General
UK Last Mile Delivery: The Kerbside is a Licensed Resource, Not Free Space
Sep 9, 2026
15 mins read

Every delivery ends with a vehicle stopping. In UK cities that stop is not a free act. The right to occupy a length of kerb is granted or withheld by the highway authority, it varies by the metre and by the hour, and exercising it without permission is a civil contravention with a published price.
Routing engines do not model this. They model the road as a network to traverse and the stop as a point to arrive at, with a service time attached. The stop’s legality is treated as the driver’s problem, and the resulting penalty charge notices land in a fleet administration budget where nobody connects them back to the plan that produced them. That accounting split is why the cost persists. The kerbside is a licensed resource, and the penalty is a routing output, not an overhead.
Key Takeaways
- London PCN levels rose in April 2025 for the first time since 2011, with higher-level Band A contraventions now £160, or £80 within the discount period.
- Enforcement is a patrol process rather than automatic, so the expected penalty per non-compliant stop is a direct function of dwell time.
- Cutting dwell from 12 minutes to 5 reduces expected penalty exposure by roughly 53% at an unchanged location.
- Starting a shift at 07:00 to beat congestion can place about a third of stops inside a 07:00 to 10:00 loading ban. The early start is the illegal one.
- The ten-minute grace period applies only in a parking bay, so the same eight-minute dwell is protected in one location and exposed a few metres away.
Why the kerbside is a constraint rather than a cost line
The price is published and it moved recently. London Councils confirmed that boroughs raised parking and traffic PCN levels in April 2025, the first increase since 2011, approved by the Mayor in February 2025. The current penalty structure sets higher-level contraventions at £160 for Band A and £140 for Band B, halved within the discount period, with lower-level contraventions at £110 and £90. Higher level covers the more serious cases such as parking on yellow lines or causing an obstruction. Lower level covers contraventions where parking was permitted but the rules were broken.
The distinction matters operationally: a delivery vehicle stopped on a yellow line is in the higher band, not the lower one.
Enforcement is more nuanced than most operators assume, and it is not automated. CCTV parking enforcement was restricted in 2015 after concerns about overzealous use, civil enforcement officers must almost always be used, and a panel of adjudicators found that PCNs issued by CCTV on red routes breached Government guidance, which limits CCTV to situations where officer enforcement is impractical.
That makes enforcement a patrol process, which is far more useful than a camera would be. A camera produces a certainty. A patrol produces a probability that rises with how long the vehicle stays, and a probability that rises with dwell is something a plan can price.
Red routes tighten it further. TfL’s guidance on red route enforcement confirms that double red lines prohibit stopping at any time and single reds prohibit stopping during signed hours, that loading is permitted only in marked loading boxes during permitted hours, and critically that the ten-minute grace period applies only to vehicles stopped in a parking bay.
The pressure is increasing. The World Economic Forum’s work on the future of the last-mile ecosystem projects substantially more delivery vehicles in inner cities by 2030 with congestion rising alongside. And the leg is already the expensive one: McKinsey’s work on out-of-home delivery puts last mile at 60% to 70% of total parcel delivery cost.
How to price and plan the kerbside
1. Convert enforcement into an expected cost per stop
If a civil enforcement officer passes a given location at random intervals averaging T minutes, the probability of being observed during a dwell of d minutes is one minus the exponential of minus d over T. Multiply by the penalty a fleet realistically pays, which for a promptly settled higher-level Band A contravention is £80.
| Dwell | Patrol every 15 min | Patrol every 30 min | Patrol every 60 min |
|---|---|---|---|
| 3 min | 18.1%, £14.50 | 9.5%, £7.61 | 4.9%, £3.90 |
| 5 min | 28.3%, £22.68 | 15.4%, £12.28 | 8.0%, £6.40 |
| 8 min | 41.3%, £33.07 | 23.4%, £18.73 | 12.5%, £9.99 |
| 12 min | 55.1%, £44.05 | 33.0%, £26.37 | 18.1%, £14.50 |
| 20 min | 73.6%, £58.91 | 48.7%, £38.93 | 28.3%, £22.68 |
Patrol intensity is not yours to set. Dwell is. That single fact moves the kerbside from an overhead into a plan variable, because the operator controls exactly the term the probability depends on.
