Last-Mile Cost Simulator

Model cost per delivery, attempt cascade, zones, own vs 3PL mix and returns. Find per-zone break-even and the optimal split. AI, 3 scenarios. Free.

Advanced simulator

Which zones should I price up or drop?

Find which zones stop being profitable, where to raise prices or stop serving, and whether your own fleet or a 3PL fits better.

Volume

Total daily orders and monthly operating days.

Delivery zones

The three zones sum to 100%. Each has its own density, distance, service time and success rate.

Σ 0%
UrbanDowntown / CBD, high density.
SuburbanMedium residential areas.
RuralLow density, long distances.

Own fleet

Vehicles, drivers and direct costs of the in-house operation.

3PL / outsourced

Share of volume handled by an external carrier. Their fees may include success, failed attempt and return.

Revenue, returns and SLA

Revenue per delivery, real returns, fixed overhead and SLA-miss penalties.

Saved configurations

Fill in your data to see the report

This simulator only generates a diagnosis, charts and recommendations when it has your real business values. Fill the editor above and the report will appear automatically.

Required data

  • Orders per day
  • Operating days per month
  • Own vehicles
  • Driver cost per hour
  • Vehicle cost per day
  • Fuel price
  • Revenue per delivery
  • Fixed monthly ops cost

Want to explore it first?

Load a realistic case to see how the report looks. You can edit any field afterwards.

Connect with other simulators

Cash Flow Simulator

Project how last-mile profit (or loss) impacts monthly cash.

Delivery Routes Simulator

Validate fleet size and density with Daganzo before changing the split.

Warehouse Capacity Simulator

If your return rate is high, size the warehouse for reverse logistics.

Methodology and assumptions

How results are calculated, what we assume when modeling, and where the method loses precision.

Formula

Cost per delivery = (Variable + Allocated fixed + SLA penalty) ÷ Successful deliveries · Break-even = Fixed ÷ (Revenue − Variable cost)

Assumptions

  • Per-attempt success rate stable (1st, 2nd, 3rd visit).
  • Returns imply full reverse logistics cost.
  • SLA with linear penalty over deliveries outside the time window.

Applicability limits

  • Does not model dynamic cross-docking or urban hubs.
  • Per-zone costs must be entered manually when the spread is over 20%.
  • For mixed fleets (in-house + 3PL) run the simulator twice and compare.

Sources

How it works

1. Declare volume and zones

Orders/day, operating days and the urban/suburban/rural mix with its own distance, service time and FTR.

2. Own fleet and 3PL

Own vehicles with costs, and the 3PL mix with its fees per success, failed attempt and return.

3. Revenue, returns and SLA

Revenue per delivery, return rate, fixed overhead and SLA penalty. The simulator crosses everything and compares 3 scenarios.

Frequently asked questions

1How is this simulator different from the Delivery Routes one?
Delivery Routes solves fleet + density with Daganzo (where to operate and with how many vehicles). This simulator focuses on per-attempt and per-zone economics, and on the own-vs-3PL decision. Use them together: first size the operation with Routes, then decide the operating model with Last-Mile.
2How does FTR affect real cost per delivery?
Cost per successful delivery is amortized over attempts. With 90% FTR you need 1.11 attempts per success; with 70% FTR that jumps to 1.43 — 29% more direct cost. If the failure also costs money (driver + fuel without revenue), real impact is higher. That is why serious operators invest in notifications, windows and PUDO before adding fleet.
3When is own fleet better than 3PL?
Rough rule: own is cheaper when density is high and volume fills the shift. 3PL is cheaper when density is low (rural), volume is irregular, or the mix requires wide geographic coverage. The simulator shows per-zone margin on both channels — if your own loses money in urban, something is off (low utilization or inflated costs).
4What do I do if my rural zone is negative?
Three routes: (1) repricing — charge an explicit rural surcharge; (2) consolidate — deliveries every 2-3 days instead of daily; (3) outsource to a rural-specialized 3PL. What does NOT work: raising the general service price to 'cover' rural — you kill your competitiveness in urban, where you actually make money.
5How do returns fit in the model?
We model two different things: failed attempts (could not deliver, retry) and real returns (you delivered but the customer sent it back). Each has its own cost. A return rate >10% is typical for apparel and can double your real cost per successful delivery — that is why many ecommerce operations never reach positive margin until they attack that number.

