Every routing vendor demonstrates the same thing: a messy map, a button, a tidy map. What they rarely explain is what the software was told to make tidy. Courier route optimisation software does not find “the best route”. It finds the best route against a specific target, and if that target is the wrong one for your operation, you will get a shorter round that loses you money.
The pressure to get this right keeps rising. Department for Transport figures show van traffic in Great Britain reached 58.8 billion vehicle miles in 2025, some 10% above pre-pandemic levels, while lorry traffic fell to 16.3 billion miles and sits over 5% below 2019. More vans, doing more drops, on roads that have not grown.
This guide is for UK courier, same-day and multi-drop operators who are past the question of whether to use routing software and are now deciding which one, and on what basis.
What is courier route optimisation software?
Courier route optimisation software calculates the best sequence and allocation of stops across a fleet, subject to constraints such as customer time windows, vehicle capacity, driver hours, access restrictions and depot start times. It answers what order the committed work should run in, and on which vehicle.
It is worth being clear about what it is not. It is not a satnav, which sequences nothing and simply navigates to the next address you give it. It is not capacity planning, which decides how much work to accept in the first place. And it is not a live traffic app, though a good planner consumes traffic data as an input.
| Tool | What it decides | Where it stops |
|---|---|---|
| Courier route optimisation software | Stop sequence and vehicle allocation across the fleet | Cannot rescue a round that was over-committed at booking |
| Courier route planning software | Daily plan construction, often with manual override | Weaker on live re-optimisation mid-round |
| Satnav or consumer mapping | Turn-by-turn navigation to one destination | No sequencing, no constraints, no fleet view |
| Capacity planning | How much work the fleet can accept | Does not sequence the accepted work |
| Manual planning | A plan built from local knowledge | Degrades fast above roughly 25–30 stops |
Key takeaway: Routing software sequences committed work. It cannot undo a bad acceptance decision made hours earlier.
What are you actually optimising for?
The objective function is the target the software minimises or maximises. Most disappointment with delivery route optimisation software traces back to an objective that did not match the commercial reality of the operation.
| Objective | Best suited to | What it costs you |
|---|---|---|
| Shortest distance | Fuel-sensitive rural and trunking work | Can produce long, slow urban rounds with poor drop density |
| Shortest total time | Urban multi-drop and same-day | May add mileage to avoid congestion, raising fuel spend |
| Lowest total cost | Mixed fleets with different cost-per-mile | Only as good as the cost data you feed it |
| Maximum on-time delivery rate | Time-window and SLA-driven contracts | Deliberately leaves slack, so utilisation looks worse |
| Balanced workload | Employed drivers and consistent rounds | Sacrifices some efficiency for fairness and retention |
A same-day courier optimising for distance is the classic mismatch. Distance-minimised rounds cluster geographically but ignore promise times, so the round looks efficient on a map and breaches half its windows. Conversely, a pallet operation optimising purely for time will happily add fifteen miles to dodge a queue that would have cost four minutes.
The practical test at demo stage: ask the vendor which objective the demo used, and ask them to re-run it under a different one. If they cannot, the objective is hard-coded, and you are buying their assumption about your business.
Key takeaway: Decide what “better” means for your contracts before you evaluate software that promises it.
UK constraints that generic route planners miss
Route optimisation for courier companies in the UK carries constraints that internationally built planners frequently ignore. These are the ones that cost real money:
- Clean Air Zone and ULEZ charges. Seven Clean Air Zones operate in England alongside London’s ULEZ, and GOV.UK confirms they charge 24 hours a day, midnight to midnight, with payment due within six days or a penalty charge notice follows. A planner that does not hold each vehicle’s emission standard against zone boundaries will route a non-compliant van through a zone for a two-minute saving and a daily charge.
- Driver hours and break rules. A sequence that is geographically elegant and legally non-compliant is not a plan.
- Kerbside and access reality. Red routes, bus lanes, weight limits, low bridges, gated estates and pedestrianised hours. Van-specific routing matters; car routing does not know your vehicle is 3.5 tonnes.
