Last-Mile Delivery Efficiency and Cost in 2026: What a First-Attempt-Rate Improvement is Actually Worth, by Density

Last-mile delivery efficiency and last-mile delivery cost both point back to the same underlying lever more often than either term acknowledges: first-attempt delivery rate, the share of deliveries completed without a second visit. It is treated as a single target applied uniformly across a network, on the assumption that a point of improvement is worth roughly the same wherever it happens. Modeling the marginal cost of a redelivery across three density archetypes found the dollar value of a one-point first-attempt-rate gain ranging from $1.76 a day in a dense urban network to $3.78 a day in a sparse regional one, a factor of 2.2, on identical stop counts. Locus, the world’s first Decision-Intelligent, Agentic TMS, prices first-attempt-rate improvements against each network’s own route density, so investment is weighted toward where it returns the most rather than spread evenly across a network where it does not pay the same everywhere.

Why a First-Attempt-Rate Target Cannot Be One Number: The Business Case

First-attempt delivery rate sits at the intersection of cost and efficiency because a failed attempt consumes both: extra vehicle time that lowers stops-per-vehicle-hour, and extra handling, redelivery and support cost that raises cost per successful drop. Industry first-attempt success rates range from 85% to 93% on average, with top-performing fleets reaching 95% to 98%, a wide enough band that most networks have real room to improve, and most improvement programs set one target across the whole operation.

That target-setting practice assumes a point of first-attempt-rate improvement is worth roughly the same wherever it lands. It is not, and the reason connects directly to the density relationship this cluster has modeled elsewhere: a redelivery inserted into an already-dense route costs a short detour, because the vehicle is passing near the address anyway. A redelivery inserted into a sparse route costs real additional distance, because there is no nearby vehicle already in the area to absorb it cheaply.

This means the same operational effort spent lifting first-attempt rate by one point returns a different dollar value depending on where in the network it happens. A program that measures its own success by the point of improvement, without tracking where that improvement occurred, cannot tell whether it captured the return available or left the larger part of it on the table by improving the wrong segment first.

The gap compounds at scale. A network running thousands of stops a day across mixed densities is not choosing between improving first-attempt rate or not; it is choosing where within the network to spend the notification, scheduling and driver-training investment that improvement requires, and that choice has a computable right answer once the density-dependent value is known.

This also reframes what a first-attempt-rate improvement program is actually for. It is not a uniform service-quality initiative applied evenly across an operation, in the way a customer-facing metric like on-time rate is often treated. It is a capital allocation decision with a computable return that differs by segment, and it deserves the same rigor applied to any other investment choice: rank the options by return, fund the highest ones first, and revisit the ranking as the network’s geography changes.

How Density Sets the Value of a First-Attempt-Rate Point

1. A failed attempt returns to the depot with the vehicle

No extra distance has been driven yet at this point, since the stop was already on the route. The parcel simply needs another attempt scheduled.

2. The re-attempt is scheduled into a future route

This is a routing decision, and the cost of that decision depends entirely on which route it lands on and how close the address sits to that route’s other stops.

3. In a dense network, a nearby vehicle almost always exists

Because stops are close together, some route on some day passes near enough to the failed address that inserting the re-attempt costs only a short detour.

4. In a sparse network, the nearest vehicle may be far away

The same insertion in a low-density area can mean sending a vehicle materially out of its way, or waiting for a route that happens to already be heading in that direction.

5. The marginal cost of the re-attempt reflects that difference directly

Because the insertion cost is what a first-attempt-rate improvement is actually buying back, the density that sets the insertion cost also sets the value of the improvement.

6. A uniform improvement target ignores this and misallocates the effort behind it

Treating every percentage point the same, regardless of where it is captured, spends equal effort on unequal returns, and the sparser segments are where the larger return is sitting unclaimed.

What the Model Shows

The model measures the marginal cost of inserting one re-attempt stop into an existing route across three density archetypes built from stated inputs rather than observed customer data: a dense urban network, a mixed suburban network, and a sparse regional network, each holding roughly 95 to 118 base stops and planned with balanced angular clustering, nearest-neighbor sequencing and local search, averaged over 60 trials per archetype with the fleet held fixed at each archetype’s own feasible size.

The marginal redelivery cost rises steadily with sparsity. It came out at $1.49 in the urban archetype, $2.71 in the mixed suburban archetype, and $3.98 in the regional archetype. The relationship moves smoothly across the three points rather than jumping between two states, consistent with density being the underlying driver rather than an urban-versus-everything-else split.

Converted into the value of a one-point first-attempt-rate gain, the range is a factor of 2.2. At each archetype’s stop count, a 1-point improvement means avoiding roughly one redelivery a day, worth $1.76 in the urban archetype, $2.71 in the suburban one, and $3.78 in the regional one. The same operational effort, the same single percentage point, returns more than double the dollar value in the sparsest network tested.

