Gordon Stokes Transport  

Pass-by and diverted traffic, and secondary trips

Transport Home Photography Home Transport Photos

Last updated - Mar 2015

Pages in this series

1 - Transport Assessment Introduction

2 - Trip generation

3 - Routes taken by traffic

4 - Pass by, diverted and other trips

5 - Other aspects

6 - Transport in Environmental Statements

Pass-by, diverted and linked trips - Summary

Not all the traffic that is predicted to use a proposed development will be 'new' to the locality, as some would be passing by anyway, or making a short diversion. Also, predictions of traffic to a mixed development could mean double counting as some people might use two land uses in the development on one trip. But it's a very inexact science and this page attempts to give guidance as to what might be realistic.

Terminology - different types of trip

The traffic prediction aim of a Transport Assessment (TA) is to assess how much extra traffic would appear on the local road network as a result of it being built. Most TAs correctly argue that not all traffic using the proposed development would be new to the road network.

  • Pass by traffic refers to people who would visit a proposed development but would have been passing right by it on their way somewhere else. For instance, you drop into a shop to pick up a newspaper and sandwich on the way to work
  • Diverted traffic refers to traffic which makes a very short diversion to visit the proposed site. The area assessed can vary according to the size of the development but is usually that which uses roads within about 400m.
  • "Linked trips" are where people do more than one activity at a destination. For TAs, they are generally only relevant to developments that are either right next door to other services, such as in a town centre, or where the development has mixed land uses, such as restaurants being built on the same site as a large shop. The reason they are important is that the Trics database estimates trips for each land use, but if someone used a shop and a cinema without moving their car, then the TA might estimate more trips than actually occurred.

The diagram below shows different trip types, using phraseology that seems to be most commonly accepted (though the numbering or diverted trips is mine).

  • A primary trip can also be called a simple trip, or a 'single purpose trip'. This is a journey from home to somewhere, and back home again (or it might be from work to somewhere and back to work again)
  • Pass-by and diverted trips are made on trips where a visit to more than location is made, sometimes known as a 'multi purpose trip' or a 'trip chain'
  • A 'linked trip' is one where someone uses two or more facilities at the same time in the same location. It can also be called a multi-purpose trip
  • The term 'secondary' trip can be used in a number of ways. The only thing that is certain is that it's NOT 'primary'! It can be used for a trip chain, or for a linked trip. Some TAs will call linked trips "secondary linked trips"
  • The designation of "Diverted (1), (2) and (3)" are my inventions and you won't find them referred to elsewhere, but I think it's useful to explain how much of a diversion is made and whether the journey will add traffic to the local network, because TAs have tended to assume they are all short diversions.

Traffic to some developments will be largely due to people who are passing by anyway – a petrol station is an obvious example. Most new developments do not generate entirely new journeys but change the day to day travel patterns that people are making already. But unless a development is

  • directly on a major road that most people in the area that it will serve will regularly drive past,
  • is of a small enough scale that passing traffic can sustain it economically, and
  • is of a nature that passing traffic is a relevant draw (such as a petrol station, or convenience store),

then it is inevitable that it will draw traffic in from a wider area, if it to succeed. The question is - "how much?" The number of trips that are calculated as not adding traffic to local roads can have a very large impact on calculations of how much extra congestion will ensue. Hence the scope for argument and disagreement between TA writers and those who doubt the accuracy of a TA is very large too.

Why are these trip types important?

When a development is likely to attract a lot of traffic to it, the effect on the local traffic network will be less if many of the trips are pass-by or diverted. The diagram below shows how different trip types are counted as adding traffic to a local network, or not, as the case may be.

So in terms of additional traffic associated with a proposed development:

  • A primary trip adds traffic throughout the zone being studied
  • A pass-by trip does not add any traffic to the local road network (because it is deemed to have been already there), though it does add traffic to a junction into the development
  • A short diverted trip adds traffic, but only up to the point where the diversion was made towards the site
  • Longer diverted trips (Types 2 and 3) should be counted as adding a new trip throughout the area being studied (but in my experience this is very seldom done which reduces the predicted traffic)

So if consultants can show that a high proportion of trips are likely to be pass-by or diverted, then the effect on junctions in the area will be less than if all trips were new to the local road network.

