Content uploaded by Fraser Neil Mcleod
Author content
All content in this area was uploaded by Fraser Neil Mcleod
Content may be subject to copyright.
102
THE IMPACT OF FAILED HOME DELIVERIES ON CARBON
EMISSIONS: ARE COLLECTION / DELIVERY POINTS
ENVIRONMENTALLY-FRIENDLY ALTERNATIVES?
Julia Edwards, Alan McKinnon, Tom Cherrett, Fraser McLeod, Liying Song
Julia Edwards1, Alan McKinnon1, Tom Cherrett2, Fraser McLeod2, Liying Song3
1Logistics Research Centre, Heriot-Watt University,
2Transportation Research Group, School of Civil and Environmental Engineering, University of
Southampton
3School of Traffic and Transportation, Beijing Jiaotong University
Introduction
There has been phenomenal expansion of online shopping in recent years, with growth rates
exceeding 35% year-on-year (IMRG, 2008). Retail sales for this channel now account for £18.5bn in
the UK (Mintel, 2008). IMRG (2008) estimates that 820 million parcels were delivered to UK online
shoppers in 2008. This delivery experience is critical to the success of online shopping.
General consensus among consumers is that online shopping is good for the environment (IMRG,
2008; Royal Mail, 2007), yet consideration needs to be given to the frequency and treatment of failed
deliveries. Not only are unsuccessful deliveries costly and time-consuming for both retailers and
carriers and inconvenient for the consumer, they also have a detrimental effect on the environment
(Webster, 2007). Increasingly many people are not-at-home to receive deliveries during the working
day, when most home delivery companies operate (Prologis, 2008). As a result, it has been estimated
that half of households are unoccupied between the hours of 9.00 and 16.00 (Retail Logistics Task
Force, 2001). Therefore, it is not surprising that the majority of respondents, who have experienced a
failed delivery, report that it was because no one was at home to receive the package (IMRG, 2008). It
is mainly parcel carriers which must cope with this failed delivery problem.
The aim of this paper is first, to assess the additional carbon emissions generated by failed delivery
(as opposed to a successful first-time delivery) on a per drop basis and second, to consider the
potential environmental savings from the use of alternative collection/delivery locations over traditional
delivery methods.
Frequency of failed delivery
There are several variations to the traditional home delivery method but generally, goods are ordered
by the customer and delivered to a location of their choosing, using relatively narrow time windows
defined by the retailer. If the delivery fails because no-one is at the address to receive the item, the
carrier may make several re-delivery attempts as part of the round. If these also fail, the recipient is left
a notification card detailing a number of options, typically that:
l The item has been left with a neighbour;
l The item has been left somewhere outside the premises (unsecured delivery);
l The item has been returned to the carrier’s depot and further instructions are required from
the intended recipient. (Under these circumstances, the recipient can request a further
redelivery attempt which could be chargeable, or could visit the carriers’ depot personally to
collect the item).
Actual first-time delivery failure rates1 among carriers vary considerably; Beveridge (2007) indicated a
range between 2-30%, depending on the carriers’ policies for dealing with ‘no-one-at-home’. Some
parcel delivery companies achieve very high first-time delivery rates (McKinnon & Tallam, 2003), as
they are prepared to leave deliveries in alternative locations, (e.g. with neighbours or in ‘unsecure’
areas around the premises). According to IMRG (2008), 84% of online shoppers report that they
would be happy for a neighbour to receive their delivery on their behalf. Other carriers require proof-
of-delivery, and consequently have a much higher delivery failure rate when no-one is available at the
delivery address. As a result of these different delivery arrangements, estimates of first-time delivery
1 IMRG (2008, p.25) define a first-time delivery failure “as a delivery for which a signature cannot be
obtained, either from the customer or a designated customer representative, and this results in the
customer's address being carded and the item returned to the delivery depot for either redelivery or
customer collection”.
14th Annual Logistics Research Network Conference, 9th – 11th September 2009, Cardiff
103
failure rates vary widely from 6 out of every 10 small-package deliveries (Retail Logistics Task Force,
2001) to a more conservative one in nine (IMRG, 2008).
Quantifying CO2 emissions for failed delivery
Methodology
This research uses the carbon audit model outlined by Edwards et al. (2009) to calculate the CO2
emissions from failed deliveries. Emissions factor data from Defra (2008), NAEI, (2008), RHA (2008)
and FTA (2008) have been used to derive an average emissions factor for van-based home deliveries.