Annualized over 250 operating days, the gap is not trivial. The urban archetype’s daily value compounds to roughly $440 a year per depot from a single point of improvement; the mixed suburban archetype’s to roughly $676; the regional archetype’s compounds to roughly $946. Across a network with many depots at each density, the difference in where the investment is spent adds up to a material reallocation, and it scales with however many points of improvement a program actually captures, not just the first one.

The industry range for first-attempt rate makes this a live decision, not a hypothetical one. With published first-attempt rates running from 85% to 93% typically, and top performers at 95% to 98%, most networks have several points of realistic improvement available, and the question this model answers is which points to chase first.

What the model does not settle. It measures the cost of a single re-attempt inserted into an otherwise-fixed set of routes, which is the right basis for evaluating where to invest in prevention, but it does not model the upstream cost of achieving the improvement itself, such as notification systems or driver training, which may or may not scale with density in the same direction as the return does. If the cost of achieving a point of improvement is also higher in sparse territory, for reasons like weaker connectivity for notifications or longer driver routes that make callback windows harder to hit, the net investment case narrows from the return figure alone. The return side of the calculation is what this model establishes; the cost side needs to be measured separately for any specific network before the two are combined into a final prioritization.

Uniform Targets and Density-Weighted Targets: Key Differences

What to Look for in Density-Weighted First-Attempt Investment

Marginal redelivery cost computed per segment, not assumed uniform

A platform should be able to state what a redelivery actually costs in a specific zone’s route geometry, rather than applying one average figure across a network with materially different densities.

First-attempt-rate targets set as dollar value, not just percentage points

A percentage-point target treats every point as equal. A dollar-value target makes the density-dependent difference explicit and directs investment toward where it returns the most.

Investment in prevention, such as notification and scheduling accuracy, prioritized by computed return

Where the tools to improve first-attempt rate are limited, whether by budget or by driver-training capacity, they should go first to the segments where the model shows the highest return per point, not to whichever segment asks loudest.

Redelivery cost tracked as its own metric, not folded into general fleet cost

Isolating the marginal cost of a re-attempt is what makes the density relationship visible at all. A platform that reports only blended cost per stop cannot separate what redeliveries specifically are costing by segment, and without that separation the density-dependent return this model demonstrates simply cannot be seen, let alone acted on.

The annualized value recalculated as density shifts

A zone’s density changes as an operation grows into new territory or a route pattern is redrawn, and the value of a first-attempt-rate point should be recalculated rather than fixed once and left stale.

Density-Weighted Investment in Practice

A leading ASEAN apparel retailer. Last mile ran almost entirely through carriers, each reporting its own status codes, with no trustworthy delivery date at checkout and hundreds of thousands of delivery and returns complaints in a single half-year. After harmonizing carrier statuses into one standard set, WISMO and returns queries fell more than 40%, a result that depended on knowing where in a multi-market network the underlying failures were concentrated rather than treating first-attempt failure as one undifferentiated problem.

A Canadian grocery brand delivering fresh and perishable goods. Home delivery across more than 30 cities through contracted third-party fleets, spanning dense urban cores and less dense surrounding suburbs within the same brand. Deliveries ran 33% faster at 15% lower fulfillment cost, with customer support resolution 10 to 20 times quicker, gains that required allocating improvement effort differently across cities of different density rather than applying one national standard.

A beverage distributor with depot-based mixed fleets. Vans, trucks and motorbikes serving thousands of small retail points a day, previously planned in spreadsheets with no visibility into where redeliveries were concentrated. Fuel consumption fell 37% and orders per delivery trip rose 22%, with route planning time down 35%, once the operation could see which parts of its territory were generating the costliest re-attempts.

Common Mistakes in Pricing First-Attempt-Rate Improvements

Setting one first-attempt-rate target across a network with mixed densities. A single target either underinvests in the sparse segments where the dollar return per point is highest, or overinvests in dense segments where the return is comparatively small, without anyone deciding to make that trade deliberately.

Measuring first-attempt-rate improvement only in percentage points. A point of improvement is not a fixed unit of value; it is worth a computable dollar amount that varies with density. Reporting only the percentage hides which improvements were actually worth pursuing first.

Assuming redelivery cost is uniform across a network. It ranged from $1.49 to $3.98 across the three archetypes modeled here, a factor of 2.7 on the marginal insertion cost alone. Treating it as one average figure erases the information needed to prioritize investment, and the error compounds because the average itself is volume-weighted toward whichever density happens to carry the most stops, which is rarely the density where the return is highest.