The problem is that estimation of how many trips will be of different types is very difficult to assess in many cases, and research has been patchy. Some surveys have been done asking people whether they were passing by a development when they visited it, or whether they made a diversion, but;

  • the level of detail to determine whether a diverted trip is Type 1, 2 or 3 (using the system above), is seldom looked at,
  • all local road networks are different, so it's difficult to gauge how big the study area should be,
  • all developments are different, so it's difficult to judge how many trips will be of each type, and
  • much of the research was conducted for reasons other than for TAs, and the methodology doesn't always show what it seems to.

Pass by, Diverted and Linked traffic literature

Some Transport Assessments quote research studies in ways that do not give an accurate picture of likely rates for a proposed development. Common mistakes are to quote high figures, but ignore key factors that relate to the propensity to make linked, pass-by or diverted trips. Important factors relate to:

  • The size of retail store. In general smaller stores are likely to be associated with more pass-by trips, not many diverted, and very few linked trips. Much of this difference is due to the time it takes to park, get to the shop, find stuff, pay for it and get back to the car. With increasing size the proportion of pass-by is likely to fall, while diverted trips are likely to increase, since larger stores are associated with people making a special trip or diverting to buy more stuff (in general).
  • Geography of the store in relation to major centres of employment (or other attraction) and whether the store is on a major radial route (or other high traffic volume route)
  • The store in relation to other existing retail or local attractions. Stores close to existing centres with many shops (over 50 or 100) are likely to have higher linked trip rates than those further away, or those close to centres with fewer other shops.
  • Most linked trips are to other retail outlets or retail related services such as banks, rather than, for example leisure facilities. Food shopping does not seem to be often related to activities such as theatre, cinema, or ice skating. And solitary activities tend not to be mixed with social ones.
  • Large foodstores which include services such as a pharmacy, banking services, clothes, and household goods would probably be associated with lower levels of linked trips since there is less likelihood of requiring other shops on a shopping trip.
  • Time of day and day of week is important. Research by MacIver and detailed analysis of the National Travel Survey shows that pass by and diverted trips are uncommon in the morning. Most such trips are made on the way home from other activities.
  • Although surveys may record quite high levels of multi-chain trips, this does not mean that the shopping element is necessarily on a direct journey from another activity. Analysis by MacIver and analysis of the National Travel Survey both point to around half of such trips being either major diversions, or to shops located ‘beyond; the other activity in relation to home.
  • Studies of linked trips have generally been carried out to assess how much a new retail outlet will boost the performance of a town or district centre, rather than whether it will reduce trip making. Hence they generally include shoppers using all modes, and there is no indication of whether shoppers are visiting other shops 'additional' to one store, or whether they are making more efficient use of travel. Most do not ask if a car was moved during the visit, or even if the other places visited were in the same centre. A high level of linked trip making may help a centre while also increasing traffic levels -retail studies are done to show synergy rather than fewer trips!

Key points from oft quoted key literature

Below is key literature that is often quoted, looking at the characteristics of the developments and areas in which they are based. The table below summarises key literature, while paragraphs below provide more detail for many of them. The comments tend to critical, not of the research itself, but how the figures are used in TAs.

Document

Key finding quoted

Comments

Ghezawi,RS, Wegmann FJ and Chatterjee A (1998) Convenience Store Trip Generation, Institution of Transport Engineers Journal, USA

Average pass by rate of 72%

Based on American convenience stores with fuel station attached with average floorspace of 200m2. Very few UK developments would have anywhere near this pass-by rate, but it is very often quoted.

Mouchel (2009) Proposed Tesco Store and Shopping Centre, West Bromwich, Working Paper 3 – Linked Trips

40% robust estimate for PM peak

This was part of a TA, arguing a level for a particular development, and may or may not be a robust estimate for that one, but that does not mean it will be robust for any other. Like many reports it is loosely based on findings from other reports.