2. Treat dwell reduction as a compliance investment
Once the penalty is a function of dwell, the value of shortening the stop can be stated.
| Dwell at a 30-minute patrol interval | Probability observed | Expected penalty | Change against a 12-minute baseline |
|---|---|---|---|
| 12 min | 33.0% | £26.37 | baseline |
| 8 min | 23.4% | £18.73 | 29% lower |
| 5 min | 15.4% | £12.28 | 53% lower |
| 3 min | 9.5% | £7.61 | 71% lower |
Cutting a twelve-minute stop to five removes just over half the expected penalty. Everything that shortens the doorstep interval, from pre-arrival notification through better manifest sequencing to access instructions captured on a previous visit, therefore has a compliance return that is never credited to it, because the saving appears in a different budget from the investment.
3. Model the time band, because the early start inverts
The standard urban response to congestion is to start earlier. In UK cities that runs directly into peak-hour loading bans, and the arithmetic is unkind.
| Shift start | Hours inside a 07:00 to 10:00 loading ban | Stops exposed, 40-stop route | Share of route |
|---|---|---|---|
| 07:00 | 3 | ~13 | 33% |
| 08:00 | 2 | ~9 | 22% |
| 10:00 | 0 | 0 | 0% |
A 07:00 start places roughly a third of the route inside a window where stopping to load is prohibited on many streets. This is the opposite of the position at a border crossing, where arriving early is worth a great deal. Here, arriving early is the contravention, and the congestion saving and the compliance cost point in opposite directions.
4. Run the sensitivity before assuming it is immaterial
The exposure depends on how many stops lack a compliant place to stop, which varies enormously by borough and street. Rather than a single figure, the honest form is a range.
| Exposed stops per day | Patrol interval | Dwell | Expected cost per day | Per vehicle per year, 250 days |
|---|---|---|---|---|
| 5 | 30 min | 8 min | £94 | £23,400 |
| 5 | 30 min | 12 min | £132 | £33,000 |
| 13 | 30 min | 8 min | £243 | £60,900 |
| 13 | 30 min | 12 min | £343 | £85,700 |
| 20 | 15 min | 12 min | £881 | £220,300 |
These assume every exposed stop is non-compliant, which is the worst case and will overstate a well-planned operation that finds legal bays for most drops. The purpose of the table is not the number. It is that the number is calculable at all, and that it scales with two variables the plan sets.
5. Understand that grace depends on where you stop, not how long
TfL is explicit that the ten-minute grace period applies to vehicles stopped in a parking bay. That produces an asymmetry most operations do not encode.
| Stopping location | Grace available |
|---|---|
| Marked loading bay, permitted hours | 10 min |
| Pay-and-display bay | 10 min |
| Single yellow line, loading permitted | None |
| Double red line | None |
| Red route, outside loading hours | None |
The same eight-minute dwell is protected in one location and fully exposed twenty metres further along the same street. That makes the choice of stopping point a compliance decision at the level of the individual stop, and it is the level at which no routing engine currently operates.
6. Capture the compliant stopping point as reusable data
Drivers already know where the loading bay is on a street they serve regularly. That knowledge is not in the system. Capturing the stopping point actually used, alongside whether it was compliant and what the dwell was, converts individual experience into a network asset. The second visit to an address should inherit the first visit’s solution.
7. Attribute penalties back to the plan
The accounting change is the one that unlocks the rest. PCNs arriving in a fleet administration budget cannot influence planning. Coded back to route, stop, time band and dwell, they become a feedback signal, and the operation can see which routes and which hours generate the exposure. Without that attribution, the cost is real and permanently invisible to the people who could remove it.
Where the UK kerbside differs
| Dimension | UK urban kerbside | US urban kerbside | Continental European access zones |
|---|---|---|---|
| What is regulated | The right to stop, by location and hour | Mostly the right to park | The right of the vehicle to enter |
| Granularity | Per length of kerb, per time band | Per block or zone | Per city or zone |
| Enforcement mode | Civil enforcement officer patrol, CCTV restricted | Officer and camera | Camera, often automated |
| Penalty character | Published, banded, appealable | Varies, often absorbed in bulk | Published, tied to vehicle class |
| Operator lever | Dwell time and stopping point | Volume-negotiated tolerance | Vehicle assignment |
| Failure visibility | PCN weeks later, unattributed | Ticket, often paid centrally | Notice, tied to plate and zone |
The UK column is the only one where the binding variable is how long the vehicle stays and precisely where it stopped, which is why the lever is the plan rather than the fleet composition. The continental European position, where the vehicle’s emission class decides whether it may enter at all, is a different problem with a different answer.
