Last-Mile Delivery Efficiency Benchmarks: What Good Looks Like in North America (2026)

What the Data Actually Says, and What It Doesn’t

Search for last-mile delivery benchmarks and you will find confident numbers everywhere: cost per drop, first-attempt failure rates, on-time percentages by sector. Most of them come from software vendors and content aggregators rather than research firms, and they circulate so widely that they start to look authoritative. For a benchmark to be worth citing, it has to trace to a credible, named source, and a surprising number of the figures operations leaders quote do not.

This reference sorts the last-mile numbers that hold up from the ones that do not. Where a reputable research firm publishes a figure, it is cited by name and year. Where only vendor estimates exist, that is stated plainly and the figure is left qualitative rather than dressed up as research. That distinction matters, because benchmarking your operation against a number that was never measured rigorously is worse than not benchmarking at all. The scope here is North America, and the figures below are the ones that survive scrutiny.

The Cost of Last-Mile

Last-mile is the most expensive leg, and by a lot. Last-mile delivery accounts for roughly 41 to 53% of total logistics or shipping cost, depending on how the boundary is drawn (Capgemini Research Institute, 2023). That single figure is why last-mile efficiency draws the attention it does: it is where the largest share of delivery cost concentrates, so small efficiency gains move a big number.

Fuel and labor dominate the operating cost (trucking proxy). No research firm publishes a clean last-mile-only operating-cost breakdown, but the American Transportation Research Institute’s 2024 Operational Costs of Trucking gives the best reputable proxy for the cost structure: driver pay is the largest component at about 44%, equipment about 28%, and fuel about 21% (roughly 48.1 cents per mile). This is line-haul trucking operating cost, not last-mile specifically, and should be labeled that way, but it establishes that labor, not fuel, is the dominant cost lever, which shapes where efficiency investment pays off.

Also Read: How Do Fuel Price Increases Affect Logistics Budgets?

Rural delivery costs roughly twice as much as urban. Absolute cost-per-drop figures ($2 to $3 in cities, much higher rural) circulate widely but are vendor or aggregator estimates and should not be cited as research. The credible, directional figure comes from US Postal Regulatory Commission analysis, which indicates rural delivery costs on the order of twice urban delivery (worth verifying against a primary PRC filing). Use the multiple, not the dollar amounts, because the multiple is the part that is actually sourced.

A failed first attempt costs about $17. Each failed first delivery attempt costs roughly $17 once redelivery, wasted mileage, and handling are counted (OrangeMantra). This is the most-cited figure of its kind; note that first-attempt failure rates (the “8 to 20% fail” figures) are vendor estimates, not research, so treat the rate as an industry estimate while the per-attempt cost is the more usable anchor.

Service and Reliability

This is where the benchmark landscape thins out. There is no reputable-research-firm table of on-time delivery rates by sector, and the first-attempt success rates and deliveries-per-hour figures that vendors publish are operational KPI content, not measured industry benchmarks. Presenting a vertical on-time table as research would be inventing precision that does not exist.

What research does establish is a shift in what customers value. McKinsey has found that reliability has overtaken speed as the top delivery priority, with speed falling from the number one factor to around number five between 2022 and 2024. The practical benchmark, then, is not a universal on-time percentage but a direction: measure your own on-time and first-attempt rates against your own SLAs and trend, and prioritize reliability, because that is what the demand-side research says customers now weigh most.

Returns and Reverse Logistics

Returns are one of the few last-mile-adjacent areas with solid, sector-level research. The US return rate is about 16.9% of retail sales overall and about 19.3% for online purchases (National Retail Federation, 2024). By category the spread is wide: apparel is returned at roughly 22% when bought online versus about 6.2% in-store (ICSC; verify against the ICSC release). For any operation with meaningful e-commerce or apparel volume, returns are a first-order cost and experience factor, not an afterthought, and these are the figures to benchmark against because they are actually measured.

Market Context

Two context figures are worth including with appropriate caution. On opportunity, McKinsey estimates that same-day delivery could unlock a $200B+ opportunity across North America and Europe over the decade, a credible, top-tier-sourced sizing of the demand shift. On market size, estimates cluster around $150 to $180 billion for the last-mile market in 2024, growing at roughly 9 to 11% annually, but these come from market-research firms (such as MarketsandMarkets and Future Market Insights) whose methodologies vary and which sit a step below analyst firms like Gartner and McKinsey in credibility. Cite them as a range and by name, not as settled fact.

Also Read: 8 Latest Trends in Last-Mile Delivery Technology (2026) | Locus

What Best-in-Class Operations Achieve

Industry benchmarks describe the average; they do not describe what a well-run operation can reach. As a directional sense of the ceiling, and labeled clearly as vendor-platform deployment outcomes rather than industry benchmarks, operations running agentic dispatch and routing have raised delivery execution rates from the mid-70s to the low-90s percent, cut manual dispatch effort by more than 80%, and sustained 99%+ on-time delivery (anonymized Locus deployments). The gap between those figures and the industry averages above is the efficiency that better planning and real-time adaptation can recover.

How to Use These Benchmarks

Benchmark against what is measured, and measure what is not. Use the sourced figures here, the 41 to 53% cost share, the trucking cost structure, the rural-versus-urban multiple, the $17 failed attempt, the NRF and ICSC return rates, as external reference points, cited by name so they carry weight. Where reputable benchmarks do not exist, on-time by vertical, first-attempt rate, deliveries per hour, do not borrow vendor numbers; instrument your own operation and trend against your own baseline. The most useful benchmark for those metrics is your operation last quarter, not a figure of uncertain origin.