eight per cent of those orders were cash on delivery. So roughly 2,900 collections happened today, and not one rupiah of that money is in a bank account. It is distributed across 180 rider bags, several hub safes, and a remittance process that will close over the next two to three days.

The dashboard has nothing to say about any of it. It was built to answer where the parcel is, and the parcel has arrived.

This is the structural blind spot in real-time visibility across Southeast Asia. High cash-on-delivery penetration is one of the features that genuinely distinguishes SEA markets from Western e-commerce, and it means a delivery transaction routinely includes payment collection and reconciliation. Yet the visibility stack is almost always specified against the parcel journey alone. Where COD is a minority payment method that asymmetry is tolerable. Where it is the majority of transactions, it means most of what your operation does each day is only half tracked.

Also Read: SEA’s $160 Billion Online Market: AI Logistics Orchestration 2026

Why real-time visibility in Southeast Asia has two clocks

Every COD order runs on two clocks that start together and finish days apart.

The delivery clock starts at dispatch and stops at handover. It is instrumented well: GPS position, status events, dwell time, proof of delivery capture, exception alerts. Latency is measured in seconds and platforms compete on it.

The cash clock starts at the moment of collection and stops when the money is reconciled against the order in a bank account. It is instrumented in batches, by finance, on a cycle measured in days, usually in a different system from the one showing the parcel.

Because those two clocks are tracked by different teams in different tools, nobody owns the interval between them. Operations closes the order at handover and moves on. Finance opens the reconciliation when a deposit lands. The days in between belong to no dashboard, and that interval is where COD leakage lives.

The useful reframe for anyone specifying a visibility platform in this region is that real-time visibility has to cover the transaction rather than the shipment. A parcel-complete event is not a transaction-complete event, and treating them as the same thing is what makes the shortfall arrive as a surprise.

The custody chain from rider to bank

Cash moves through four handoffs on its way to reconciliation. Traceability degrades at each one, and the degradation is not gradual.

Customer to rider. The collection itself. Evidence is usually the rider’s own entry in the driver app, which means the system’s record of a payment is an assertion by the person holding the money. Where the app records only a delivery status and not an amount, there is no independent record of what was collected at all.

Rider to hub or agent. End-of-shift cash-up, frequently a manual count against a printed manifest. This is where discrepancies first become visible, and where they are hardest to attribute, because the count covers a whole shift rather than an order.

Hub to remittance. Cash is aggregated into a single deposit. This is the point at which per-order attribution is lost, and it is lost completely rather than partially. A deposit is one number.

Bank to order reconciliation. Matching a lump deposit back to individual order IDs, typically at T+2 or later, and often by inference rather than by reference.

The pattern is the important part. Traceability is per-order at the first handoff and per-batch by the third. Once aggregation happens, a dispute about a specific order cannot be settled from the record, only argued from it. This is why COD disputes in SEA operations tend to be resolved by relationship and rider tenure rather than by evidence.

Also Read: ePOD Explained: Electronic Proof of Delivery 2026

When proof of delivery becomes financial evidence

In a prepaid market, proof of delivery proves that goods arrived. In a COD market it also proves that money was collected, which means it has to satisfy finance and, in a dispute, an evidentiary standard rather than an operational one.

That changes the capture requirements in ways most driver apps do not accommodate.

The amount collected has to be recorded, not merely the fact of delivery. Partial collection has to be representable, because it happens and a binary delivered flag cannot express it. The payment instrument has to be captured, since a QR payment and a cash payment settle on entirely different rails. And refusal after collection, where a customer pays and then rejects the item at the door, needs a distinct state, because operationally it is a return and financially it is a refund obligation.

Digital collection deserves specific attention here, because it is often presented as the solution and is really a change of shape. QR-based and e-wallet payment at the doorstep removes physical cash from the rider, which genuinely reduces float and shrink. It also introduces a payment service provider settling on its own timetable, so the operation now reconciles two rails with different lags instead of one. A visibility design that assumes COD means cash will misreport the digital half.

Also Read: Carrier Orchestration for SEA Reverse Logistics: A Playbook

Why the reconciliation asymmetry makes leakage look like success

Here is the mechanism that makes this expensive rather than merely untidy.

Consider two orders at the end of the day. In the first, the parcel was delivered, cash was collected, and the money will reach the bank on schedule. In the second, the parcel was delivered, cash was collected, and the money will not arrive.

On the dashboard these two orders are identical. Both are marked delivered. Both are green. Both count towards the 97.7%. There is no field in which they differ, because the only thing that differs is a future event the visibility system does not model.

Compare that with a failed delivery, which generates an exception the moment it happens, appears in a queue, alerts a coordinator, and gets triaged within the hour.

So the operation has built precise, immediate detection for the cheaper failure and no detection at all for the more expensive one. The shortfall eventually appears, but it appears in aggregate, weeks later, in a finance reconciliation, at which point per-order attribution has already been destroyed by the custody chain above. The information needed to investigate it stopped existing before anyone knew to look.

This is an information design problem rather than a control problem. Two genuinely different states are being rendered the same way.

