.webp)
A D2C brand founder we spoke to last month was shipping 1,200 orders a day across the South East. Her monthly Evri and Yodel bill was £84,000, around 18 percent of order value. Her support team was fielding 240 "where is my order" tickets a week. Her churn rate was climbing.
She asked whether it was time to buy vans and hire drivers.
Under the right conditions, the answer is yes. The McKinsey Center for eCommerce puts last-mile at 28 to 53 percent of total shipping costs. McKinsey's consumer research shows 90 percent of shoppers will wait 2 to 3 days if the promised window is met, and 84 percent will abandon a brand after a failed delivery. Reliability now matters more than speed, and that plays to owned last-mile done well.
Under the wrong conditions, owning last-mile is the fastest way to burn a Series B. Most D2C brands who try do it too early.
This article is a candid guide for D2C brands weighing owned last-mile against outsourced carriers in 2026. Real cost economics, what the software has to do, how proof of delivery and route optimisation work in production, and how to keep carriers as a fallback while your operation matures.
When Owning Last Mile Actually Beats Outsourcing
Three tests before you commit capital.
Delivery density. Own economics work when parcels-per-square-mile in a depot's coverage area support a full van route with reasonable stops. Below 250 deliveries per day per depot, van utilisation is too low and per-parcel cost exceeds outsourced. Above 400 per day, van utilisation is efficient and per-parcel cost drops below outsourced. Between 250 and 400 depends on your specific geography and margin structure.
Service level differentiation you actually need. If Yodel or Evri next-day is acceptable to your customers, own-fleet is a cost play only. If you need same-day, specific one-hour windows, or on-brand driver experience, own-fleet is also a customer play. The customer angle is often where ROI actually lands.
Depot geography that matches your customer base. Owned last-mile needs enough depots close enough to your customer clusters. UK D2C brands serving London, the South East, and the Midlands need at least three depots. A US D2C brand doing similar volume across the North East needs five to eight. Thinly spread nationally? Own last-mile does not fit.
Fail any of these three tests and keep the carriers. Focus on getting more out of your Shopify Ship, ShipStation, or Metapack setup and revisit in twelve months.
Real 2026 Cost of Ownership: Vans, Drivers, and Software
Real UK numbers for one depot running 8 vans at 40 to 60 deliveries per van per day.
Vans. Small van lease £300 to £450 per month; electric £400 to £600. For 8 vans, £2.4k to £4.8k monthly.
Drivers. UK driver salary £28k to £38k per year including National Insurance and pension. For 8 drivers plus a dispatcher, £21k to £30k monthly.
Fuel. £150 to £250 per week per van. £5k to £8k monthly.
Insurance. £2k to £4k per van per year. £1.3k to £2.7k monthly.
Maintenance and telematics. £1k to £3k per van per year plus £10 to £25 per vehicle per month software. £800 to £2.2k monthly.
Depot rent. £3k to £8k monthly outside London.
Software. SaaS £1k to £3k monthly for 8 vans. Custom, once built, runs £500 to £2k in hosting and maintenance.
Total monthly cost for one 8-van depot: roughly £34k to £58k for around 12,000 parcels per month. Per-parcel operational cost: £3 to £5. Add depreciation and amortised software build, and per-parcel all-in lands £3.50 to £6.50.
Compare to outsourced UK: £5 to £8 per parcel with Yodel or Evri at moderate volume; £8 to £14 with DPD or Royal Mail 24 at premium tiers. The saving is real, but it takes 12 to 18 months to realise after factoring in the software build and depot ramp.
What Last Mile Software Actually Has to Do
Six modules any serious last-mile platform must cover.
Order intake and depot allocation. Orders arrive from Shopify, marketplaces, and your OMS. The software assigns each order to a depot based on address, capacity, and priority.
Route optimisation. For each depot's daily orders, the software computes optimal routes across available vans, respecting time windows, van capacity, driver breaks, and traffic. Google's Route Optimization API is the leading production choice for custom builds, alongside commercial options (Route4Me, OptimoRoute) and open source (OR-Tools, VROOM). Realistic impact: 10 to 20 percent fewer miles, 15 to 25 percent more stops per van, 5 to 10 percent fuel saving.
Driver app. Mobile app for drivers to see routes, capture POD, report exceptions, and get support.
Dispatcher dashboard. Real-time view of all vans, current and next stops, ETAs, and an exception queue for live intervention.
Consumer tracking. A branded page showing live driver location on delivery day, narrowing ETA as the van approaches, driver name and photo when 10 minutes away, and POD retrieval after delivery. The DoorDash pattern raises first-attempt success rate materially.
Exception handling. Missed deliveries, damaged goods, wrong address, customer not home. Clear workflow to resolve, refund or reschedule, and communicate to the customer.
Off-the-shelf platforms cover most of this. Custom becomes the answer when you need specific integrations (video POD, complex exception rules) or you have scale that justifies the build cost.
Proof of Delivery: Photo, Signature, or Video
Four POD methods in production use in 2026.
GPS-timestamped photo. Driver takes a photo of the parcel at the delivery location, tagged with GPS and timestamp. Now the industry standard for standard parcels. Table stakes.
On-glass signature. Customer signs on the driver's device. Standard for regulated categories (alcohol, prescription products, high-value electronics) where legal proof of receipt matters. Slower per stop but non-negotiable for certain SKUs.
Video POD. A short delivery video, or the packing video from despatch linked to the delivery event. Materially reduces fraud in high-fraud categories (electronics, luxury goods, fashion). This is where TrackVid credibility applies directly.
NFC or QR tap-to-confirm. Customer taps a physical NFC or scans a QR at delivery. Popular in Continental Europe, gaining UK adoption. Adds friction but strong proof.
