Design principles for mass cycling - Embassy Policy

This is a draft version of our policy position on general design principles for cycling. We welcome comments below!

Executive Summary

The conditions required for mass cycling are well known: people are likely to choose cycling for their journey if they believe their route will be safe, comfortable and convenient.

In practice, this means providing cycling facilities that are separate from walking and driving facilities, and it means designing our urban and rural environments so that cycling is an obvious choice, minimising interactions with motor traffic, wherever people choose to cycle. The focus should be on providing attractive and safe conditions, rather than attempting to mitigate hostile conditions through training, or education, or changes to insurance law.

Mobility scooter and cycling de Bilt

When creating facilities for cycling, imagination, care and thought - and practical examples of designs now emerging in the United Kingdom - should allow us to be able to emulate the cycling conditions in countries like Denmark and the Netherlands. In particular, the principles of the Dutch system of Sustainable Safety should underpin designing for cycling.

We would call this Intrinsic Safety - safety built-in to the design of the road and street environment.

Consistent application of these principles will make our roads and streets safe, attractive and coherent - not just for cycling, but for all modes of transport that use them.

The main principles of Sustainable Safety [or Intrinsic Safety] are as follows.

  • Single function roads - roads and streets should have a single function, either as access, distributor, or through roads.
  • Homogeneity - the mass, speed and direction of objects/people/vehicles should be equalised as much as possible on roads and streets.
  • Instantly recognisable road design - roads and streets should be self-explaining, conveying unambiguous messages about the kind of behaviour expected. Design should be consistent, and instantly recognisable.
  • Forgiving environments – roads and streets should be forgiving environments, taking into account human fallibility, limiting the consequences of mistakes by users in any transport mode.

General policy requirements

1) Policy on the use of road and street space

The principle of single function roads means that a clear division needs to be made between ‘roads for transport’ and ‘streets for people’, where ‘people’ refers to the entire social, leisure and commercial uses of streets.

The function of a street can’t be changed simply by changing its appearance. On ‘streets for people’, the objective would be to reduce motor traffic to just those vehicles which need access to the homes, shops and businesses on those streets. These are what the Dutch would term ‘access roads’.

Gouda city centre

On ‘streets for transport’, cycling should be designed for in its own right, built-in to the road and street environment, rather than simply bodged into pedestrian-specific design, or treated like motor traffic.

We recognise that at present in the UK many streets and roads are multifunctional, serving both transport and ‘place’ functions. Over time these roads and streets should be moved into one category, or the other. Highway Authorities and Local Authorities should be given a duty to categorise roads according to function using the principles of Sustainable Safety -

  • Access Roads - giving access to houses, schools, shops and businesses
  • Distributor Roads - taking traffic locally from through routes to access roads, and
  • Through Roads - carrying large volumes of traffic between centres of population, or past them.

Where roads currently have a mixed purpose, over time planning decisions should reinforce the primary categorisation and minimise any inconsistent uses, with the long-term objective of moving towards ‘monofunctional’ streets and roads. Not only is this is essential for determining appropriate cycling provision, more generally it would make road transport more efficient, help revitalise town centres and encourage other forms of active travel.

2) Road design standards

The principle of Homogeneity means that highway authorities should have a set of mandatory standards for the accommodation of cycling, according to the above categories of road (access, distributor, through road), taking into account traffic levels and speeds.

The classification of each road/street type should reflect the amount and composition of motor traffic using it. In Britain this would most likely involve moving what are currently distributor roads into an access road role, by lowering or removing through traffic. Those roads that remain assigned as distributor roads should have separate cycle provision along them. By this principle, cycling - a low speed, low mass mode of transport - should not share space, anywhere, with motor traffic travelling at higher speeds, or in significant volumes.

These mandatory standards would incorporate ‘separation requirements’, defining the treatment needed in all cases, including at junctions and roundabouts, taking the guesswork out of cycle provision, and creating the same consistency for the use of cycling infrastructure as drivers usually experience on the roads. An example of separation standards, based on Dutch principles, can be found here.