The analysis presented here is based on a typical van2 home delivery round of 120 drops and 50-
miles distance. Having already noted the wide variations in failed deliveries, three ‘first-time’ failed
delivery ratios were considered:
• 10%, similar to the 12% assumed by Weltevreden & Rotem-Mindali (2008), and recorded by
IMRG (2008);
• 30%, in line with findings by McLeod & Cherrett (2006), Song et al. (2009) and Belet et al.
(2009);
• 50% failure rate (worst case scenario) of a magnitude noted by Retail Logistics Task Force
(2001).
Low first-time failure rates of approximately 2%, achieved by parcel companies practicing unsecured
deliveries, and couriers who have made prior arrangements with regular customers in the event of
failures are not considered here. In these delivery cases, there will be only a very marginal increase in
carbon emissions.
For each of the three different first-time failure rates, emissions are also calculated for a second
attempted drop (the customer is notified of the intended re-delivery). Normally, the two delivery
attempts are on consecutive days, which results in a high percentage of second, re-delivery attempts
also failing, assumed to be 50% in this research (in accordance with McLeod & Cherrett, 2009). Often,
the customer will then travel to the local depot to collect the missed package rather than arrange for
another delivery attempt which could be at additional cost. The CO2 implications of the above
scenarios have been analysed.
Only 5% of people would wish to nominate a local depot as an alternative delivery address (Retail
Logistics Task Force. 2001) implying that the location of a typical depot is inconvenient for the majority
of online customers. Therefore, we can assume that most customers collect undelivered packages
either by car or by public transport with previous surveys suggesting that the vast majority of these
trips (87%) are by car (McLeod & Cherrett, 2006). Defra’s emission factors (CO2 per km travelled) for
average car and bus journeys have been used to calculate the emissions generated by these
individual trips (Defra, 2007).
Three different collection depot scenarios were modelled (being 15km, 25km and 40km away from the
average home) are examined, in accordance with Clements (2005); McLeod & Cherrett (2009) and
Song et al. (2009) respectively. Further, as customers wherever possible will wish to minimise both the
time and expense involved in retrieving a missed delivery from a depot, we assume that they will try to
combine collection with other activities (a non-dedicated trip). Few previous studies have examined
consumer travel behaviour in relation to missed deliveries. Turner (2007) found that out of a sample
of 167 visitors collecting failed home deliveries at a Royal Mail sorting office, 62% drove and 60%
claimed to be making dedicated trips (no trip chaining on either leg). 84% said that this was their usual
response to a failed home delivery (personal collection). Parcel carrier depots are often more distant
and not as accessible as Royal Mail sorting offices. As a result, we assume that only 50% of the
overall return trip from customer’s home to the depot is allocated to the collection of the missed
package, and that most people will travel by car.
Results
Edwards et al. (2009) modelled carbon emissions for a standard home delivery round. Assuming
average delivery rates of 120 drops over a typical 50km round, it was calculated that a successful first-
time drop emitted 181gCO2. Using this emissions value per drop as a base, re-calculated emissions
generated for the different first-time failed delivery rates are shown in Table 1. The additional gCO2
2 A van denotes a light goods vehicle up to 3.5-tonnes maximum permissible gross vehicle weight of
van-type construction on a car chassis that operates on diesel fuel.
104
attributable to delivery failures arise from subsequent delivery attempts and the extra vehicle mileage
they generate. We assume that re-delivery attempts as part of the same round are subsumed within
the initial round length.
100%
successful
first-time
delivery
10%
failure
rate
30%
failure
rate
50%
failure
rate
gCO2 per
drop 181g 199g 235g 271g
Table 1: First-time van-based delivery: Emissions (gCO2) per drop
Emissions of CO2 per average drop increase from 181g for 100% first-time delivery to the worst-case
scenario of 271g when one in two deliveries are assumed to fail. Moreover, when a parcel carrier
finds no one at home at the first attempt, most delivery companies repeat the delivery 24-hours later.
Customers normally out during the working day, therefore, have very little advance notice to arrange
receipt of the package; consequently, the second attempt often fails, compounding the effects of the
initial failed delivery (Table 2). It has been assumed that regardless of initial delivery failure rates, half
of all second attempts fail.
1st delivery attempt failure rate
(plus 50% 2nd delivery failure)
10% failure
rate 30% failure
rate 50% failure
rate
gCO2 per
drop 208g 262g 316g
% increase
over base
case 15% 45% 75%
Table 2: Re-delivery factoring in a 50% failure rate: Emissions (gCO2) per drop
When a carrier experiences a 50% delivery failure rate for both first and second delivery attempts,
each drop in a typical round would be allocated 316gCO2 or approximately three-quarters more CO2
per drop than a successful first-time delivery.