- Narrow time windows. UK retail and trade customers increasingly specify two-hour windows, and treating them as soft preferences produces plans that fail on contact with the day.
- Rural drop density. A round in Powys and a round in Peckham need different objectives, and a single global setting will serve one of them badly.
Key takeaway: Ask which UK-specific constraints the planner models natively, not which it can be configured to approximate.
What distance savings are actually worth
Vendors quote percentage distance reductions. Those percentages are worth converting into money before they influence a decision. Realistic gains against a competent manual planner run 5–12%; larger claims usually reflect a poor baseline rather than exceptional software. GOV.UK weekly road fuel price data put diesel at roughly 165 pence per litre in mid-July 2026, having peaked above 190 pence in April, with fuel duty fixed at 52.95 pence throughout. Build a business case on a single week’s price and it will look very different a quarter later.
Multi-stop route optimisation: what changes above 30 stops
Multi-stop route optimisation software becomes genuinely necessary at the point where the number of possible sequences exceeds what a person, or an exhaustive calculation, can evaluate. That threshold arrives earlier than most operators expect.
The combinatorics are unforgiving. Working from a fixed depot, ten stops allow over 360,000 possible orders and twenty allow more than 100 quadrillion. No solver evaluates them all. Good software uses heuristics to return a very strong answer quickly, which has three practical consequences:
- Two runs can differ. A heuristic result is not unique. Small input changes can reorder a round, which unsettles drivers unless you explain why.
- Constraint quality beats solver power. Accurate service time, correct time windows and real vehicle profiles improve results far more than a faster algorithm.
- Re-optimisation matters more than the first plan. Days deviate. Dynamic re-routing when a drop fails or a job is added is where daily value accrues.
Below about 25 stops, an experienced planner with local knowledge is often competitive. Above 30, they reliably are not, and the gap widens with every additional stop and constraint.
How to evaluate courier route planning software
Score any delivery route planning software against these before the conversation turns to price:
Modelling accuracy
- Vehicle profiles by weight, height and emission standard, not one generic vehicle
- Service time configurable by customer or drop type
- Time windows enforced as hard constraints, with visible breach warnings
- Historic traffic by time of day, not a flat average speed
Operational fit
- Selectable objective function, changeable by round or contract
- Dynamic re-routing during the day when jobs are added or fail
- Driver app that delivers the sequence, with ePOD captured against it
- Planner override that is recorded, so you can measure trust in the plan
Proof and integration
- Planned against actual reporting on distance, time and drops
- Telematics or GPS feed so actuals are measured, not self-reported
- API links to your order source, so plans build from live jobs
The planned-against-actual line is the one that separates serious platforms from map-drawing tools. Without it you cannot tell whether the plan was good or the driver simply ignored it. Pair the planner with your wider route optimisation software and dispatch stack so the same constraints govern both.
Why route optimisation projects fail in practice
The software is rarely the problem. Five patterns account for most disappointing rollouts:
- Dirty address data. Ungeocoded or approximate postcodes send vans to the middle of a postcode sector rather than a door.
- One global service time. A flat five minutes across residential drops, trade counters and site deliveries makes every plan wrong in both directions.
- Driver override without measurement. Experienced drivers deviate, which is sometimes correct. If deviations are never recorded, the model never learns and trust never builds.
- Optimising a bad acceptance decision. If the round was over-committed before planning began, better sequencing only redistributes the failure.
- No feedback loop. Service times and traffic drift. A model never retrained on actuals decays within months.
The driver point deserves emphasis. Driver adoption is the single biggest predictor of whether a routing project delivers, and it is won by showing drivers the plan respects their constraints, not by mandating compliance.
Summary
- Courier route optimisation software sequences committed work; it cannot fix over-acceptance.
- The objective function decides everything. Match it to your contracts before comparing vendors.
- UK-specific constraints, especially Clean Air Zones and vehicle-appropriate routing, are where generic planners leak money.
- Time savings are typically worth three to four times the fuel savings.
- Above 30 stops, constraint quality matters more than solver sophistication.
- Measure planned against actual, or you will never know whether the plan worked.
