Last-mile delivery is frequently cited as 40 to 50 percent of total supply chain cost in retail logistics studies. That number is widely quoted, but it bundles together costs that are very different in nature and very different in their amenability to reduction. Understanding which portion of last-mile cost is fixed, which is variable, and which is directly addressable through dispatch operations is what separates meaningful cost reduction from wishful thinking.
This post tries to be concrete about that cost breakdown in the context of Brazilian retail pickup operations, because we see the numbers directly through the stores we work with.
The Three Components of Last-Mile Cost
Last-mile delivery cost in a store-pickup model breaks into three components: carrier cost (what you pay the platform or courier per order), staff cost (the time your counter team spends per order on dispatch, picking, and handoff), and failure cost (the cost of failed or delayed dispatches: re-delivery attempts, customer complaints, courier reallocation).
Carrier cost gets the most attention because it shows up on invoices. But for stores operating through gig-courier platforms in Sao Paulo, per-order carrier rates are largely determined by platform pricing and competitive dynamics. There is limited room to negotiate as an individual store, particularly for growing retailers who don't have the volume to command preferential rates. Carrier cost reduction at the platform level requires either higher volume or multi-platform competition between providers, which takes time to build.
Staff cost is more controllable but often undercounted. The time a dispatcher spends evaluating which order to dispatch next, locating a bag, confirming a courier's identity, and logging the departure is real labor. At R$30 per hour for counter labor (a rough approximation for Sao Paulo retail), an average of 3 minutes of staff time per order adds R$1.50 per order in staff cost. Across 80 orders per day, that's R$120 per day or roughly R$3,600 per month in labor attributable to dispatch handling. Reducing that per-order staff time to under 1 minute saves R$2,400 per month without touching carrier rates.
Failure cost is the most variable and often the largest controllable cost. When a courier arrives and can't complete a pickup because the order isn't ready, you've paid for that courier arrival without getting a dispatch. Depending on how your platform agreement works, you may owe a partial fee, and you've consumed a curbside slot and a portion of a staff member's attention. A 20% courier arrival failure rate (the order wasn't ready when the courier arrived) adds roughly R$0.50 to R$1.00 in effective courier cost per completed order, depending on platform mechanics.
What Drives Carrier Arrival Failures
The primary driver of courier arrival failures is order assignment before the order is ready for pickup. When a courier is assigned to an order that's still in picking, there's a race between the courier's arrival and the order's bag status. If the courier wins, the order fails. If the bag wins, the dispatch succeeds.
In manual dispatch operations, order assignment often happens early. A dispatcher sees a courier approaching on the app and assigns them the closest available order to avoid losing the courier to a different job. If that order is still in picking, the assignment was premature. The dispatcher was optimizing for courier retention (the visible risk) at the cost of ready-status accuracy (the hidden risk).
Automated dispatch sequencing addresses this by making bag-ready status a required input to courier assignment. An order in "picking" status cannot be assigned to an incoming courier. This prevents the premature assignment problem entirely but requires reliable status updates from the picking staff. If picking staff are inconsistent about confirming ready status in the system, the automation can't help.
The Batching Economics
Batching, consolidating multiple orders for one courier, changes the cost calculation meaningfully. A courier delivering three orders in one trip to the same district zone produces three courier dispatches for roughly the cost of 1.3 to 1.5 standard dispatches (accounting for the slightly longer trip time). The per-order carrier cost drops by 40 to 50 percent for the batched orders.
The limitation is that batching introduces wait time for orders that are ready before the batch is full. The operations question is: what is the right exchange rate between per-order carrier cost reduction and additional customer wait time? For orders with a 30-minute delivery SLA, batching is a bad trade. For orders with a 90-minute window, it's usually favorable if the batch forms within 20 to 25 minutes.
Most retail stores we work with have a mix of both SLA types but don't route them through separate dispatch logic. All orders go through the same dispatch queue, which means the 90-minute orders eat the same carrier cost as the 30-minute orders. Separating them into SLA-appropriate dispatch paths, with batching enabled only for the appropriate tier, is one of the most direct available cost levers for stores that have both order types.
What Dispatch Improvement Can and Cannot Do
I want to be direct about the limits here. Dispatch sequencing optimization is primarily a labor cost and failure cost lever. It does not reduce base carrier rates. For stores where carrier cost is the dominant component of last-mile cost, dispatch optimization will improve margins but not dramatically. The savings come from staff efficiency gains and failure-rate reduction, not from renegotiating platform economics.
The stores where dispatch improvement produces the strongest cost outcomes are those with high failure rates (above 15% courier arrival failures) or high per-order staff handling time (above 4 minutes per order). Those are the conditions where the failure cost and labor cost components of last-mile expense are large enough that sequencing improvements produce measurable savings within the first month.
For stores already running below 10% failure rates and under 2 minutes per-order handling time, the marginal cost reduction from further dispatch optimization is smaller. The investment in tooling and configuration may still be worth it for customer experience reasons, but the purely economic case is less clear. It's worth doing the calculation before assuming dispatch technology will pay for itself on cost reduction alone.