Related simulators

Delivery Routes SimulatorWarehouse Capacity SimulatorCash Flow Simulator

Complete guide

Last-mile delivery: why it's 53% of shipping cost and how to optimize it

The last mile — the final leg from DC or dark store to the customer's door — is the most expensive segment of the logistics chain. Capgemini, ARC Advisory and CSCMP studies agree: it represents 41% to 53% of total shipping cost in B2C e-commerce operations. In dense urban (Manhattan, Chicago Loop, Brooklyn, London Zone 1, CDMX, Bogotá) the share can hit 60%. Structural causes: low stop density per vehicle (20-50 on B2C vs 200+ consolidated B2B2C), high no-home rates (first-time delivery failure), returns, tight time windows, and rising velocity expectations (same-day, next-day).

Key formulas and metrics

Last-mile cost per drop = Total operating cost ÷ Successful deliveries

Delivery density = Stops delivered ÷ Kilometers driven

First-time delivery success rate (FTDR) = Successful first-attempt deliveries ÷ Total attempts × 100

Return rate = Parcels returned ÷ Parcels dispatched × 100

Benchmarks: last-mile cost per drop US USD 6-12, LatAm urban USD 2.50-5.50 (Rappi, 99minutos, Uber Eats Delivery); healthy urban FTDR 88-94%, degraded <85%; e-commerce return rate general 15-20%, fashion 25-40%, electronics 8-12%.

Operating models: in-house vs outsourced vs hybrid

In-house fleet: vehicles, drivers, dispatch under operator control. Pros: total experience control, vehicle branding, granular data; cons: CAPEX, HR management, high fixed cost that doesn't scale with variable volume. Typical break-even: stable volumes >200-300 daily deliveries in a defined zone.

Outsourced (3PL / carrier): UPS, FedEx, USPS, DHL in US/UK; Estafeta, DHL, 99minutos in Mexico; Servientrega, Coordinadora in Colombia; Rappi Turbo, Cornershop for on-demand. Pros: variabilization, coverage, carrier scale economies; cons: carrier margin, less experience control, capacity dependency during peaks.

Crowdsourced delivery (gig economy): Uber Direct, DoorDash Drive, Amazon Flex, Roadie, Rappi Turbo. Pros: ultra-variabilization, elastic capacity; cons: variable quality, high turnover, shifting regulation (rider law debates in Europe, labor reform in Mexico).

Hybrid model: in-house fleet for dense zones + 3PL for periphery + crowdsource for peaks. Amazon Logistics, Walmart US and MercadoLibre use hybrid structurally — most efficient for geographically and temporally heterogeneous demand.

PUDO and dark stores: alternatives to home delivery

PUDO (Pick-Up / Drop-Off): network of pickup points — Amazon Lockers, UPS Access Point, FedEx OnSite, OXXO in Mexico, MercadoLibre points, Parcelly in the UK. Advantage: density collapses cost (50-200 deliveries at one point vs 50 deliveries at 50 doors). Cost per parcel 30-60% lower than home delivery. LatAm penetration still low (8-14%) vs Europe (30-45% Germany, 25-35% France); US sits around 15-25% depending on metro.

Dark stores / micro-fulfillment: mini urban warehouses 2-5 km from final customer. Getir, Gorillas (closed), Jokr models; in LatAm Rappi Turbo, Merqueo. Drops delivery time to 10-30 min and lowers cost per drop through ultra-concentrated geography. Economically validated in high density (Manhattan, Brooklyn, Chicago Loop, Madrid, CDMX metro core) with >400 daily orders per dark store.

First-time delivery failure: the hidden cost

A failed attempt costs 1.5x-3x the original shipment. Top cause in US/LatAm: customer not home (home delivery). FTDR lift tactics: (1) tight time windows with customer confirmation (4-6h instead of 8-12h); (2) in-transit notifications (SMS 30-60 min out); (3) PIN/OTP at the door; (4) authorized-neighbor fallback; (5) PUDO as fallback option. Carriers moving FTDR from 82% to 92% typically cut total last-mile cost 14-22%.