These road types should be instantly recognisable, meaning that drivers in particular would know what kind of behaviour is expected on a given road or street. Through roads and distributor roads should be designed in very different ways from access roads, with separate provision, or parallel routes, for cycling.

Equally, road and street design should be forgiving, allowing for human fallibility and mistakes. For instance, the higher the (measured) speeds on a given through or distributor road, the greater the degree of separation from walking and cycling. On access roads where space is shared between cycle and motor traffic, design should ensure that the volume and speed of motor traffic is kept low, again, to minimise the consequences of human error.

Such a framework would replace all current DfT guidance on cycling. These standards would be comprehensive and non-negotiable, and could influence the function and categorisation of roads.

For example, if there is (genuinely) ‘not enough space’ on a through road for cycleways or cycle lanes, it is the road which would ultimately have to change in categorisation and character: by having the number of traffic lanes reduced (including making roads one-way for motor vehicles), or by being taken out of the through-traffic network entirely, and having traffic levels reduced to the point where cycle tracks or lanes are not needed. In this way all roads and streets would ultimately be made cycle-friendly.

In general, a duty on highway and local authorities to assign a single function to their roads and streets should govern this process, making it clear which roads and streets are access roads, and which are through roads, in turn determining whether cycle-specific provision is required.

3) Implementation

Whenever roads are built, re-engineered, resurfaced or rebuilt, the Highway or Local Authority responsible would have a duty to build the appropriate cycle facilities consistent with that category of road, according to the new standards. Thus the normal cycle of road maintenance and rebuilding will simultaneously push the development of cycling infrastructure, making best use of resources. There should be no road schemes, of any kind, which do not include high quality cycling facilities designed into them from the start.

The responsible Highway or Local Authorities should also have long-term plans in place to reclassify, redesign or rebuild the existing highway network, in line with the principle of monofunctionality, with high-quality cycling infrastructure developed alongside distributor and through roads, and motor traffic removed or redistributed away from designated access roads.

These road and street types should again be instantly recognisable, and forgiving, responding to the fact that human beings are fallible and will make mistakes.

Cycle-specific policy requirements

Beyond these general requirements, cycling needs specific policy measures for it be an attractive, convenient and safe mode of transport, for all potential users. These measures can be grouped under five main requirements.

1) Safety

Measures to ensure that cycling is a safe mode of transport flow from the Sustainable Safety requirements detailed above. In essence, the number of encounters with motor traffic should be minimised, both along road sections, and at junctions. Where speed differences are high, cycling should be physically separated. Where the mass of motor traffic is great, again, cycling should be physically separated. (Notably, this means that cycling should not be accommodated within bus lanes).

Where cycling does encounter motor traffic (for instance, on access roads, and at minor junctions) speed differences should be minimised as much as possible, by means of the layout of the street, or junction. Conflicts at major junctions should be avoided by signal separation, or grade separation, depending on the speed of motor traffic.

2) Comfort

Principally, this means insulating cycling from the effects of motor traffic. Nobody should have to cycle with the flow of motor traffic on roads and streets carrying more than 2000 PCU per day, which equates to approximately 200 PCU per hour, at peak times.

To make cycling a comfortable experience, Motor traffic should either be reduced on these streets to below that threshold (on access roads), or, cycling should be physically separated from that motor traffic (on distributor or through roads), with cycle-specific infrastructure.

Comfort also involves ensuring that

  • cycling takes place on smooth surfaces, without undulations;
  • routes are sufficiently wide to accommodate (future) demand, and minimise delay;
  • routes are not excessively steep;
  • delay or waiting is minimised;
  • ensuring that routes do not involve excessive waiting or delay, excessive gradients, or sharp corners that involve loss of momentum, and that
  • nuisance from weather (wind, rain etc) is taken into consideration in design.

Cycling infrastructure should be designed to suit cycling. Cycling should not be accommodated on pedestrian-specific infrastructure. Walking and cycling are distinct modes of transport, with separate requirements, and should be designed for separately, giving each mode of transport sufficient space for comfort. 