Normally after two failed delivery attempts the parcel carrier returns the package to a local depot for
the individual to collect. While 15% of online shoppers claim not to visit a parcel depot to collect a
missed delivery (IMRG, 2008), Department for Transport (2009) estimate that around 3% of all home
delivery recipients make a trip to retrieve an item left at a post-office, depot or outlet. In comparison, in
a survey of 379 households across West Sussex, Song (2008) found that 21% of respondents claimed
to have travelled to a carrier’s depot for the purpose of collecting failed home deliveries between 3 and
11 times per year. When these individual trips are undertaken by private cars or public transport
(buses), the carbon intensity of online retailing rises steeply.
Distance to local depot
gCO2 per trip 15km 25km 40km
Car 3,113g 5,188g 8,300g
Bus 1,340g 2,234g 3,574g
Table 3: Emissions (gCO2) per consumer trip to a local depot to collect a missed delivery3
Table 3 shows the CO2 emissions produced by a customer travelling to a local depot of varying
distance from the point of origin to collect packages. Generally, local depots, often on the outskirts of
urban centres for ease of access to the road network, serve much wider areas than their immediate
3 It was assumed that a customer used a standard car of unknown fuel
14th Annual Logistics Research Network Conference, 9th – 11th September 2009, Cardiff
105
conurbation. Consequently, the greater the distance a customer has to travel to the local depot, the
more CO2 will be generated by that trip. Clearly, using a car to make a 40km journey to bring back a
missed delivery will produce 8,300gCO2 (or the equivalent of 26 re-delivery attempts by delivery van,
when half of first and second-time deliveries fail) (Table 3). When the bus is the chosen mode of
travel to a depot, the consumer (assuming the bus has average patronage of 9.2 passengers) will be
responsible for 3,574g (or 11 times the amount of CO2 generated by the two missed parcel deliveries)
for the 40km journey. Even the shorter distance of 15km, representing for example an individual’s city
centre-to-suburb trip to collect a package, will generate 3,113gCO2 by car and 1,340gCO2 when the
journey is undertaken by bus. The individual trip to the local depot accounts for the vast majority of
the CO2 emissions associated with the failed delivery. Therefore, minimising the emissions associated
with personal consumer travel to the depot is key to mitigating the overall environmental impact of
failed deliveries.
Alternative collection/delivery points (CDPs)
Failed deliveries are clearly undesirable from a number of viewpoints:
l The carrier incurs additional costs in trying to make further attempts to deliver (related to the
associated logistics and call centre charges which have been estimated to be up to £38.50
for each delivery failure (IMRG, 2006)),
l The customer is inconvenienced and may have to travel to the carrier’s depot personally to
collect the item,
l There are wider environmental impacts due to the additional vehicle trips made by carriers
and consumers.
With such a large proportion of failed deliveries, different delivery solutions have emerged, allowing
carriers to drop consignments without the need to obtain the final customer’s signature. One solution
is to deliver packages to secure storage boxes, sometimes fitted as an integral part of the house, or
secured to an outer wall, or for communal use in the form of locker banks (Giraffe Marketing, 2004;
Bearbox, 2004; ByBox, 2004). Another solution is for the carrier to take failed deliveries to a local
attended ‘collection/delivery point’ (CDP), which could be a shop, a petrol station or a post office. The
customer is left a card, or could in principle be sent an email or a text message, to inform them where
to collect their package. This delivery method is now used by Royal Mail and Parcel Force in the UK
using specific post offices as CDPs whilst other examples use grocery stores, newsagents and petrol
stations. Kiala has 5000 CDPs across Europe with outlets in the UK, Belgium, Germany, Austria,
Spain, The Netherlands, Luxembourg and France while Pickpoint has a network of CDPs across
Germany. A similar system using convenience stores operates successfully in Japan (Chopra, &
Meindl, 2003).
A delivery policy which automatically took failed first-time home deliveries to the customer’s nearest
CDP, if such a network of attended or unattended points were available, could benefit all parties and
reduce the aggregate mileage associated with either redelivering or collecting failed consignments.
This would need the customer’s agreement in advance and for a preferred CDP to be identified at the
point of sale, to be used by the carrier in the event of the failed first-time delivery.