Reverse logistics: returns as its own workstream

Returns isn't a subset of last mile — it's a parallel workstream with its own economics. Reverse logistics cost = pickup + transport + inspection + restock or liquidation + inventory write-off (if unsellable). In fashion e-commerce a return can eat 40-70% of the original order margin. Mitigation architecture: (1) locker/PUDO returns (customer drops off at OXXO, UPS Access Point, Amazon Locker — no pickup cost); (2) carrier bulk consolidation (pick up returns at the same stop where you deliver new orders); (3) inspection-at-origin with instant credit (skip round-trip to central warehouse for low-value items); (4) liquidation partnerships with B-stock channels (B-Stock, Liquidation.com) for items that don't re-enter inventory. Companies that industrialize reverse logistics recover 8-15 margin points vs those that treat it as ad hoc.

Worked example: US fashion e-commerce, 8,000 shipments/month

100% outsourced operation with national US carrier: average USD 7.40/parcel. Volume 8,000 monthly shipments, return rate 28% (fashion typical), FTDR 84%.

Comparative simulation:

  • Status quo (100% carrier): USD 7.40 × 8,000 = USD 59,200/month. Returns at USD 7.40 × 2,240 × 1.8 (return factor) = USD 29,840. Total USD 89,040. Cost per successful delivery: USD 15.50.
  • Hybrid with in-house fleet NYC metro + carrier rest: dense zones (40% volume) at USD 4.00/parcel, FTDR 92%; rest USD 7.40 via carrier. Monthly cost USD 50,480, returns 21%, total USD 73,560. Cost per successful delivery USD 11.65.
  • Hybrid + PUDO option with USD 2 customer discount: 22% of volume migrates to PUDO, effective FTDR 96%. Monthly cost USD 59,900, returns 15%, total USD 70,530. Cost per successful delivery USD 10.40.

Annualized savings scenario 3 vs status quo: USD 222,000 + NPS improvement of +9 points from flexibility.

Density economics: the fundamental driver of last-mile cost

Every last-mile cost model eventually reduces to stops per route — the number of deliveries per vehicle per day. The economics are non-linear:

  • Urban dense (Manhattan, Chicago Loop, CDMX Cuauhtémoc): 35–55 stops/route. Cost per drop USD 3.50–6.00. In-house fleet competitive above 200 daily shipments per zone.
  • Suburban: 18–28 stops/route. Cost per drop USD 6.00–9.50. Carrier usually wins unless the operator has stable, route-optimized volume.
  • Rural / low-density: 8–14 stops/route. Cost per drop USD 10–18. Carrier or PUDO consolidation is almost always cheaper than in-house.

The math is straightforward: a driver's all-in cost (salary, benefits, vehicle TCO) runs USD 250–350/day in the US. At 45 stops, that's USD 6–8 per stop before any variable cost. At 12 stops it's USD 21–29 per stop. This is why Amazon's dense-zone economics are unbeatable — they engineer density before signing new routes.

2026 LATAM context: competitive pressure and new entrants

In Mexico, Colombia, Brazil and Argentina, last-mile is undergoing structural change in 2026:

  • MercadoLibre Envíos Flex allows sellers to use their own vehicles for same-day delivery in exchange for preferential algorithm placement, creating a gig-last-mile model competing with Rappi and 99minutos.
  • Rappi Turbo (10-minute dark-store model) has expanded to 14 cities across LatAm; its dense urban coverage is pushing standard e-commerce carriers to cut 2-hour delivery prices.
  • InDrive Delivery entered last-mile in Mexico and Colombia with a driver-bidding model that undercuts established 3PLs on price by 15–25% in off-peak hours.
  • OXXO Envíos (Femsa network) has made 20,000+ convenience stores into PUDO points across Mexico — the fastest PUDO network growth in LatAm.

For mid-market e-commerce operators in LatAm, the 2026 competitive reality is that outsourced last-mile prices are under structural pressure downward from platform competition — but quality variance is also increasing, making FTDR monitoring more critical than ever.