Only in areas where few people are walking is sharing acceptable, and this should be on infrastructure designed for cycling. The threshold for acceptable sharing should be 100 pedestrians per metre of width, per hour. Above this threshold, walking and cycling should be separated.

These comfort (and safety) criteria should inform 'separation standards' - a typical example for within urban areas, below.

Cycle provision in built-up areas

3) Directness

Routes for cycling should be direct, both in terms of distance, and in terms of time.

In distance terms, main cycle routes should deviate from the 'crow flies' distance by no more than a factor of 1.2. Excessive bending and winding of routes should be minimised. At junctions movements should not be illogical or circuitous.

In terms of time, stops and delay for people cycling should be minimised as much as possible, particularly on main routes. This includes a design speed of 20mph for cycling infrastructure.

4) Attractiveness

Cycling infrastructure should be socially safe. It should be well-lit, with social surveillance from the surrounding area, and well-maintained.

Cycleway Houten

5) Coherence

In urban areas, cycle routes should form a dense network, that connects people with the places they want to go to. In practice this will mean a grid with routes no more than 250m apart. In rural areas, cycle routes should form direct connections between settlements and amenities.


Perhaps most importantly of all, these attributes should not be traded off against each other. No-one should have to choose between these attributes when making a journey by bicycle; between fast, direct and convenient routes, and routes that are safe and comfortable. Every route should be suitable for all potential users, fast or slow, experienced or new to cycling, old or young - direct, safe, attractive and comfortable.


"For example, if there is (genuinely) ‘not enough space’ on a through road for cycleways or cycle lanes, it is the road which would ultimately have to change in categorisation and character: by having the number of traffic lanes reduced (including making roads one-way for motor vehicles), or by being taken out of the through-traffic network entirely, and having traffic levels reduced to the point where cycle tracks or lanes are not needed. In this way ALL ROADS AND STREETS would ultimately be made cycle-friendly."

I think this is too radical. For example, let's say that we have a road that connects cities that are 20 km apart, but on some sections it's too narrow for cycle paths/lanes. Closing this road for through car traffic would mean that the route for cars would become 40 km long. The road carries 3000 cars a day and estimated cycle traffic, when cycle paths are built, is 30 cyclists per day. Sometimes a small gain for cycling would mean disproportionate looses for other modes.

Two small comments on section 2 "Comfort":

The point "ensuring that routes do not involve excessive waiting or delay..." seems to duplicate the previous two points.

The number 100 pedestrians per metre of width, per hour before providing separate infrastructure sounds fairly high to me: on a 2m wide cyclepath that would mean more than 3 pedestrians per minute. Not sure what the CROW recommendations are, but I'd think that in NL this is typically lower.

Exactly - David Hembrow says "Almost none is "too many". That's why you always find a separate path for pedestrians anywhere that appreciable pedestrian flows are expected", whilst 100 pedestrians per metre of width, per hour would mean 300 pedestrians per hour on a 3 m wide cycle path (two way cycle paths usually shouldn't be narrower), which is approximately 3000 pedestrians per day! (At least if we can estimate number of pedestrians similarly as we do with PCU.) Thousands of pedestrians certainly isn't "almost none"! Again, probably "testing" these recommendations in practice would reveal such mistakes.

The topic of Sustainable Safety has been covered in many places, but there are significant differences. For example, there seems to be quite a bit of confusion with distributor roads.

First, we have a definition from SWOV:
"connecting roads have been defined and are called distributor roads. This road type has a
flow function on road segments and an exchange function at intersections, and connects through roads with access roads, AS WELL AS THROUGH ROADS AND ACCESS ROADS AMONG EACH OTHER."

Then from Bicycle Dutch blog:
"Distributor Roads which connect through roads AND local access roads." (still correct definition if you understand what it means)

Then from Pedestrianise London blog:
"Distributor roads (aka district access roads) - streets that provide access to estate access roads, non-direct through routes." (OK, but... "non-direct through routes"? What's this?)