Also Read: Best Last Mile Delivery Tracking Software Platforms in 2026

Three ways operations track COD against real-time visibility

The row that decides the others is per-order attribution. Everything downstream, dispute resolution, leakage detection, rider accountability, follows from whether the link between a specific collection and a specific order survives to the bank. It is preserved by design or lost by default, and the point at which it is lost is early.

Rider cash float, and what actually caps it

Float exposure is straightforward to compute and rarely computed: number of riders holding cash, multiplied by average daily collection per rider, multiplied by average days to remittance. For an operation running 180 riders in one city, that number is usually large enough to interest a CFO who has never been shown it.

Three levers reduce it, and two of them cost something.

Increasing remittance frequency shortens the clock directly and consumes rider productive time, because a mid-shift trip to a hub is minutes not spent delivering. In a dense urban network that trade may be worth it. Across dispersed or archipelagic geographies it usually is not.

Capping per-rider collection limits worst-case exposure and constrains routing, since the planner must now treat accumulated cash as a vehicle constraint alongside capacity and time windows. That is implementable and it does narrow the solution space.

Raising digital collection share is the only lever that reduces float without costing delivery capacity, which is why it is the structural answer rather than the operational one. It moves the problem from custody to reconciliation, and reconciliation is a data problem you can solve once.

What to measure

Cash-in-transit days. The weighted average interval from collection to bank reconciliation. This is the headline number for the second clock and almost no visibility dashboard reports it.

Remittance lag distribution by rider cohort. Split by tenure. Aggregate averages hide the concentration, and the concentration is usually in the newest cohort.

Unmatched collection rate. Collections that cannot be attributed to a specific order at reconciliation. This measures how much of your custody chain has already failed.

Order-level settlement closure within N days. The share of COD orders fully closed, parcel and payment, inside a stated window. This is the metric that makes a transaction-complete definition operational.

Digital collection share, trended. The structural lever. Track it alongside float exposure so the relationship is visible to whoever approves the incentives.

Also Read: Real-Time Delivery Visibility: 7 KPIs to Track in 2026

How Locus closes the gap between delivery and settlement

Locus, the world’s first Decision-Intelligent, Agentic TMS, treats settlement as part of the execution record rather than a downstream finance process, which is what allows a COD order to carry one identity from dispatch through to reconciliation. Its Dispatch, Capacity, Carrier, and Settlement agents, coordinated by an Orchestrator, run a continuous Sense-Decide-Execute-Learn loop against a model of more than 250 real-world constraints.

Three capabilities address the chain described above. The Settlement Agent runs invoice creation, reconciliation, and payment release as a single digital workflow rather than as separate steps in separate systems, which is where per-order attribution normally breaks. The Carrier Agent holds every transporter contract and rate structure as a live source of truth and reconciles each claim against it, so a discrepancy is caught against terms rather than argued after payment. And the control tower software provides order-level and milestone-level real-time visibility across owned fleet, contracted 3PLs, and parcel partners, which is the surface on which an open settlement item can sit next to the delivery event it belongs to.

Locus is recognized by Gartner for seven consecutive years, featured in the 2026 Hype Cycle for Supply Chain Execution and Logistics Technologies, named a Leader in TMS by QKS Group (SPARK Matrix), and ranked #1 in Route Planning on G2’s 2026 Best Software Awards. In October 2025, Ingka Investments, the investment arm of Ingka Group, the world’s largest IKEA retailer, acquired Locus. Locus continues to operate independently. Further analyst recognition is published in full.

Two deployments show the scale and the mechanism separately.

A global FMCG operation across ten Asian countries runs more than 1,000 distributors and a rider pool exceeding 5,000, reaching over 1.8 million retail outlets. At that scale the custody question is not theoretical: 5,000 riders each holding a day of collections is a working capital position, not an operational detail. The deployment reported 3X ROI, more than 12,000 trips saved per month, and over $4 billion in orders optimised, which indicates the rider network was being actively managed rather than merely tracked.

A paint industry leader automated reconciliation across 160 depots and more than 1,500 carrier invoices a month using the Settlement, Carrier, and Orchestrator agents. This is carrier settlement rather than COD, and the transferable part is the mechanism rather than the use case: checking every claim automatically against contracted terms caught 5% to 6% variance that manual review had been missing, and cut payment cycles from 30 to 45 days down to 7 to 10. A 5% to 6% variance in a process nobody believed was leaking is the closest available proxy for what an unmeasured COD chain is doing.

Request a Locus COD settlement visibility assessment to calculate your current float exposure, measure cash-in-transit days, and establish where per-order attribution is being lost in your custody chain.

Ask when the money moved, not when the parcel did

Take yesterday’s completed orders and answer one question: for the COD share of them, where is the money right now.

If the answer is a process description rather than a number, the operation is running real-time visibility on half of each transaction. That is not a reporting gap. It is the reason the monthly shortfall is a surprise every month, and the reason it cannot be attributed when it arrives.

Start by adding amount collected and payment instrument to the proof-of-delivery capture. It is a small change to a driver app, it costs nothing operationally, and it is the first point in the chain where the record either exists or does not.