Evaluation of the e-scooter trial in the West of England

So far as the Safety Theme is concerned, there have been five main sources of data including trial operator data, collision data from Avon and Somerset Police, data from two hospital studies, and primary data from video observations of road user interactions and survey data from e-scooter users and non-users.

The Transport Policy Theme is evaluated using trial operator data, and data from five primary sources as follows: on-street surveys, follow up in-depth interviews of users and non-users, stakeholder interviews and objective measures of e-scooter parking quality. Stakeholders were drawn from organisations that were involved with the trial, including transport authorities and emergency services.

The Wider Impacts Theme has drawn on trial operator data and three sources of primary data: on-street surveys, follow up in-depth interviews with users, and stakeholder interviews. In addition, the carbon analysis has drawn information from the literature on vehicle emissions.

The Management Theme, which includes the issue of parking as a major sub-category of interest, used trial operator data and four primary data sources as follows: video observations of parking locations; follow-along surveys with non-users, stakeholder interviews, and in-depth interviews with users.

E-scooter usage. There have been over nine million rides since the start of the trial in October 2020 to the end of February 2023. The users are predominantly younger adults (49% of all rides across the trial areas were made by 18-24 year-olds up to April 2022 with only 1% of rides made by those 55 and above).

Of the users who reported their gender as male or female, 2.8 times more rides have been made by males than females. Rental e-scooters are serving larger number of young adults without access to a car or bicycle. They are being used to a similar extent for work and education purposes and for social and leisure purposes.

Rides are largely of short distances with one half of trips less than a 25 minute walk (2.1 km) in Bristol and one half of them less than a 19 minute walk (1.6 km) in Bath. Long-Term Rental (LTR) users contribute to 5% of total rides and distance travelled in the trial area even though they only make up 1% of total users.

About 15% of registered e-scooter users are active users who use the e-scooters at least once per week but this still represents a substantial number of people (42,200 people in April 2022). Most rides have been paid for on a pay as-you-go basis (56%). Daily and monthly passes represented respectively 18% and 26% of the rides.

Available data from trial operator surveys suggests that about 60% (range of 58% -65%) of e-scooter trips are replacing walking, cycling and bus use with about 30% (range of 24% - 37%) replacing car, taxi and ride-hailing.

The latest data from the trial operator’s 2022 Winter Survey suggests that the modes replaced by an e-scooter in Bristol in descending order are walking (32%), bus (18%), car (15%), bicycle (14%), taxi and ride-hail (9%). Only a small minority of trips (1%) would not have been made if an e-scooter was not available.

The surveys had an under-representation of young e-scooter users and frequent users and both of these are less likely to have used a car instead of an e-scooter. Car replacement is greater amongst older users and less frequent users while bus replacement is more common among 18–34-year-olds and more frequent users. This points to the survey mode substitution figures overestimating car and taxi substitution and underestimating bus substitution.

In-depth interviews of 13 e-scooter users suggested e-scooters are replacing walking, cycling and bus and taxi use. The e-scooter has become the first mode of choice for travel within Bristol for some users and an option that is selectively chosen for particular situations by others. Rental e-scooters have been added to people’s urban transport menu of options with the relative amount they are used varying from person to person.

Trial e-scooters appear to be riskier in the West of England area than cycling in Great Britain urban areas by a factor of about 1.8 based on number of injuries per distance ridden.

The e-scooter injury rate in the West of England is 0.53 casualties per 100,000 km ridden and the national urban cycling rate is 0.294 per 100,000 km. Note that this comparison is based on those injuries reported in the STATS19 data. However, there are several challenges to the accuracy of the estimate: there is collision and injury under-reporting and difficulty with estimating 12 distance travelled.

It should be stressed that this is a comparison of e-scooter injury rates in Bristol with the Great Britain urban cycling rate. It was a requirement of the evaluation that such a comparison be made, and it should be used with caution.

There was a total of 97 people injured in the 86 collisions involving all scooters (trial, illegal and not identifiability either) as recorded in the Police data for 2021. Nine e-scooter riders suffered serious injury and all remaining 88 injuries were slight. 13 collisions involved pedestrians as the casualty, 8 involved cyclists or a cycle pillion as the casualty, 6 involved injuries to the e-scooter pillion and 3 were drivers. 46 casualties are reported in 43 collisions which were reported to involve trial e-scooters.

Overall, 64% of collisions were not at junctions, which compares with the proportion of cycle collisions not at junctions in 2017 to 2019 in Great Britain of 26%. There may be some over-representation of collisions away from junctions as compared with cycling. While the STATS19 data is likely to under-report single vehicle (i.e. e-scooter only injuries), most injuries reported in the hospital study data result from falls, with these possibly being equivalent to single vehicle incidents.

Most injuries occur to the upper and lower limbs and the head and face. Only about one in ten riders say they feel unsafe riding an e-scooter and nearly seven in ten saying they feel safe. There are differences between demographic groups with older people, women and infrequent users feeling less safe

The minority of 409 respondents to the Experience Survey suggested they feel discriminated against by the presence of e-scooters.

