The last mile now accounts for 53% of total shipping costs, up from 41% in 2018. For a two-van operation, that figure shows up in the day’s results. The gap between a route that finishes by 3 p.m. and one that runs past dark, burns an extra tank, and pushes a driver into overtime often comes down to a single planning decision. Small teams feel this faster than large carriers do, because they have fewer trucks to absorb a bad plan.
Most small delivery operations still build routes by hand. A dispatcher looks at the day’s stops, groups them by rough geography, and hands each driver a list. The method works until the stop count climbs past a dozen, at which point the human brain stops finding the shortest path and starts defaulting to the order the addresses were entered.
The Real Cost of a Hand-Built Route
Two numbers explain why manual planning gets expensive. Delivery vehicles in stop-and-go urban work average about 6.5 miles per gallon and burn close to a gallon of fuel per idling hour. Empty miles, the distance driven with nothing to deliver, reached 16.7% of all miles logged in 2024. A route that doubles back, or sends a driver across town and then back to a stop two blocks from the depot, generates both problems at once.
A small team cannot spread that waste across a hundred trucks. Five extra miles per driver per day, across three drivers and 250 working days, comes to 3,750 miles a year of avoidable driving. At urban delivery economy, that translates into fuel burned and hours paid that the operation never gets back. Labor is the largest line in last-mile work, close to 50% of the total, so every extra hour on the road is paid twice.
The Math Behind Stop Sequencing
Route planning tools solve a version of the traveling salesman problem, which asks for the shortest path that visits every stop once and returns to the start. A person can solve this for five or six stops by eye. Past that, the number of possible orderings grows too large to check by hand. A computer checks them in seconds.
The savings are documented at scale. UPS built its own routing system, ORION, which evaluates more than 200,000 route options for a single driver’s day before it settles on one, and reported cutting roughly 100 million miles a year, which the company tied to about $300 million in annual savings.
Mapping Software in a Small Operation
A small fleet does not need a custom system built by a logistics department. It needs a tool that imports a list of addresses, plots them, and returns a drivable order. This is the practical role of mapping software for a team that may run anywhere from two to twenty vehicles, alongside the spreadsheets and printed manifests a dispatcher already works from. Its job is to replace the guesswork in the sequencing step while leaving the rest of the operation alone.
The setup is usually a matter of uploading a spreadsheet. Most tools accept a column of addresses, geocode them onto a map, and let a dispatcher set constraints such as a depot start point, a return point, and time windows for stops that must land in a certain part of the day. Geocoding accuracy matters here. An address placed on the wrong side of a divided highway can add ten minutes to a stop, so a tool that flags its low-confidence matches saves a driver from a wasted loop before the day even starts.
Constraints Beyond Raw Distance
A customer who only accepts deliveries before noon changes the order. A van with a weight limit changes which stops can ride together. A driver who knows that one bridge backs up at 8 a.m. has information the map does not.
Good planning tools let a dispatcher encode these rules rather than fight them. Time windows, vehicle capacity, and required stop order become inputs the optimizer respects. Traffic is the variable that breaks a clean plan, and it is getting worse: the average American driver lost 43 hours to congestion in 2024. A sequence that is shortest among the routes that are actually legal and possible is the only kind of shortest that helps a driver.
Reading the Map After the Run
When every stop is logged with a time, a small team can see which routes ran long, which neighborhoods cost more to serve, and which new customers on the edge of the service area are worth the detour. After a month of those logs, “the west side is slow” stops being a hunch and becomes a number a dispatcher can act on.
Cutting deadhead miles, the stretches a vehicle drives with nothing aboard, is one of the fastest ways a fleet lowers both fuel use and emissions, and a small team only sees those stretches once the runs are logged.
This is where a small operation can out-think a larger one. The owner often rides along, dispatches, and reads the data personally. Decisions that take a committee at a national carrier take one person an afternoon at a local one.
A Low-Risk Way to Start
The numbers to write down are plain: total miles per van, clock time from the first stop to the last, and the count of deliveries that missed their window. Three figures across one week settle the question with the team’s own data rather than a vendor’s claim.
If the planned routes finish earlier and burn less fuel, the case makes itself. If they do not, the operation has learned something specific about its own stops that a sales pitch never would have surfaced. Either way the dispatcher ends the week with sharper information than it started with.