Chasing first-attempt-rate improvement in the segment that is easiest to fix rather than the one with the highest return. Dense urban segments are often the easiest to improve, because notification infrastructure and driver familiarity are usually strongest there, but this model shows the regional segment paying more per point improved. Easiest and highest-return are different rankings, and a program that defaults to the easiest wins first is optimizing for visible early progress rather than for the largest total return.

How Locus Prices First-Attempt-Rate Improvements by Density

Locus, the world’s first Decision-Intelligent, Agentic TMS, computes the marginal cost of a redelivery from a network’s own route geometry, which is what makes it possible to price a first-attempt-rate improvement in dollars rather than treat every percentage point as equivalent. The route planning and dispatch layer reasons across more than 250 real-world operating constraints including vehicle capacity and time windows for every segment it plans, so the same data used to build routes also produces the segment-specific insertion cost this model is built on, without a separate analysis exercise. Six governance mechanisms covering explainability, traceability, evaluation, autonomy levels, execution sandbox and human-in-the-loop keep every cost figure traceable to the routes and constraints behind it, and the Control Tower carries the executed record for redeliveries specifically, so the realized value of a first-attempt-rate improvement can be measured against the modeled figure rather than assumed to match it.

The platform reasons across those constraints over 1.5B+ deliveries for 360+ enterprise customers in 30+ countries at 99.99% uptime, with $320M+ in aggregate logistics cost savings, 800M+ miles reduced and 17M+ kg of CO2 avoided. Locus has been recognized by Gartner for seven consecutive years, including the 2026 Gartner Hype Cycle for Supply Chain Execution and Logistics Technologies and the 2026 Gartner Market Guide for Multicarrier Parcel Management Solutions, where ShipFlex is featured as a Representative Vendor. Locus holds Leader designation in the QKS SPARK Matrix for Transportation Management Systems 2025 and the #1 position for Route Planning in 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.

Two deployments show first-attempt investment directed by evidence rather than applied uniformly. A leading ASEAN apparel retailer ran last mile through carriers with no harmonized status data and no trustworthy checkout date, generating hundreds of thousands of delivery and returns complaints in one half-year across a multi-market network with very different densities city to city. After moving onto Locus, carrier onboarding fell from three months to three days, a 97% improvement, and WISMO and returns queries dropped more than 40%, once the platform could see where failures were concentrated rather than treating the whole network as one undifferentiated problem. A Canadian grocery brand delivering fresh and perishable goods to homes across 30+ cities through contracted third-party fleets achieved 33% faster deliveries, 15% lower fulfillment cost and customer support resolution 10 to 20 times faster, results that depended on knowing which cities’ geographies made first-attempt investment worth the most.

A first-attempt-rate improvement is not worth the same everywhere, and treating it as a single percentage-point target hides a factor-of-2.2 difference in what that point actually returns. Modeling the marginal cost of a redelivery across three densities found it worth $1.76 a day in a dense urban archetype and $3.78 a day in a sparse regional one, on identical stop counts, with a mixed suburban archetype sitting proportionally between them. The pattern confirms density, not an urban-versus-rural binary, is the variable that matters, and it is computable for any specific network from its own route geometry. Locus prices first-attempt-rate improvements this way by default, so investment goes where the return is largest rather than where it is easiest to claim. Request a Locus first-attempt investment review to see where your own network’s highest-return segments sit.

FAQs

Is a one-point improvement in first-attempt delivery rate worth the same everywhere in a network? No. Modeling the marginal cost of a redelivery across three density archetypes found the value of a one-point improvement ranging from $1.76 a day in a dense urban network to $3.78 a day in a sparse regional one, a factor of 2.2, on identical stop counts.

Why does redelivery cost more in a sparse network than a dense one? Because inserting a re-attempt into an existing route costs whatever detour is required to reach the address, and a dense network almost always has a nearby vehicle already passing close by, while a sparse network may need real additional distance to reach the same stop.

What is a realistic first-attempt delivery rate target? Industry figures put typical first-attempt success between 85% and 93%, with top-performing fleets reaching 95% to 98%. Most networks therefore have several realistic points of improvement available, which makes where to capture them, not whether to pursue them, the more useful question.

Should first-attempt-rate improvement investment be spread evenly across a network? No. Because the dollar value of a point of improvement varies with density, spreading investment evenly, or by volume, misallocates it relative to where the highest return actually sits, which in this model was the sparsest segment tested rather than the densest.

How much does a redelivery actually cost? It depends heavily on density. In this model the marginal cost of inserting one re-attempt into an existing route ran from $1.49 in a dense urban archetype to $3.98 in a sparse regional one, a factor of 2.7 on the underlying insertion cost.

How should a network decide where to invest in reducing failed deliveries first? Compute the marginal redelivery cost for each segment from its own route geometry, convert that into the dollar value of a first-attempt-rate point for that segment, and direct notification, scheduling and driver-training investment toward the segments with the highest computed return rather than the ones that are easiest to improve.