Veloq
Last mile

Last-mile delivery software for grocery: why the slot decides the cost

Brian Fleming··8 min read

A $100 basket of groceries earns a typical North American grocer about $4 when the customer pushes the cart down the aisle and drives it home. Pick the same basket in the store, drive it to the customer's door, and McKinsey's estimate says the grocer loses about $13. Same eggs. Same milk. A seventeen-dollar swing, and most of it happens after the bags leave the building.

That's the problem last-mile delivery software exists to solve, and grocers know it. When McKinsey asked grocery executives which lever matters most for e-commerce profitability over the next two to three years, 52% picked reducing last-mile costs. Bigger baskets came second at 45%. Automating fulfillment, the thing with robots and press releases, came in at 38%.

So the money is in the van. The question is what the software in the van, and behind it, has to do.

What is last-mile delivery software?

Last-mile delivery software plans and runs the final leg of an order, from the dispatch dock to the customer's door. In retail, it usually covers six jobs: deciding which delivery windows to offer, turning orders into routes, assigning routes to vehicles and drivers, giving drivers an app to navigate and confirm delivery, telling the customer when to expect them, and reporting what each drop cost.

Parcel carriers have run versions of this for decades. Grocery is harder, for reasons any store manager could list before lunch. The customer chooses the time, not the carrier. The baskets are big and heavy. Some of the order is frozen, some is chilled, and the bread can't ride under the watermelon. A missed parcel gets redelivered tomorrow. A missed grocery order gets refunded, and the ice cream goes in the bin.

A grocery buyer needs software that understands time windows, temperature zones, and basket volume, and that talks to the operation picking the order.

Why is the last mile the most expensive part of online grocery?

Because delivery is now how most people shop online for groceries. McKinsey's 2026 North American research puts delivery at close to two-thirds of online grocery orders, with pickup making up the rest. Brick Meets Click found that same-day orders made up nearly 80% of US grocery delivery orders in the first quarter of 2026, and that ultra-fast delivery, an hour or less, was already 18%.

Every one of those orders needs a vehicle, a driver, and a slice of that driver's shift. Picking cost can be squeezed with better store processes or automation. Delivery cost is set by geography, time, and how many stops you can string together in an hour. You can't automate your way out of a traffic jam on Route 1, and every extra minute a driver spends between stops comes straight out of the order.

Which levers move last-mile cost?

Four of them do most of the work.

Drop density. The more deliveries a route completes per hour, the less each one costs. Density depends on how many customers order in an area, and on whether the windows they book let you group neighbors onto the same route.

Slot design. The windows you offer decide how much freedom the router has. Offer every customer every window, and you get vans crossing town to hit a 7 p.m. slot on one side and a 7:15 slot on the other. Price, limit, or close windows by area, and you steer demand toward routes that work.

Sequencing. Within a route, stop order matters, and so does the vehicle's load plan. Frozen goes last into the van and first off it at the right stop. Basic, and easy to get wrong at scale.

Fleet mix. Own vans, contracted delivery partners, third-party couriers. Each has a different cost per hour and a different level of control over the customer experience.

Notice that two of the four are decided before a single route exists. That's why the most important feature in grocery last-mile software sits at checkout.

How do delivery slots affect cost and customer retention?

The slot is where cost and loyalty meet, and plenty of grocers still treat it as a website setting.

Shoppers have firm views on the window. In McKinsey's 2026 consumer survey, 52% of shoppers who prefer delivery named scheduling flexibility as a reason, second only to saving time at 67%. Asked when they expect an order to arrive, 39% said the same day, 29% said within two to four hours, and 14% said within an hour.

So customers want to choose the time, and they want it soon. Wide windows give the router slack. They also ask the customer to stay home for two hours, which is a lot to ask of anyone with a job, a dog, or a sense of self-worth.

Narrow windows are harder to run, and the payoff is a customer who can plan around you. Rohlik Group offers 15-minute delivery windows across its markets, and more than 94% of its orders arrive on time and error-free.

The catch is that a 15-minute promise only works if the software behind it is honest. A narrow window the van can't hit is worse than a wide one. Slot capacity has to come from live picking and routing capacity, recalculated as orders arrive, so the customer is only offered windows the operation can keep.