Five criteria for evaluating UK kerbside capability
1. Can the plan hold a stopping point separate from the delivery address? The address is where the goods go. The stopping point is where the vehicle waits, and they are different locations with different rules.
2. Are loading restrictions modelled as time-banded constraints? Ask whether a stop can be prohibited between specific hours at a specific location, and whether re-optimisation re-checks that when the route runs late.
3. Does the platform record dwell separately from service time? Service time is what the job takes. Dwell is how long the vehicle occupies the kerb. They differ whenever the driver leaves the vehicle, which is most of the time.
4. Can PCNs be attributed to route, stop and time band? Without that link the cost cannot inform the plan, and the only available response is to pay it.
5. Does it learn stopping points from execution? A compliant stopping point discovered on one visit should be available on every subsequent visit to that address.
What this looks like in enterprise deployments
Neither of the deployments below is a UK kerbside implementation, and it would be wrong to present them as one. Both show the mechanism the problem needs: location-specific rules and per-stop knowledge held inside the plan rather than in the driver’s head.
A Fortune 50 parcel and freight enterprise operating across a 120-country network centralised dispatch across 51 sites and a 4,500-strong pool split between captive and third-party capacity, lifting weekly execution rate from 75% to 92% and surfacing more than $14 million in unused capacity. The gain came from removing a boundary that stopped one plan seeing the whole resource, which is the same structure as a kerbside whose availability is known locally and nowhere else.
A leading North American retailer running multi-hundred stores across ocean, rail and road replaced six legacy systems with a single orchestration layer, reaching over 99% on-time store delivery with route compliance above 95% and exceptions resolved in under two hours. Route compliance above 95% is the relevant figure: a plan that specifies a stopping point is only worth building if drivers follow it, and compliance is how you know whether the constraint was respected or merely recorded.
Four mistakes in UK urban delivery planning
Treating PCNs as a cost of doing business. They are a function of two variables the plan controls, dwell and stopping point, and calling them unavoidable removes the incentive to model either.
Starting earlier to beat congestion. In UK cities the early hours are precisely when loading bans bite, so the congestion saving is bought with compliance exposure.
Assuming a grace period exists. It applies in a parking bay. On a yellow line or a red route there is none, and the operational habit of a quick stop carries no protection there.
Booking dwell inside service time. If the two are one number, the kerbside occupancy that generates the penalty is invisible in every report.
How Locus supports stop-level compliance planning
Locus, the world’s first Decision-Intelligent, Agentic TMS, plans against constraints at stop level rather than at network level, which is the granularity this problem requires. The route planning system sequences against more than 250 real-world operating constraints including time windows, access restrictions, vehicle class restrictions, site access rules and driver hours, so a stop can carry its own permitted hours and its own vehicle conditions rather than inheriting a single city-wide rule.
Because plans are produced in roughly two minutes and re-optimised continuously as execution events arrive, a route that drifts towards a restricted window can be re-sequenced while the decision still has value. That matters here more than in most geographies, because in the UK the contravention is usually created by timing rather than by a planning error.
Two capabilities do the measurement work. The Driver Companion App carries turn-by-turn navigation and electronic proof of delivery with photo, scan and chain of custody, which is what allows the stopping point and the dwell to be captured as data rather than remembered. Analytics across more than 250 operational metrics, with location analytics mapping task density and order distribution, is where the attribution back to route, stop and time band becomes possible.
One boundary is worth stating plainly. Locus does not publish or maintain the UK kerbside ruleset, and no routing platform should be assumed to. Loading restrictions are set and revised by individual highway authorities on their own timetables, at a granularity of individual streets, so a UK deployment needs a maintained source for that data, whether the operator’s own survey, a commercial kerbside dataset or authority feeds, mapped into stop-level attributes during implementation. That mapping should be a named requirement rather than an assumption, and it is the part most likely to be underscoped. Locus also does not provide legal advice on contravention or appeal.