An finally in this article (in fact, I believe some other people express it similarly):
"Distributor Roads - taking traffic locally from through routes to access roads"
This isn't really a correct definition, because it would mean that e. g. a road which connects two districts in a city (i. e. connects access roads between that districts) shouldn't be classified as a distributor road (And what instead? Through roads shouldn't, in principle, be connected directly with access roads, whereas access roads' main role is, of course, access to buildings).

Taking inspiration from the Dutch is a good thing, but it should be noted that in the Netharlands these guidelines are not mandatory, while in cycle-hostile countries we need as much mandatory rules as possible, because otherwise engineers and officials would very often try to find excuses for not following the recommendations. On the other hand, we cannot legislate everything. A classic example is described on Bicycle Dutch blog:

"In one of the former main streets that was supposed to be blocked, the residents were against the closure in such a large majority that the city decided to reverse that decision. The street did become a 30km/h zone, but with a direct access to a main arterial road (number 4 on the map above). This is in conflict with the rule that no 30km/h street has access to a main artery, but these exceptions to plans will always be there in reality."

Closing every single junction with an access road would probably be even unreasonable, because a through road with one or two junctions with access roads isn't necessarily less safe than some other through road which has only junctions with distributor/through roads, and nobody wants to make long detours just "for the sake of the idea". What really counts is the number and density of junctions, not the type of the roads crossed.

The table linked here does allow junctions with access roads on through roads (yet another inconsistency), but only "grade separated or TCS" (even if traffic volumes are very low) which probably wouldn't make sense in Utrecht's case either - traffic lights are not only more expensive (especially with regards to maintenance) but also less preferred by road users than a roundabout, and in this case they could only solve a non-existent problem.

So my idea is to look on a map and try to categorise as much roads as possible, and then check if proposed rules work in real life - are there enough exceptions, but also do they solve real problems we have?

"In distance terms, main cycle routes should deviate from the 'crow flies' distance by no more than a factor of 1.2."

This doesn't sound realistic to me, or maybe needs more explanation. I calculated this factor for some of the trips I frequently make, and the average was 1.43 by bike, and 1.44 by car. Only one of my trips was close to this, at 1.21 by car (1.35 by bike.) The rest ranged up to 1.64 by bike, 1.73 by car, but that's the shortest trip, a ~1.8 mile dogleg to cover 1.1 miles. The trips around 1.4 don't seem frustratingly inefficient to me.



@Andy S

That's right, the 1.2 factor should apply to bike vs. car (or pedestrian) trip, not "crow flies" distance.

As a lifelong cyclist and someone who has spent half my adult life in the Netherlands (Utrecht) and half in the UK (Berkshire, London and Cambridgeshire) I think I am fairly well qualified to comment.

This is a very good translation of how to implement something similar to what the NL do in the UK and looking at the critisim in the comments, what's getting lost in translation is that this would have to be a long term strategy of renewal and implementing the principles to all new developments. We are dealing with multi purpose chaos, as the peice says and it would take time to change this.

Cambridgeshire is a county where there is alot of farmland being turned into new developments, plus there are major roads that are long overdue both upgrading but in some cases, much as I am against building for more road transport, they could really do with dualing and straigtening them out a bit. East Anglia is notoriously bad for both it's levels of traffic violence and ignoring the fact that it could be a cycling utopia - it's more or less flat, the land is not hard to adapt - it was only a century or so ago it was drained and as a holiday destination, it's seriously under provided for for cyclists once you leave the chaos of Cambridge.

If I look at all the bits around Cambridge and Ely that simply weren't there 20 or so years ago, you can see how rapidly things grow and are growing - by implementing these principles in the new developments, it would still be there in 20 years, growing, experimenting doing what Milton Keynes should have done - they started, and just gave up! You don't just give up on a property, if it doesn't work, you fix it, if that fails or expires due to irrelevance or gets overtaken by something new and better, you fix it again!

London is constantly being developed and adapted, it has budgets that are often large and largely squandered on projects that don't live up to their expectations and given car ownership must be pretty much the lowest in the country, should be possible to do something about this with a legally well set out strategy like this. Back to the lost in translation, I mentioned - with a long term plan in mind and with each set of scheduled renewal and redevelopment, it could be done in a generation. 