Route optimization: the software layer

Modern last-mile operators cannot compete without route optimization software. The difference between manually assigned routes and algorithmically optimized routes on a fleet of 20+ vehicles:

  • Distance reduction: 15–25% fewer kilometers per route (OptimoRoute, Route4Me, Onfleet, Circuit benchmarks 2024).
  • Time window compliance: optimized routes hit customer time windows 94–97% vs 78–85% manually.
  • Failed-delivery reduction: tight time windows enabled by optimization cut first-attempt failures 8–14 points.
  • Driver overtime: optimized routes equalize workload across drivers, reducing overtime by 20–35%.

For a 30-vehicle fleet running 250 stops/day, a 20% distance reduction at USD 3.80/gallon (diesel) on vehicles getting 15 MPG saves roughly USD 18,500/year — and the FTDR improvement typically saves more. ROI on mid-market route software: 2–4 months.

Dynamic pickup consolidation and micro-hubs

An underused tactic for high-density urban e-commerce: micro-hubs — small urban staging points (parking structures, shared retail back-rooms, pop-up trailers) 1–3 km from the customer cluster. The pattern: a van delivers 80 parcels to the micro-hub, then riders on e-bikes or on foot cover the final 500m–1.5km radius. The result:

  • 40–60% reduction in parking and idling time vs van-to-door.
  • E-bike delivery cost per drop: USD 1.50–3.00 vs USD 5–9 for van.
  • Emissions reduction 70–85% on the micro-hub leg (relevant for city logistics restrictions — Paris, Amsterdam, Barcelona low-emission zones limit diesel vans; CDMX no-circula rules are tightening).

Amazon, DHL, and UPS all operate micro-hub programs in European cities. In LatAm, Rappi Turbo uses dark stores as micro-hubs by design. For a D2C brand delivering 400+ daily orders in a 5 km urban radius, a micro-hub model can reduce cost per delivery from USD 7.50 to USD 3.80 — nearly halving last-mile cost on that zone.

Common mistakes and red flags

  • No FTDR tracking by zone and driver: without per-driver, per-zone FTDR data, you cannot identify whether failures are structural (bad time windows) or operational (specific driver issue).
  • Using a single carrier with no backup: peak season capacity constraints at your primary carrier will strand shipments; always have a secondary and a crowdsource escalation path.
  • Treating return logistics as an afterthought: reverse logistics cost is not zero. Map it per channel (carrier pickup vs PUDO drop) and model it separately.
  • Oversizing the in-house fleet for average volume: size the owned fleet for the 60th-percentile daily volume, use crowdsource for the peaks. Oversizing ties up CAPEX in idle vehicles.

Conclusion

Last mile isn't optimized with a single provider but with architecture: mix of fleet, carrier, crowdsource and PUDO calibrated by zone and product type. The simulator lets you load volume by zone, compare models, quantify the impact of lifting FTDR and reducing returns, and find the break-even where in-house fleet beats the 3PL. For mid-market e-commerce between 3,000 and 40,000 monthly shipments, it's the operating decision that most levers margin — every dollar recovered in last-mile drops straight to the bottom line.

Illustrative case

Composite case for instructional purposes: combines sector dynamics with realistic figures. Names are fictional and do not represent a specific company.

Case study: US DTC apparel brand, 11,400 monthly shipments

Stitch & Shore (name modified) is a US DTC apparel e-commerce that closed 2024 with 11,400 monthly shipments and an operating margin of 6% — below the 11% direct-competition benchmark. Internal diagnosis pointed to last mile: average USD 9.20/parcel (100% outsourced carrier), FTDR 81%, return rate 34% (fashion typical) and a delivery NPS of 42 — meaningfully low.

The team loaded 90 days of operating data into the simulator by delivery zone. Analysis revealed: (1) 38% of volume concentrated in NYC metro with stops <10 miles apart — obvious in-house fleet candidate; (2) FTDR degraded by 8-12h windows without 30-min notification; (3) high return rate from lack of locker/store-return policy — customers were returning via carrier (double logistics cost) instead of exchanging at physical store or locker.

Q1 2025 implementation: (A) in-house fleet of 8 vehicles + 12 riders in NYC metro with 3-hour windows and 30-min SMS; (B) retained national carrier for rest of country; (C) partnership with 2,400 UPS Access Point lockers + 180 private lockers for pickup and direct returns.