Trics and Safeway report 95/3 Traffic and parking at food retailing

Usually one or two at the higher end of the spectrum for linked trips

The study was of several stores, each of which had different relationships to their local centre, and very different linked trip rates

Wrigley, Prof Neil, Dr. Dionysia Lambiri and Katherine Cudworth (2010) Revisiting the Impact of Large Foodstores on Market Towns and District Centres, University of Southampton (Available as Full Report, Summary Slides, December 2010, and Executive Summary)

Various rates of linked trips reported, usually from the summary report

The summary reports only touch the surface of the complexity of the findings, which are very different in different case studies. More is written about this and other work by Neil Wrigley in the text after this table. In particular they argue that linked trips boost town centre performance, which implies linked trips are a result of synergy, and not necessarily reducing overal trip making.

Wrigley N, Lambiri, D and Cudworth K (2009) Linked trips and town centre viability. Town and County Planning 2009, pp 433-438

Not often quoted

Quite in depth detail on Shepton Mallett case study. Shows figures quoted are for all modes, with walk linked trip rate much higher. Walk rate was 47% but car rate was only 27%

Wrigley N, Cudworth K and Lambiri, D (2010) Further evidence on linked trips and foodstore development. Town and County Planning 2010, pp187-193

Not often quoted

This gives more detail on the Somerset market towns case studies than is in the summary report

MacIver A (1999a) Transportation Impact Assessment and trip-chaining behaviour, Traffic Engineering and Control March 1999 pp129-134

Seldom quoted

This offers a re-analysis of data from a Sainsburys study of several supermarkets in Swindon and shows that around half of 'pass-by and diverted' trips were actually an extension of another trip, so should be counted as a primary trip

MacIver A (1999b) Transportation Impact Analysis: Forecasting Travel Demand, TEC May 1999, pp 262-266

Not often quoted or quoted disparigingly as not being based on evidence

This suggests likely pass-by rates for supermarkets in different locations, based on work in the preceding paper. TAs sometimes state that the figures are not based on evidence, though this is no less true than many other reports quoted

TRICS Consortium (1995) TRICS Research Report 95/2 “Pass-By and Diverted Trips: A Resume”

Quoted selectively as a source of figures, but also as now being out of date

This was the best general review of literature and evidence on pass-by and diverted rates, and being produced by Trics, was the nearest to 'guidance' that existed, until the next entry below ...

TRICS Consortium (2014) TRICS Research Report 14/1 “Pass-By & Diverted Trips”

Not yet worked its way through to many TAs

The report looks at literature in a somewhat uncritical manner, includes some new research, but concludes that each case has to be looked at separately with some guidance as to what factors might affect rates

Guy (2007) Planning for retail development – a critical view of the British experience. Routledge, Oxford

Only generally quoted as underplaying extent of linked trip making

Tables provide summaries of other research with levels of linkage of trips for car users and non car users, and the types of trip for which linking is most commonly made (see below)

The next paragraphs provide some more detail on literature referred to above.

Ghezawi,RS, Wegmann FJ and Chatterjee A (1998) Convenience Store Trip Generation, Institution of Transport Engineers Journal, USA
This report is often quoted in TAs for just about any retail development as showing as average pass-by rate of 72% (with the implication that the consultants are being very 'robust' by using a lower figure). This study was of convenience stores with an average floorspace of 200m2, in the USA, all with petrol pumps as part of the store. It is doubtful that this is applicable to any UK developments, but since pass-by rates are likely to be much higher for smaller stores, and the fact that all of these sold petrol, to quote them for larger supermarkets in the UK would be misleading. It's usually quoted to show that the chosen level for a TA is 'generous' in not using such a high figure. The next paper (below) seems to back up these findings.
Available as a pdf via a web search

Johnson, Kevin L; Hammond, Matthew I (2001) Trip-generation characteristics for convenience stores, Institution of Transport Engineers Journal, USA
This study found Trip Generation rates significantly different than ITE's trip generation rates, sixth edition rates for the convenience store with gasoline pumps category during the peak hours, plus signifantly higher pass-by trip percentages. As a result,the authors suggest consideration of a new land use code.
Abstract available on web