What should grocers look for in last-mile delivery software?

Here's the checklist I'd bring to a vendor demo.

Slot offers built from live capacity. Ask how the system decides which windows to show. If the answer involves a static table someone updates on Mondays, keep walking.

Integration with picking. Routes and pick waves should be planned together. If the warehouse or store picks in the wrong order, the perfect route leaves late.

Multi-temperature handling. The driver app should track chilled and frozen totes separately and record handoff, so a cold-chain problem shows up in the data before it shows up in a complaint.

Real-time re-planning. Drivers get sick, vans break, a customer adds three items an hour before cutoff. The system should re-plan without a dispatcher redoing the day by hand.

ETA accuracy, measured. Ask for on-time rates from live customers, not simulations.

Fleet flexibility. Own fleet, partners, and third-party couriers, managed from one place, with cost per drop visible for each.

Cost reporting by drop. If you can't see what each delivery costs, you can't fix slot design or pricing.

What comes next: AI routing and autonomous delivery?

BCG's 2026 analysis of grocery delivery automation sees change arriving in waves. In the first, already underway, AI improves forecasting, routing, and substitution and cuts inefficiencies by 10% to 20%. Autonomous vans and sidewalk robots sit further out: BCG puts them more than ten years away, at which point they could cut last-mile costs by 30% to 50%.

Some of it is already on the street. Walmart announced its one-millionth drone delivery in May 2026, with an average delivery time of 23 minutes. Drones carry small baskets, so they suit the top-up order more than the weekly shop, but they're one more vehicle type for the dispatch system to plan.

For a buyer today, the practical point is that the routing and slotting software you choose now is the system that will eventually dispatch those vehicles too. Pick one that can add a new vehicle type through configuration.

Build, buy, or rent from a marketplace?

Grocers have three ways to get delivery done. They can build their own stack, buy software and run their own operation, or hand delivery to a marketplace that brings its own shoppers and drivers.

Marketplaces are the fastest route to coverage, and plenty of large grocers use them alongside their own fleets. The trade-off is control: over the delivery window, the customer experience at the door, and the data on what customers buy and when. Building in-house gives full control and a long list of things to maintain. Buying software and running your own fleet sits in between, and it's where most grocers with steady volume end up.

The right mix depends on density. In dense areas with steady demand, owned delivery tends to pay back. In thin areas, a partner may be cheaper than a half-empty van.

Questions grocers ask about last-mile delivery software

What's the difference between last-mile delivery software and route optimization software?

Route optimization is one module. Last-mile delivery software also covers slot management, dispatch, the driver app, customer ETAs, and proof of delivery.

How much does last-mile delivery cost a grocer?

It depends on density and slot design. McKinsey's estimate of a $100 basket moving from a $4 profit in store to a $13 loss when picked in store and delivered shows how much is at stake.

Can last-mile delivery software work with store-based fulfillment?

Yes. Routes can start from stores, micro-fulfillment centers, or dedicated fulfillment centers. The software needs to know which node is picking each order.

Does same-day delivery cost more than next-day?

Not necessarily. Same-day with good density and narrow, well-managed slots can cost less per drop than next-day with scattered orders. Our report on profitable same-day grocery delivery walks through the numbers.

Planning the slot, the pick, and the route together

The grocers who make delivery pay do the boring part well. They decide which windows to offer based on what the vans can actually do, pick orders in the sequence the routes need, and measure every drop.

That's the approach behind Veloq's last-mile software, which runs same-day delivery for Rohlik Group across five countries. The same system offers the slot, plans the pick, and builds the route. When Rohlik's German business moved from next-day to same-day on the platform, Bringmeister's monthly orders went from 57,756 to 134,860 in the same fulfillment center.

Seventeen dollars is a lot to give up between the store and the doorstep. Much of it gets decided at checkout, the moment the customer picks a window. See how Veloq's last-mile software works.

Sources

See the platform behind same-day grocery

Talk to the team that runs fulfillment, routing, and storefront software for Rohlik Group's twelve fulfillment centers.

Book Demo Workshop