Design your POD stack for the categories you actually ship. A clothing and homeware brand can probably run on GPS-timestamped photo alone. A brand shipping premium electronics or wine should build video and signature into the flow.
What We Learned Adding Video POD to TrackVid
TrackVid is our video proof and claim management platform for ecommerce sellers. Video POD is one of our core features. Two lessons transfer directly to a last-mile build.
Video storage costs will eat you if you do not design retention. A 30-second delivery video at HD is 5 to 10 MB. At 400 deliveries per day per depot, that is 60 GB per week of raw video. We settled on 90-day standard retention with extended retention only for videos linked to a claim or dispute. Cost per depot lands £30 to £80 per month instead of £300 to £800 without a retention policy. Design retention from the start.
Linking video to order reliably is not trivial. In TrackVid, a packing video for a specific order can be retrieved by order ID, SKU, tracking number, or customer email. In a last-mile build, you need the same for delivery videos: the customer disputes on their end, and you need to retrieve the correct video in under 30 seconds. Design the data model for retrieval from day one.
You can see TrackVid alongside our portfolio of shipped work. If you want to talk about how the same engineering translates into your D2C last-mile platform, book a last-mile strategy call with WhiteStone.
Realistic Software Cost Bands for a Custom Build
Three phases we walk D2C brands through if they commit to custom.
Proof of concept: £30k to £70k over 8 weeks. One narrow use case (usually route optimisation plus a driver app for a single test route). Offline data, no consumer-facing tracking. Purpose is to prove the software works for your depot geography and driver workflow.
Pilot: £80k to £180k over 3 to 5 months. Full driver app, dispatcher dashboard, basic consumer tracking, POD capture, integration with your OMS or Shopify. One depot, running alongside carriers for the first 3 months. Measure lift honestly.
Production: £250k to £500k over 6 to 10 months. Multi-depot deployment, full consumer tracking (live map, ETA narrowing, POD retrieval), exception workflow, reporting, and finance integration.
Add £500 to £2,000 per month in ongoing platform run cost. Route optimisation API usage lands at £0.02 to £0.08 per delivery on Google Routes.
For most D2C brands under 10,000 deliveries per week, SaaS platforms (OptimoRoute at £15 to £30 per vehicle per month plus per-delivery fees, or Onfleet, or Bringg) are the right answer. Custom becomes defensible above 15,000 deliveries per week or when you have unusual POD, tracking, or integration requirements SaaS cannot cover.
How to Keep Carriers as a Fallback
Even brands with fully-built own last-mile keep carrier fallback for four reasons: peak spikes (Black Friday, Christmas), locations outside depot coverage, weather disruption, and single-item high-value shipments where carrier tracking is what the customer expects.
Design your platform for carrier fallback from day one. Three principles.
Every order can route to any carrier or your own fleet. Routing happens at order intake based on address, priority, capacity, and configurable rules. Not hardcoded to own fleet.
One tracking experience regardless of carrier. The customer sees your branded page whether the parcel is on your van, in DPD's network, or with Royal Mail. Carrier tracking events feed into your platform and get normalised into your customer-facing format.
Cost visibility per parcel per route. Your platform reports cost of every parcel by carrier or own fleet. Finance and ops can see whether the routing decision was economically correct after the fact.
This design also protects you if you decide, six months in, that owning last-mile is not right for your business. Reverting to full carrier delivery becomes a configuration change, not a rebuild.
Frequently Asked Questions
Should a D2C brand own last-mile or outsource it?
Own last-mile beats outsourced above roughly 400 deliveries per day in a defined geography, when you need service-level differentiation, and when your customer base clusters around depots you can reasonably operate. Below 250 per day, outsourced almost always wins. Between 250 and 400, hybrid models often work best.
What is a realistic last-mile software cost?
Custom build: POC £30k to £70k over 8 weeks, pilot £80k to £180k over 3 to 5 months, production £250k to £500k over 6 to 10 months. SaaS alternatives: OptimoRoute, Onfleet, Bringg, or DispatchTrack at £8 to £25 per vehicle per month plus per-delivery fees. Most brands under 15,000 deliveries per week should use SaaS.
How does route optimisation work in 2026?
Modern route optimisation solves the Vehicle Routing Problem with constraints (time windows, load limits, driver breaks). Google's Route Optimization API is the leading production choice, alongside commercial platforms (Route4Me, WorkWave) and open source (OR-Tools, VROOM). Realistic impact: 10 to 20 percent fewer miles, 15 to 25 percent more stops per van per day, 5 to 10 percent fuel saving.
How do we handle proof of delivery reliably?
Four methods. GPS-timestamped photo is table stakes for standard parcels. On-glass signature is required for regulated categories (alcohol, prescriptions, high-value electronics). Video POD reduces fraud materially in high-fraud categories. NFC or QR tap-to-confirm adds friction but strong legal proof. Design your POD stack around the categories you ship, not a one-size-fits-all approach.
How do we integrate with existing carriers as a backup?
Design so every order can route to any carrier or your own fleet, based on rules at order intake (address, priority, capacity). Show the customer one branded tracking experience regardless of who is carrying the parcel; normalise carrier events into your format. Track cost per parcel per carrier so finance can see the decision quality after the fact.
The One Thing to Remember
Owning last-mile only pays back at genuine density. Below 250 deliveries per day per depot, keep the carriers and focus on getting more out of your Shopify or OMS setup. Above 400, the maths works if the software is right and the depot is run well. Between those numbers, hybrid models are usually the safer answer while you build up.
If you want a candid conversation about whether your specific D2C operation should own or outsource, browse our custom software development services or come straight to the call.