Mouchel (2009) Proposed Tesco Store and Shopping Centre, West Bromwich, Working Paper 3 – Linked Trips
This paper is often quoted as showing 40% linked trips being considered a 'robust' estimate during a weekday PM period - it is often implied that 'robust' means it is also robust for the proposed development. It is based on three appendices - Tesco research, a Neil Wrigley paper and a paper written by Oxford University for the National Retail Planning Forum (NRPF). The Tesco research quoted as a basis for this figure does not give store sizes, but from descriptions it seems likely that they were probably less than 5000m2 in size. The services and facilities used included public toilets which accounted for between 4% and 25% of trips involving ‘other facilities and services used’. The paper also draws on research on ‘linked trips’ by Wrigley et al (see notes below). The NRPF report takes extracts from a more general paper which concludes that the research on linked trips is does not generally have its methodology well described so it is difficult to draw robust conclusions.
Found on the web with difficulty

Trics and Safeway report 95/3 Traffic and parking at food retailing
This report is based on one research study of Safeway stores opened in the early 1990s in 9 locations.
They are all described as either ‘town centre’ or ‘out of centre’ with the out of centre stores ranging from being in established neighbourhood centres (e.g.Harwood, close to Bolton in Lancashire), to ones built on new housing estates (e.g Cheltenham). The level of vehicle use and linked trip making varies greatly, with linked trips by users of all modes varying from 8% to 37% with an average of 24% on Friday peaks, from 12% to 50% with an average of 21% on Saturday peaks. The lowest figures were found at Reading and Malvern, and the highest in both cases was found at Harwood in Lancashire. The text reports that “In many of the out of centre stores this is a difficult figure to calculate as the definition of adjacent facilities is difficult to define. For instance, in Harwood, a bank has been built as an adjacent retail unit and this is heavily frequented by customers, whereas in Tewkesbury a range of cash dispensers is located within the store”. The smallest stores (about 3,700m2 gross were Harwood and Tewkesbury).
Not freely available - only via Trics consortium

Wrigley, Prof Neil, Dr. Dionysia Lambiri and Katherine Cudworth (2010) Revisiting the Impact of Large Foodstores on Market Towns and District Centres, University of Southampton (Available as Full Report, Summary Slides, December 2010, and Executive Summary)
Wrigley N, Lambiri, D and Cudworth K (2009) Linked trips and town centre viability. Town and County Planning 2009, pp 433-438 (further detail on Shepton Mallett case study)
Wrigley N, Cudworth K and Lambiri, D (2010) Further evidence on linked trips and foodstore development. Town and County Planning 2010, pp187-193
These reports are all based on a large research study carried out between 2006 and 2009, funded by Tesco with the aim of assessing the retail impact of new supermarkets in market towns in Somerset, and in District Centres in Greater Manchester.
The key point in relation to linked trips is that they are concerned with general shopping behaviour by people visiting the stores rather than vehicle trips. They are concerned with the synergy of providing more shopping and other services close together, with the question being "do shops in close proximity increase patronage?". Thus, rather than a survey of individual trips the researchers were interested in whether people “always”, “frequently” or “occasionally” combine trips to the new supermarkets with trips to other facilities close by. The very high figures quoted in some Transport Assessments therefore have to be treated with caution.

Wrigley et al found high levels of people who use the new supermarkets also sometimes using other services in the town on the same trip. But it should be pointed out that:-