Where I live currently, Utrecht, it's the 4th largest city and on many levels gets used to experiement both infrastructure and socially on a large enough scale to be meaningful, but at a cheaper scale to be viable - this way Utrechters do sometimes feel like gineau pigs but often are proud to be getting to try out something that will benefit us all in the future. Places like Brighton, Cambridge, Bristol, Reading, could be used in simlar ways to trial provision that could work in London, not the other way round - try to push it though in London, fail, and not see anything happen.

I live in a large, 1960's brutalist development, famous in Holland for being overpopulated with foreigners, crime and is 70% social housing, sounds horrible right? No actually its really nice! Since the problems started 30 years ago, the council set to work with projects and ideas that would ultimately go on to be used in similar problem areas of Rotterdam and Amsterdam. Overvecht is Hollands first peaceful neighbourhood, with mediators, teams of social workers and project workers who manage the many different aspects. Overvecht has excellent cycle provision, trees and play parks on every corner, heathcare centres, the 2nd largest shopping centre in the city, plus every half mile or so, more local shops, lots and lots of schools - one of which has four locations all with the dutch equivelent of Ofsted outstanding, despite being 70% children who are 'black'.

My point is, it's not standing still, they messed up, there's no denying it, they tried to build too many affordable houses at once, you could say, the UK did it too. The difference is, they saw the value of not letting it fester and infect, they are winning and it's actually really safe, friendly and coming from someone who has lived in some of the nicest parts of the UK, I would go as far as to say, my quality of life here is much higher, my stress levels are lower.

For those in the UK who want to see better cycle provision, you have to think holistically, put people first across the spectrum and not just think of the bikes - it needs to work for everyone if it's to work at all, which of course, also means compromising. But compromise everywhere but safety, because if you don't feel safe, and safe to let your kids out on the street to play, you cannot be happy.


A few possible tweaks occurred to me:

- in the spirit of unbundling, you could use an access/distributor/through route classification to describe, design and most importantly, discuss cycle infrastructure as a separate thing to the road classification. For example, if you classified the E-W Superhighway as a through route for central London, or NCN4 in Surrey (after a lot of work, obviously), then you'd want to put in cycle distributors at various points. Etc.

- adequate cycle parking needs mentioning somewhere (under attractiveness?)

- it might be worth emphasisng that making a route as close to crow flies as possible would also involve identifying and removing critical existing severance.

On the last point, I leave you with this from Goes in the Netherlands.

That's a non-motor level crossing in the distance. In the extreme case that you wanted to get from one end of the path to the other, the motor:bicycle trip distance ratio would be 35:1 !

I welcome the proposals for a statutory duty on authorities to categorise roads according to function and for mandatory standards for infrastructure. I would like to see further duties on the relevant authorities:


  • to prepare a detailed action plan, based on an assessment of the state of the existing infrastructure and the work necessary, for achieving full compliance with mandatory standards. These standards should include a reasonably achievable timetable for compliance.

  • to make a formal statement that they recognise that cycling infrastructure is as important as that provided for motor vehicles and pedestrians.


Without these it is likely that local political considerations will prevent the achievement of the policy requirements.


This is a statement of policy, so does not carry much detail. But more could be included about construction standards. Cycling infrastructure needs to be built without obstructions. Too often tracks are partly blocked by direction sign poles, lighting columns, trees, bollards etc. Cycling surfaces should not just be smooth, they should be hard, that is, paved.


There is a significant omission - under the heading ‘Comfort’ there is no reference to the importance of ongoing maintenance of infrastructure. Good infrastructure will not remain so unless the relevant authorities commit resources to maintenance just as they do with motoring infrastructure. Surfaces will deteriorate. Cycle tracks will accumulate litter and be affected by ice and snow, leaf fall, mud etc. They may be used for parking or for storage of dustbins. Hedge and tree growth etc will block them. Local businesses will put advertising signs in them. Road repairs will block them. All these problems must be firmly nipped in the bud, just as they would be if they occurred on the carriageway. Sometimes diversions will be needed and should be of a good standard and well signposted.