Disabled people were more likely to feel discriminated against than non-disabled people (21% versus 13%). Those who identified as BAME or other non-white ethnic groups were also more likely to feel discriminated against (18% of BAME respondents and 36% of those of other ethnicities, compared to 11% for white people).

Younger people feel more discriminated against because e-scooter use needs the rider to hold a provisional driving licence.

The data available suggests that e-scooters are being used as part of a longer journey involving bus or rail for between 10% and 20% of journeys involving e-scooters.

In-depth interviews have revealed that e-scooters are a very useful form of transport for getting to and from a railway station where long distance travel is involved, as this avoids stress from unreliable bus connections and otherwise having to leave a bicycle at the station where it may not
be secure

E-scooter provision is not clearly linked to deprivation in Bristol but is more related to centrality with greater concentrations of parking zones in the centre of the city and a corridor connecting the centre to the northern suburbs of Bristol.

Suburban and peripheral areas have been relatively less well served, particularly in the north-west and south of the city and these include some of the most deprived areas of the city. Expansion of the operating zone to the north-west in March 2022 and to the south in December 2022 has resulted in a more equitable distribution of e-scooter parking zones across the city.

In Bath the operating zone covered a central part of the city before being expanded to cover most of the city in the summer of 2022.

With regard to users, access to e-scooters is perceived overall as easy (87% of responses), with some differences across ages and disabilities: younger users (aged 18-29) find access easier than older users, and non-disabled find it easier to access e-scooters than disabled users

The high proportion of e-scooter trips reported to be work-related (36%) demonstrates that the rental e-scooters are directly supporting economic activity, while the use of e-scooters for running errands (e.g. shopping) and visiting gym/sports venues (combined total of 19%) shows they are indirectly supporting businesses in the area.

Trips for social engagements and leisure (combined total of 39%) might also support businesses. Four in ten Bristol respondents (39%) to the trial operator’s surveys said that e-scooters enabled travel to places not previously possible and 31% of Bath respondents. In Bristol, this figure is highest in the 18-24 age group (53%) and decreases across the age groups. This provides an indication that e–scooters are widening the travel horizons of younger users and enabling them to access new opportunities.

There was no evidence from the video-based interactions study that there were any situations when the presence of e-scooter riders on the carriageway impacted on the flow of motor traffic.

This objective evidence is supported by the views of stakeholders who generally had no concerns about the operation of the network with the additional presence of e-scooters, although there was a lone more nuanced view was that the numbers may cause issues at certain times of the day in some locations.

It was suggested that e-scooters will have increased the overall capacity on the network for moving people, and this would be as a result of e-scooters take up less space within the highway.

As well as enhancing network efficiency, it was noted that e-scooters enhance individual trip efficiency because they are faster than walking for trips that switch from walking, and reach nearer to ultimate origins and destinations typically than other modes.

Nearly one-quarter of Experience Survey respondents (23%) thought that e-scooters contribute to health.

It is likely that many respondents interpreted this question as asking about physical health. It was recognised in interviews with e-scooter users that using an e-scooter is less exercise than walking and cycling, but some interviewees said that e-scooters encouraged them to go out when they might not have otherwise (not just for leisure but also for education, work or personal business reasons).

Almost half of e-scooter users (45%) thought that e-scooters contribute to their well-being. A number of in-depth interviewees highlighted that e-scooters facilitated them exploring the city and discovering new places. For some interviewees the e-scooters made it easier to visit family and friends in other parts of the city.

The estimation of net carbon dioxide emissions changes from mode switching to e-scooters, using lifecycle estimates per passenger kilometres for different modes, indicate that there are carbon emissions savings.

The estimates have wide variability and this is based on significant variability in figures for mode replacement from different surveys. The level of the savings is driven significantly by the proportion of walk trips that divert to e-scooter use. The estimates have adopted an e-scooter lifecycle carbon emissions factor of 65.2 gCO2eq per passenger kilometre.

The estimates of net saving for the year 2021 range from 6 to 238 tonnes net saving in Bristol for 4.15 million rides, and from 5 to 7 tonnes saving in Bath for 206,000 rides. These are based on mean lifecycle values. The variability results from a range in vehicle carbon emissions estimates and mode replacement estimates. There are potential additional variabilities beyond these, including possible differences in distances travelled by the replacement modes (e.g. replaced car trips possibly being longer than replaced walk trips) and differences in distances of the actual e-scooter journey as compared with a journey made by the replaced mode.

There was some recognition amongst some stakeholders that an e-scooter is potentially a good monitoring device for environmental measurements such as air quality, and infrastructure measurements such as road condition monitoring. It was recognised that analysis of this sort of data is a significant undertaking.

The issue with introducing methods for collecting additional data is mainly concerned with where the costs lie. Analysis of the data would typically be undertaken by a specialist sub-contractor, and this would logically be at a transport or highway authority’s expense.

There were 902 parking locations in Bristol and South Gloucestershire and 117 in Bath. Parking boundaries are defined by geofence, and in a few locations in South Gloucestershire also by a white painted boundary.