  • While 80% of users of Somerset market town supermarket users did at least occasionally link trips, only 51% of Lancashire district centre ones did so.
  • The only centre where all subdivisions of linking frequency were recorded was Ilminster where 26% always combined, 35% frequently did so, and 21% occasionally.
  • The text reports that linking was higher for non car users than car users. The only centre for which this data is presented in Shepton Mallett where an average of 32% always or frequently combined. In this case 47% of people walking to the store linked trips, while only 27% of car users did. (TAs are concerned with car journeys in this respect)
  • The highest figures for market towns were for Crewkerne where it was not possible to interview in the supermarket car park so it is recorded that those arriving by car for a ‘primary’ trip would not have been interviewed (thus increasing the apparent percentage linking trips).
  • Conclusions are that factors leading to high rates of linked trip making include the relative attractiveness of the existing centre. In these cases all centres had at least 70 other retail units, with most having between 120 and 138. Using these figures for supermarkets in centres with smaller numbers of shops would be misleading.
  • Geographical isolation from other retail outlets would also seem to be a factor. The Somerset market towns tend to be the only retail opportunities of such a size for many miles, whereas the district centres with much lower linked trip rates are all much closer to other retail centres, including what can be called town or city centres with even larger retail offers.
  • In TAs linked trips are assumed to mean that fewer trips are made than the sum of what would be expected from individual land uses. Wrigley's research points to town centres benefiting by synergy of retail outlets together, with the implication that a person attracted to one land use will visit others while they are there, rather than necessarily link their trips in order to reduce trips.

Using a formula whereby “always or frequently” is assumed to provide a linked trip in 6 out of 10 trips and “occasionally” for 2 in 10 trips, the data suggests that for all trips the market town linked trip rate would be about 35%, and for car trips about 29%. For the District centres the figures would be 20% for all and 17% for car users.

None of the above casts doubt on Wrigley’s conclusions which relate to the effects that the development of town centre superstores have on the retail viability of the existing centres in which they are placed, but casts serious doubt on using figures from the reports (especially the summary reports without reference to the full report) for assuming linked trip behaviour for Transport Assessments.

MacIver A (1999a) Transportation Impact Assessment and trip-chaining behaviour, Traffic Engineering and Control March 1999 pp129-134
MacIver A (1999b) Transportation Impact Analysis: Forecasting Travel Demand, TEC May 1999, pp 262-266
These reports are based on earlier research commissioned by Sainsbury’s that were used in the 1995 Trics guidance note on pass-by and diverted trips. The first (March 1999) contains detailed analysis of trip chains from the travel diaries. It concludes that for a large proportion of trips involving another location than the foodstore, the foodstore part was not on the way to, or home from the other activity, but an ‘extra leg’ on top. So, to conclude that multi-trip chains involving food should be described as pass-by or diverted is faulty. He found that about 50% of the linked trips could be defined as pass-by or diverted for Transport Assessment purposes.

The second paper takes results from the first and applies them to likely proportions of primary, transferred and pass-by and diverted trips. He suggests that the proportion of pass-by and diverted trips is affected by local geography, and suggests:

  • For superstores located on major commuting routes into larger urban areas, the pass-by proportion may range from 25% to 35%, depending on the levels of traffic flow. More populous areas likely to generate higher levels of pass-by
  • On other less significant commuting routes , in out of town locations and in urban areas with smaller populations, one can assume 15-25%
  • In town centres and on non-primary routes about 10%
  • In locations with little propensity to generate pass-by trips the proportion can be as low as 5%

These figures are often criticised in TAs as not being based on evidence, though they are scaled by the findings of MacIver’s research from the Swindon Sainsbury’s study, and based on considered judgement.

TRICS Consortium (1995) TRICS Research Report 95/2 “Pass-By and Diverted Trips: A Resume”
Until very recently this was the most wide ranging review of information on the subject. Much of the paper is concerned with definitions, which varied widely between Transport Assessments, and still often do. The broad conclusion was that around 60% to 65% of traffic to supermarkets being built at the time (early 1990s when supermarkets up to around 5000m2 were typical) would be primary or transferred (and hence additions to the local network). The other 35% would be pass-by or diverted, depending on locational characteristics.
Not freely available - only via Trics consortium

TRICS Consortium (2014) TRICS Research Report 14/1 “Pass-By & Diverted Trips”
This report updates the 1995/2 report, now regarded as outdated, though the literature shows that there has not been a large number of surveys or analysis that add much to knowledge (besides Neil Wrigley's work). It discusses the possible effects of internet shopping trends, though is quite sensibly inconclusive about how growth in internet shopping would impact on frequency of physical shopping or propensity for pass-by trips. It also contains analysis of the Trics database that provides implications as to how pass-by and diverted rates may vary according to store size and location. It concludes that pass-by and diverted trip rates need to be set with regard to analysis of the local situation and in discussion with Local Authorities.
Not freely available - only via Trics consortium