Eight-month results: NYC cost per drop USD 9.20 → USD 4.50 (−51%); consolidated FTDR 81% → 93%; return rate 34% → 23% (locker exchange avoids carrier return in 40% of cases); cost per successful delivery USD 14.00 → USD 9.30 (−34%); delivery NPS 42 → 71 (+29 points). Operating margin moved from 6% to 10.8% in 12 months — USD 560,000 additional annualized on stable revenue base.

From theory to calculation

When you need more than a quick calculation, our advanced simulators model full scenarios with your data.

See advanced simulators

Sector reference ranges

Indicative ranges based on public sector literature and operational observation. Your business may differ — use the numbers as a starting point, not as a target.

MetricValueSource
Last mile as % of total shipping cost53%Capgemini Last-Mile Delivery Report 2024
Cost per drop — urban last mile, LatAmUSD $3.50-6.5099minutos / Rappi disclosures + ARC Advisory 2024
Healthy first-time delivery success rate — urban92-95%DHL Last-Mile Benchmarks 2024
Incremental cost of a failed delivery attemptUSD $4.50-8.00CSCMP State of Logistics 2024
PUDO penetration — LatAm vs Europe8-14% vs 30-45%GlobalData Last-Mile Market Report 2024
E-commerce return rate — fashion25-40%Shopify Commerce Trends 2024 / NRF 2024

Frequently asked questions

1What percentage of shipping cost is last-mile delivery?
Between 41% and 53% of total shipping cost per Capgemini and CSCMP 2024. In dense urban (Manhattan, Chicago Loop, London Zone 1) it can exceed 60%. It's the most expensive segment from low stop density (20-50 on B2C vs 200+ consolidated), tight time windows, failed attempts, returns and rising velocity expectations.
2At what volume does an in-house fleet make sense?
Typically >200-300 daily deliveries concentrated in dense zones with stop radius <10 miles. Below that threshold, fixed cost of vehicles, drivers and dispatch doesn't amortize vs marginal cost of a 3PL (which variabilizes with volume). For operations with strong seasonal demand, hybrid almost always wins: in-house fleet for constant base + 3PL and crowdsource for peaks.
3What is a good first-time delivery success rate (FTDR)?
88-94% in healthy urban; below 85% signals a problem. Main levers: 3-6h time windows (instead of 8-12h), in-transit notifications (SMS 30-60 min out), authorized-neighbor fallback, PIN/OTP at the door, and PUDO as alternative. A failed attempt costs 1.5x-3x the original shipment, so lifting FTDR 5 points typically reduces total cost 10-15%.
4What is PUDO?
Pick-Up / Drop-Off points — network of points where customers pick up or return parcels: Amazon Lockers, UPS Access Point, FedEx OnSite, OXXO in Mexico, Parcelly in UK. Key advantage: density collapses cost per drop (50-200 deliveries at one point vs 50 at 50 doors). Cost per package 30-60% lower than home delivery. US penetration 15-25%; LatAm 8-14%; Europe 30-45%.
5What are dark stores?
Mini urban warehouses not open to the public, 2-5 km from final customer, built for fast fulfillment (10-30 min). Getir, Jokr, and in LatAm Rappi Turbo, Merqueo model. Economically viable in high density (>400 daily orders per dark store) and core urban. Enable same-day and same-hour delivery with competitive cost per drop through ultra-concentrated geography.
6Does crowdsourced delivery work for e-commerce?
Yes for peaks and variable demand, not as backbone. Platforms (Uber Direct, DoorDash Drive, Amazon Flex, Roadie) give immediate capacity elasticity but carry variable quality and high turnover. Successful model is hybrid: backbone with in-house fleet or stable 3PL + crowdsource for specific window peaks (season, daily peaks, unplanned same-day). Growing regulatory risk in US, LatAm and Europe (gig-worker classification, rider reform).
7How do I reduce return rate in e-commerce?
Four measured-impact levers: (1) interactive sizing with 3D models or augmented reality (−15-25% fashion); (2) locker/PUDO return instead of logistics pickup (avoids 40% of carrier returns); (3) rich product description with video and reviews (closes expectation-reality gap); (4) fit prediction based on customer history (−10-18% size-driven returns). Benchmark fashion return rate 25-40%; electronics 8-12%; general e-commerce 15-20%.

Keep exploring

Tools from the same topical cluster. Use them together to close the loop on your analysis.