There are a few metal corrals at locations such as universities. The trial operator data suggests mostly good compliance (88%), with 11% instances categorised as “not ideal”, and 1% as illegal or unclassifiable.

An on-street survey in central Bristol found that 1.3% of e-scooters blocked the pathway used by pedestrians, and 15.8% were located on pedestrian pathways requiring pedestrians to swerve around them. Video observations show that e-scooters can interfere with pathways for walking sometimes for extended periods of time.

Responses from nine follow-along participants of diverse ages and disability statuses revealed that the issues created for pedestrians are barriers to access, a sense of risk, and a sense of loss of pedestrian space. One participant reported that e-scooters make their world a frightening place.

The footway, the carriageway and cycle tracks are all part of the highway. Cycle tracks provide the most appropriate infrastructure for e-scooters because e-scooter characteristics are most like cycles. The video-based interactions analysis has revealed a very high number of 39,369 occasions when near-misses involving e-scooters or cyclists occurred in 36 hours of video at eight sites. This number is almost as high as the total number of e-scooters and cyclists passing through the site and reflect multiple near-misses within the frame of the video.

The data suggest that e-scooters are less likely to have interactions with pedestrians than cyclists are. The number of near-misses between e-scooters is very high during carriageway riding, with a total of 37,305 near-misses being with motor vehicles (95%).

Surface roughness and potholes, and inadequate reinstatements of the highway pavement by utilities, mean there is a greater likelihood of falls from e-scooters.

From the point of view of e-scooter users, a better infrastructure for e-scooters and other micro-mobility transport was identified as urgent by interviewees if e-scooters are to thrive. From the trial operator’s Summer 2021 Survey, infrastructure (quality of roads, having enough cycle lanes) is regarded as important for safety while riding an e-scooter by most users (about four in five) with a significant minority dissatisfied with it (about one in four). This suggests it as a priority for intervention if usage is to increase.

Digital data can play a large part in planning, managing and operating e-scooters, including integration with other applications such as mobility-as-a-service information systems. Refined detail is possible, such as knowing the ultimate origin of travellers based on when they open the app to search for an e-scooter and good knowledge about routes taken. However, these data, it was suggested by stakeholders, could and should be more available to the authorities than they have been in the trial.

The legislative, regulatory, and licensing frameworks have been generally effective. Most stakeholders welcome the fact that riders are required to hold a provisional driving 15 licence and the responses framed the advantages, including riders being subject to the same penalties as drivers for being riding under the influence of alcohol.

No stakeholders called for private e-scooters to be made legal, with the suggestion from some that their sale be illegal. Any development in national legislation should not create additional time and cost penalties for highway authorities in relation to traffic regulation orders.

The advent of geofences has brought into focus the need for possible development in the law and regulation concerned with the ‘digital highway’. Highway authorities’ duties include controlling highway space and hence they need powers over those who may be defining virtual spaces. From the point of view of e-scooter users, interviewees felt shared e-scooters could make a valuable contribution to cleaner urban mobility but were concerned misuse of the scheme could lead to it being withdrawn and would welcome better enforcement of illegal practices.

As with the bus industry and for any privately determined and operated service, there is a dichotomy between profit maximisation and maximisation of social benefit.

From a societal point of view, there is an equity reason for ensuring a hire scheme is available to as many people of all ages and abilities as possible, and as many geographic areas and socio-economic groupings as possible.

Some stakeholders thought the solution to the problem is to make sure that a contract with an operator is explicit about the scope and coverage of services they need to provide, and there are mechanisms for checking delivery of that service. It was evident from responses provided by stakeholders that their time and effort in relation to assisting making the trial a success was a hidden subsidy to the operator. The suggestion was that as a minimum authorities need to cover additional costs and that future commercial models should reflect such transactions.

Stakeholders were satisfied with the overall operations and governance of the trial, while recognising that a lot has developed and been learned over the trial period. To improve such governance for future operations, it was suggested that inter-authority agreements be put in place.

It was also recognised that a more tightly specified contract with an operator is required, possibly with incentives and penalties for poor performance as discussed above. An inter-authority agreement and operator contract should be sufficiently flexible to allow for development and change as a result of technological capability.

A key to successful operational management is access to good monitoring data, which has evolved in nature over the trial. Stakeholders widely shared the view that data sharing for monitoring of operational performance could be further enhanced, perhaps particularly in relation to parking. One stakeholder thought that the operator needed to better understand the local travel environment and local authorities, and to this end would require a locally based project manager. From the point of view of e-scooter users, interviewees felt operational matters requiring attention were the management of parking, the existence of ‘dead zones’ where e-scooters cannot be used, and rules and messaging implying e-scooters are aimed at a young demographic.

A number of stakeholders noted that the cost of using e-scooters needs to be competitive with the bus if users are to continue using them.

The primary methodology for the trial operator to communicate with users has been through the mobile application. There were conflicting views from stakeholders about whether the messaging had reached saturation point, but there was concern that it is not possible to say how the messaging is being received by users.

Some highway authorities see value in expanding the use of in-app messaging to advise of wider highway related matters, and facilities to do this would need to be identified in a contract with an operator.