Quantifying CO2 implications from using alternative collection/delivery points (CDPs)
The type of alternative CDP and typical distances from a consumer’s home are listed in Table 4, after
Song et al. (2009) who looked at the impacts on householder collection mileage from using different
densities of CDP across West Sussex. The increased CO2 per average drop (compared with a
successful standard delivery) reflects the additional distance travelled when a parcel carrier deviates
from a round to drop failed deliveries at the CDP before returning to base. It was assumed that no
deliveries were taken back to the depot, as they would either be delivered to the intended recipient
first-time or taken to their nearest CDP, for the customer to collect. Extending an average delivery
round by 6.5km to deposit failed deliveries at a Tesco Extra store would result in each drop being
allocated an extra 23gCO2 while the relatively short additional mileage to a post office (1.2km) would
add just 4gCO2 to each drop. The carrier benefits from having denser networks of CDPs as this
minimises the length of the diversions that vehicle make in the event of delivery failures.
106
The results suggested that compared with a standard failed delivery4, all the alternative CDPs offered
significant CO2 reductions due to the consumer having to travel less distance (on average) to collect
failed first-time deliveries, against travelling to the carriers local depot. Even greater CO2 savings
could be achieved by incorporating a collection into a normal travel routine. Belet et al. (2009)
reported that end consumers could reduce CO2 emissions per parcel by as much as 83% when items
were retrieved from CDPs as part of a trip chain.
Overall, the results suggested that the local post office was the most environmentally-favourable
location for a CDP, as a package left there would be responsible for just 13% of the CO2 generated by
a collection from a carrier’s local depot. This reflects the relatively high density of post offices and
short average distance to customers’ homes (1.2km) across the sample West Sussex householders.
A high proportion of customers can walk to post-offices to collect packages, and 40% of householders
in the West Sussex sample stated that walking would be their preferred mode to collect packages from
their local CDP (Song et al., 2009). Additional CO2 emissions associated with the walking trip are very
difficult to calculate and are normally excluded from carbon auditing calculations.
Rail commuters could also collect packages from railway stations as part of their daily commute, in
effect eliminating the additional customer trip to the CDP. Large supermarkets as potential CDPs,
although located the greatest distance from the average householder (6.5km), are regular destinations
for shoppers and the collection of a failed delivery could easily be incorporated into a more general
grocery trip (with minimal additional travel on the part of the consumer). Moreover, unlike post offices,
many of the larger supermarkets have extended (and sometimes 24-hour) opening hours.
Type of
collection/
delivery
point
Average
distance
(km) from
customer’s
home
CO2 per
average
drop
(including
additional
km to
C/DP)
CO2 for consumer trip
to Collection/Delivery
Points
Car Bus
% CO2 per
C/DP drop
compared
with
traditional
delivery
Tesco
Extra 6.5 204g 1,349g 581g 47%
Other
supermark
ets1
1.6 186g
332g 143g
16%
Average
supermark
et
4.0 195g
830g 357g
31%
Post office 1.2 185g 249g 107g 13%
Railway
station 3.2 192g 664g 664g 26%
1From ASDA, Morrison, Sainsbury, Waitrose
Table 4: Emissions (gCO2) per consumer trip to alternative CDPs
Conclusions
Various traditional failed delivery scenarios have been examined, based on different realistic failure
rates for home delivery. When only the freight component of failed delivery is assessed, the additional
CO2 from the second delivery attempt increased the emissions per drop by between 9 and 75%.
However, the vast majority of emissions associated with traditional failed delivery arise from the
personal trip to the local depot by an individual collecting the missed package. In the worst case
scenario, 8,300gCO2 is produced by a car journey of 40km to a parcel carrier’s depot, which is
equivalent to 26 re-delivery attempts by a delivery van. In addition to the environmental effects, these
collection trips are inconvenient and costly for the customer.
As a result, viable alternative delivery methods have the support of consumers: 78% of respondents
said that they would use a carbon-friendly delivery alternative over a traditional method when available
(IMRG, 2008). This research has highlighted the potential CO2 savings from the use of alternative
4 The calculation for a standard failed delivery assumes 30% of first-time deliveries fail, 50% of the re-
attempts fail and a customer, having been notified of the unsuccessful second delivery attempt, makes
a 15km trip by car to collect the package from the holding local depot.
14th Annual Logistics Research Network Conference, 9th – 11th September 2009, Cardiff
107
CDPs for failed deliveries. Supermarkets, railway stations and post offices each offer distinctive
benefits for consumers, and all lessen the CO2 emissions from failed home deliveries. Post offices,
owing to their extensive network, present the greatest savings. It should be remembered that smaller
CDPs (e.g. post offices and corner shops) have limited capacity for receiving packages and if such a
system was to be adopted more widely, particularly as a first-time delivery address, storage, handling
and security issues might arise.