Last updated: April 30, 2026 · Reviewed by the Simúlalo editorial team. Figures and benchmarks are indicative; verify with your own data before deciding.

View methodology

How this simulator was reviewed

What you'll see, what it prevents, and where you shouldn't trust it

Every simulator on Simúlalo ships with the same editorial structure: two hypothetical worked examples with numbers, the errors it helps you avoid, the model's declared limitations, and a visible financial disclaimer. The review is signed and dated.

Hypothetical caseCase A

An e-commerce that finds out 'same-day' loses $42 per delivery

An e-commerce offers 'same-day' in CDMX at $89/delivery cost (dedicated rider, higher FTR, hub investment). Customer charge is $69 plus product. The simulator decomposes: logistics revenue $69, cost $89 = loss $20. After adding returns (4.5% on same-day vs 2.1% on standard) and re-delivery costs, the real loss is $42/delivery. Volume: 720 same-day deliveries/mo = $30,240/mo loss. The decision: charge $109 for same-day or limit availability to high-density postal codes where cost falls to $58.

Illustrative figures. Does not represent a real company or an investment recommendation.

Hypothetical caseCase B

A marketplace that postpones expansion where break-even requires 14 deliveries/day

A marketplace evaluates opening its own operation in a secondary city of 380,000 people. Initial study projects 2,400 orders/mo by month 6. The simulator, with hub fixed cost ($120,000/mo), 6 riders, 5% returns, and 48h SLA, computes break-even at 14 deliveries/day per rider — and the projection is 13. The decision: postpone the proprietary expansion until organic volume exceeds 3,000 orders/mo; meanwhile operate with a local 3PL with smaller margin but controlled risk.

Illustrative figures. Does not represent a real company or an investment recommendation.

Common mistakes it helps you avoid

Things a team or decision-maker might assume that this simulator forces you to verify before committing.

  • Calculating last-mile cost on successful delivery only, without adding the failures: if FTR is 88%, the remaining 12% returns, retries, or refunds, and that multiplies cost.
  • Subsidizing 'free shipping' without measuring the margin impact: free shipping lifts conversion, but if unit cost doesn't amortize with higher AOV (average order value), the business breaks.
  • Mistaking SLA for capability: an aggressive SLA (same-day) is a commercial commitment, not an operational capacity — if your operation can't sustain it, ratings drop and the loss compounds.
  • Treating all zones as equal: unit cost in central zones can be 3-5x lower than periphery. Profitability per zone defines the coverage strategy.

Model limitations

What the simulator does not do, and where you need a professional or a specialized tool.

  • Does not optimize routes. For dynamic routing use OR-tools, OptimoRoute, OnFleet, or similar. This simulator models aggregate cost from declared densities and SLAs.
  • Does not include specific operational frictions: load caps per motorcycle, hour restrictions in pedestrian zones, hospital delivery windows.
  • 3PL costs are market reference values. For a real decision, request quotes by zone and volume.
  • Break-even calculation assumes you can scale fleet linearly. In practice, hiring and training riders takes 30-60 days.

When NOT to use this simulator

If you're going to define the logistics proposal for fundraising or a business case for enterprise investment, this simulator is a pre-analysis tool. You'll need real per-zone quotes, historical FTR data, channel cannibalization analysis, and volume projections backed by paid marketing. Use it to prepare the conversation with the VP of Operations or COO; not to present to the investor.

Financial notice

Results are illustrative estimates and do not constitute financial, tax, accounting, or legal advice. Use the results as a reference point and validate important decisions with a certified professional.

Editorial review

Reviewed by the Simúlalo editorial team

This simulator was reviewed by the people listed below before being published. The review covers the declared formula, the model's assumptions, the explicit limitations, and the absence of unsupported financial claims.

They are part of the Simúlalo editorial team, focused on building financial tools that are clear, educational, and easy to interpret.

Last updated: We update this page when the methodology, sources used, or simulator structure change.

This tool uses standard financial formulas and user-supplied data. To explain concepts like rates, credit, risk, or cash flow we consult public and official sources (Banxico, SAT, CONDUSEF, CNBV, Banco de España, IFRS, BIS, among others). Simúlalo is not affiliated with, sponsored by, or endorsed by these institutions.

See all logistics tools