Guy, Cliff (2007) Planning for retail development – a critical view of the British experience. Routledge, Oxford
This book, a wide ranging review of retail issues, contains summaries of other research on linked trip propensity in different situations. Table 7.5, p183 is based on research by Carmen Hass-Klau looking at linked trip rates. For car shoppers it shows linked trips rates varying for town centres (25-48%), for District centres (1-19%), and for out of town centres (5-14%). In each case the rates for non car shoppers the rates are higher (22-50%, 10-28%, and 15-28% respectively). While this research is suggested in some TAs as downplaying linked trip rates it actually accord well with Wrigley's findings when taken across all cases.

Table 7.6, p184, based on Bennison et al (2000) shows the facilities most likely to result in linked trips (at town centre, edge of town, district centre and out of centre stores) being banks (22%, 12%, 7% and 7% respectively). Non food stores show 21%, 9%, 12% and 8%, while other food stores show 15% at town centre and 6% at all others. "Other services" showed only 3, 2, 2 and 2% while "sports/ entertainment" showed 5, 4, 3 and3%. The Bennison work was based on the "Tesco, 2000" research (as far as I can ascertain).
Certain pages available on Google Books - otherwise available to buy

National Retail Planning Forum (2004) Linked trips and the viability of centres of retail activity. (Written by Oxford Institute of Retail Management)
This paper is mainly a literature review with most quantitative evidence taken from the "Somerfield Linked Shopping Trip Survey, 1996" which overall found that "46% of the Somerfield shoppers undertake linked shopping on the same trip". But no details are given of store sizes or the car/ non-car split, whether a 'trip' means parking in the same location, or maybe another stop on the same outing. It also quotes Tesco research showing linked trips on 40% of main food trips and 29% of top-up trips. Again no detail is given of survey methodology. A summary of the work in the NRPF journal (which I can't seem to locate right now) drew the main conclusion that it was difficult to draw conclusions about the amount of linked trip making because research did not show what methodology had been used.
Report is an appendix to Mouchel 2009

Comments arising from literature

Two things are noteworthy from the literature:-

  • The literature is often selectively quoted in Transport Assessments, with oft repeated quotes. To me they imply that the authors haven't looked at the relevant literature, but have copied other TAs that appear to strengthe their case. If I'm guilty of going the other direction to a certain extent in what's above then so be it - but I think I'm being fair, and at least I have read a lot of it!
  • Literature from the academic sector tends to play down the contribution of pass-by and linked trips relative to that originating from the development industry

Also some litarature points to findings which may give implications that are not as straightforward as they initially seem. The Wrigley analysis is taken by Transport Assessment writers to imply that developments close together will reduce trip making levels, but Wrigley argues that the juxtaposition of retail activity boosts the vitality of the area. Thus, while some might argue that linked trip making may reduce trips by, say, 40% of what each would generate, the synergy may actually lead to people being attracted to both facilities, but using the other as well. Thus, the total trips may only be slightly less than the sum of the two individual ones. Indeed some argue that development of one land use, eg a cinema, is important in attracting people to another, eg restaurants. ... So - therefore MORE trips.

Summary of literature findings

The only real conclusion that can be drawn is that there is not enough evidence to provide any accurate estimates of pass by, diverted and linked trip rates or percentages for developments. However I feel that it is fair to say that any rates over about 25-30% for pass by and diverted and similar levels for linked trip making should be treated with scepticism. Any TA providing figures of this level or higher should be accompanied by a strongly argued case. And referring to precedent of other cases where such levels were not questioned by Local Authorities does not imply that the figures are realistic. It is more likely that the Local Authority were understaffed and/ or bamboozled by the TA.

Analysis of data on pass-by, diverted and linked trips

A section will hopefully be placed here at some point describing special analysis carried out on the GB National Travel Survey, showing the percentage of different types of shopping trip that may be pass-by or diverted trips.

The next page is concerned with some aspects of TAs that may be of interest or use

Back to Transport Index

Gordon Stokes, 2015