While the research presented here considers only the potential CO2 savings for failed deliveries, CDPs
can also be well-positioned to receive products that consumers are returning. To date the take-up of
alternative CDPs has been slow, though with potentially a million online shoppers a year having to
collect orders that carriers were not able to delivery to the home (Department for Transport, 2009),
CDPs could make a major contribution to reducing the environmental impact of online shopping.
References
• Clements, A. (2005), Nobody home, Retail Week, pp.22-23, 4 March 2005.
• Bearbox (2004), Available online at http://www.bearbox.com/, accessed on 6 October 2005.
• Belet, P., W.Mensaert, l.Sys & S.Verstrepen (2009), Carbon footprint comparison of parcel
delivery via pick-up points versus home delivery: Case Kiala Belgium, Green Transportation &
Logistics Summit, Brussels, March 18th, 2009.
• Beveridge, I. (2007), Personal Communications, Meeting with Beveridge Associates, 19th July
2007.
• ByBox (2004), Available online at http://www.bybox.com/, accessed on 6 October 2005.
• Chopra, S. & P.Meindl (2003), Supply Chain Management, Upper Saddle River, NJ, Prentice
Hall.
• Defra (2008), Guidelines for Company Reporting on Greenhouse Gas Emissions, Annexes,
updated April 2008.
• Defra (2007), Passenger transport emissions factors, Methodology Paper, June 2007.
• Department for Transport (2009), Public experiences of home working and internet shopping,
Department for Transport: London.
• Edwards, J.B., A.C.McKinnon & S.L.Cullinane (2009), Carbon Auditing the ‘Last Mile’: Modelling
the Environmental Impacts of Conventional and Online Non-food Shopping, Green Logistics
Report, Heriot-Watt University.
• Freight Transport Association (2007), The Manager’s Guide to Distribution Costs, FTA.
• Giraffe Marketing (2004), Available online at http://www.hippo-box.co.uk/, accessed on 10
October 2005.
• IMRG (2006), Valuing Home Delivery; A Cost–Benefit Analysis, IMRG, London.
• IMRG (2008), Valuing Home Delivery; E-retail Industry Review, IMRG, London.
• McKinnon, A.C. & D.Tallam (2003), Unattended delivery to the home: assessment of the security
implications, International Journal of Retail and Distribution Management, Vol.31, No.1, pp.30-41.
• McLeod, F. & T.J.Cherrett (2006), Optimising vehicles undertaking waste collection, Final report
for the Department for Transport, September 2006, unpublished, Department for Transport,
London.
• McLeod, F. & T.J.Cherrett (2009), Quantifying the environmental benefits of collection/delivery
points, Journal of Operational Research Society (forthcoming).
• National Atmospheric Emissions Inventory (2008), Vehicle Emissions, Version 8, NAEI.
• Prologis (2008), Internet Retailing: Opportunities and Challenges for the UK’s Distribution
Property Markets, Prologis Supply Chain Research Reports, Winter 2008.
• Retail Logistics Task Force (2001), @ Your Home: New Markets for Customer Service and
Delivery, Foresight Programme, Department of Trade and Industry, London, (available online at
http://www.foresight.gov.uk. accessed 09/02/2005).
• Road Haulage Association (2008), RHA Cost Tables 2008, RHA.
• Royal Mail (2007), Home Shopper Tracker 2007, RAPID Marketing Services.
• Song, L. (2008), Transport and Environmental Impacts of Current Home Delivery Services and
the Benefits of Alternative Measures, PhD Thesis, Faculty of Engineering, Science and
Mathematics, School of Civil Engineering and the Environment, University of Southampton.
• Song, L., T.Cherrett, F.McLeod & W.Guan (2009), Addressing the last mile problem – the
transport impacts of collection / delivery points, paper given to the 88th Annual Meeting of the
Transportation Research Board, Washington.
108
• Turner, M.W. (2007), ‘Sorry, we tried to deliver but you weren’t in’ – Quantifying the impacts of
failed first time deliveries’, MSc Transportation Planning and Engineering Dissertation, School of
Civil Engineering and the Environment, University of Southampton.
• Webster, B. (2007), Boom in Internet Shopping may be adding to Carbon Dioxide Emissions, The
Times, 7th July 2007.
• Weltevreden, J.W. & T.O.Rotem-Mindali (2008), Mobility effects of b2c and c2c e-commerce: A
literature review and assessment, Working First Version, Netherlands Institute of Spatial
Research.
.