Locus supports more than 360 enterprise customers across 30-plus countries, with over 1.5 billion deliveries optimised, more than $320 million in documented client logistics savings and 99.99% uptime. It has been recognised by Gartner for seven consecutive years, featured in the 2026 Hype Cycle for Supply Chain Execution and Logistics Technologies, named a Leader in TMS by QKS Group (SPARK Matrix), and ranked #1 in Route Planning on G2’s 2026 Best Software Awards.
In October 2025, Ingka Investments, the investment arm of Ingka Group, the world’s largest IKEA retailer, acquired Locus. Locus continues to operate independently.
So how should UK last mile delivery treat the kerbside? As a licensed resource whose availability varies by location and by hour, and whose price is a routing output rather than an administrative overhead. London higher-level PCNs now stand at £160, or £80 promptly settled, and because enforcement is a patrol process rather than a camera, the expected penalty per non-compliant stop is a direct function of dwell: about £18.73 at eight minutes against a thirty-minute patrol interval, rising to £38.93 at twenty. Cutting a twelve-minute stop to five removes 53% of that exposure, while starting at 07:00 to beat congestion places roughly a third of a forty-stop route inside a peak loading ban. Locus models stop-level access restrictions and time windows inside a 250-plus constraint set, captures stopping point and dwell through its driver workflow and ePOD so the exposure becomes measurable, re-optimises in roughly two minutes so a route drifting into a restricted window can be corrected, and provides the analytics to attribute penalties back to route, stop and time band. Pair it with a maintained kerbside dataset and the penalty stops being a fixed cost and becomes a plan variable. Request a Locus route planning assessment to price your own kerbside exposure.
FAQs
How much is a delivery parking penalty in London? Higher-level contraventions, which include parking on yellow lines or causing an obstruction, are £160 for Band A boroughs and £140 for Band B, halved to £80 and £70 if settled within the discount period. Lower-level contraventions such as overstaying in a pay-and-display bay are £110 and £90. London Councils raised these levels in April 2025, the first increase since 2011.
Are UK delivery parking penalties issued automatically by camera? Mostly not. CCTV parking enforcement was restricted in 2015 and civil enforcement officers must almost always be used, with adjudicators finding that CCTV-issued red route PCNs breached Government guidance limiting CCTV to cases where officer enforcement is impractical. In practice enforcement is a patrol process, which is why exposure depends on how long a vehicle stays.
Does the ten-minute grace period cover a delivery stop? Only in a parking bay. TfL is explicit that the grace period applies to vehicles stopped in a parking bay, so a stop on a single yellow line, a double red line or a red route outside its loading hours carries no grace at all. The same short dwell can be protected or fully exposed depending on where the vehicle stopped.
Is it better to start urban delivery routes earlier? Not in most UK cities. Peak-hour loading bans commonly run through the early morning, so a 07:00 start on a nine-hour shift can place roughly a third of a forty-stop route inside a restricted window. The congestion saving and the compliance exposure move in opposite directions, and the trade-off should be made explicitly rather than by default.
How do you reduce kerbside penalty exposure? Shorten dwell and choose the stopping point deliberately. Against a thirty-minute patrol interval, cutting a stop from twelve minutes to five reduces expected penalty exposure by about 53%. Pre-arrival notification, better manifest sequencing and reusing a compliant stopping point recorded on a previous visit all contribute, and none of them is usually credited as a compliance measure.
Should PCNs be tracked against routes? Yes, and most operations do not. A penalty charge notice that arrives in a fleet administration budget cannot change a plan. Coded back to route, stop, time band and dwell it becomes a feedback signal that shows which routes and which hours create the exposure, which is the precondition for removing it.
Ishan, a knowledge navigator at heart, has more than a decade crafting content strategies for B2B tech, with a strong focus on logistics SaaS. He blends AI with human creativity to turn complex ideas into compelling narratives.
Related Tags:
General
Cold Chain Last Mile: The Excursion Happens in the Last Fifty Meters
Cold chain is monitored to the vehicle door and stops there. A bare chilled item leaves spec in about 13 minutes on a doorstep. What to measure instead.
Read more
General
B2B Last Mile Delivery: The Order is Not Final Until the Shopkeeper Agrees
In general trade the buyer can reduce the order on arrival. One door-level conversation breaks three records in three systems. How to bound and capture it.
Read moreInsights Worth Your Time
UK Last Mile Delivery: The Kerbside is a Licensed Resource, Not Free Space