Exploring the relationship between urban freight demand and the purchasing behaviour of a University

Clean urban freight, within the domain of urban mobility, is a key part of European Union (EU) policy to enable environmentally sound, socially inclusive and economically viable cities. In addition it is expected to contribute to a 60% reduction in greenhouse gas emissions (GHG). Over 60% of the EU population currently lives in urban areas, contributing circa 85% of the GDP [1]. The 2011 European Commission’s Transport White Paper set ambitious scenarios and a vision of future urban freight transport to “halve the use of ‘conventionally-fuelled’ cars in urban transport by 2030, phase them out in cities by 2050 [and] achieve essentially CO₂-free city logistics in major urban centres by 2030” [2]. Some of the scenarios include minimising the number of freight movements and the distance over which they are carried out; using low emission vehicles/trucks; and making maximum use of intelligent transport systems (ITS) to increase delivery efficiency.

However, these scenarios and most previous policy and research have focused on the transport chain operators and the initiatives of shippers and/or suppliers of goods. Only in recent years has a conceptual model focused on the ‘receiver’ end of the supply chain been discussed amongst European urban freight researchers [3‐5]. A few examples have emerged of the practical implementation of this approach, the most influential being the adoption by Transport for London (TfL) of Delivery and Servicing Plans (DSP) in the London Freight Plan. DSPs were used at their Palestra Offices in Southwark [6] and later on a macro scale, for the 2012 Olympics. A DSP is the management and organisation of multiple operations of last mile deliveries from a demand based perspective, optimised by efficient freight planning from the perspective of the receiving organisation [5]. It provides a framework to better manage freight movement, by adjusting the conventional working methods of urban freight stakeholders to meet both the inbound logistics needs of the receiver and wider sustainability objectives. It may be more effective than previous interventions at the transport operator link in the chain, since the delivery company is usually carrying out the express wishes of their customer – in most cases the supplier of goods or services. The supplier, in turn, is executing a supply to meet the express wishes of their client - usually the receiver. Whilst purchasing for large organisations does certainly have a documented and often regulated sustainability agenda [7, 8] and Newcastle University does indeed have sustainability targets in its procurement policy, the relationship between an institutional organisation and that organisation as a freight receiver bears further investigation.

Despite the potential for freight ‘receiver’ stakeholders to play a key role in managing urban freight demand, very little is known about which factors influence their freight demand. This is especially true of a large organisation with demand generated by multiple departments and groups. Understanding the drivers that influence freight demand generation, in a large public organisation such as a University, would help inform the decision-making process to improve urban freight management towards a sustainable freight system.

In this research, we aimed to examine the relationship between urban freight demand (in the form of purchase order (PO) demand made at organisational level) and the purchasing behaviour of staff, using a Newcastle University as a case study. The study was expected to bring new insights to understand the differences underlying the purchasing behaviour of staff. The University had at this time over 80 buildings, holding 144 schools/departments/institutes, any of which could raise purchase orders via 466 expert buyers and 1058 shoppers (also known as requisitioners).

Focussing on the ‘receiver’ role raises multiple research questions about purchasing, both formal and private, and inbound deliveries at Universities. For example:

This study used a mixed research methods approach, including interviews with the procurement manager and questionnaire surveys of all staff. The interview with procurement manager was needed to better understand the policy and practice of purchasing at the organisation level, and then used to inform the design of the questionnaire. The questionnaire was purposed to gain a better understanding of the relationship between freight activity and purchasing, both formal and private. It was distributed via the Human Resources (HR) staff mailing list and the staff website. The design of questionnaire included buying behaviour (e.g. frequency, time, type of product, value of goods), opinions on the University coherent campus strategy, environmental issues, and socio-demographic characteristics. A better understanding of the relationship across these variables was expected to lead the University to embrace policies to better manage inbound delivery flows to the campus, reducing the visual, safety, congestion, economic and operational impacts. It was not assumed that the University wished to prevent private purchasing at work, and the possible economic, social and environmental sustainability benefits of such practice were looked for, as equally as the disbenefits.

Recent studies addressing urban freight issues have begun to highlight the relative importance of the role of ‘receiver’, as opposed to ‘carriers’, in improving the sustainability of the logistics supply chain. Logistics is an induced activity and derives from purchases. Sustainable purchasing is the pursuit of sustainable development objectives through the purchasing and supply chain process, incorporating social, environmental and economic aspects [7, 8]. Therefore it would be rational to explore the linked activities, to see the extent to which sustainable inbound logistics can be enhanced by the procurement practice that induces it.

The strategic importance of the purchasing function in the business environment, notably in the private sector, has been acknowledged and recognised to contribute to economic advantages [9]. Management, company image, customers, carriers and the public policies are some of the identified key drivers of sustainable purchasing, within the private sector. In the public sector, sustainable purchasing practices also include explicit factors such as environment, diversity, human rights, philanthropy and safety - key drivers to this sector [10]. The identified key principles of public sector procurement are transparency, accountability, and achieving value for money for citizens and taxpayers.

A benchmarking exercise on improving sustainable purchasing practices within the private sector has demonstrated eleven key activities in engaging suppliers, as can be seen in Table 1:

The importance of training and education was highlighted in private sector business practice [11], however purchasing decisions about third party logistics (3PL) are still made mainly based on traditional criteria such as price, quality and timely delivery [12]. A study looking at the 3PL consideration on adopting green initiatives demonstrated that government support, management and the entrepreneurship are the key drivers, while a well-defined regulations framework and financial incentives were the barriers [13].

A number of recent studies, mainly Swedish based, used the process approach to investigate the anticipated change. For example, the role of environmental (management) standards (e.g. ISO 14001) towards the public sector’s sustainability agenda [14]; the role of stakeholders in shaping municipal energy and climate planning processes [15, 16]; and the role of public procurement in setting up a coordinated freight distribution of municipal goods [17, 18]. The potential use of invoice data as a proxy for delivery, or using traffic counts as indicators, is discussed in Zunder et al. [5] and Aditjandra et al [19]. In addition, it is notable that the adoption of green public procurement has long been promoted within the EU [20] and that the UK professional Chartered Institute of Purchasing and Supply includes sustainability using such standards as the SA8000® Standard in its education and training materials [21, 22].

Coordinated freight distribution as discussed above is a concept of city logistics that incorporates the use of an urban consolidation centre (UCC) in combination with other measures, including encouraging the procurement function (of municipalities) to engage in-house operation of last-mile distribution and to use smaller or clean energy vehicles. The impact on the procurement function is the potential change in contract cost and in the agreement between the public authorities and suppliers/service providers (3PL), as product and logistic costs can be seen differently when last mile logistics is done in-house.

It is logical that the importance of procurement in both private and public sectors has the drivers, factors and corporate importance to drive a sustainable approach to inbound logistics through the mechanism of purchasing. There is however very little known of the purchasing behaviour at individual level within a large organisation. This paper addresses just that question, to gain a better understanding of the mechanism of purchasing that influences delivery activity and to better explore the degree to which sustainable procurement policy is being translated to sustainable logistics.

Modelling latent variables at individual level using multivariate analysis (such as regression) in transport policy research has long been recognised as a robust way to gain insights into understanding relationships between multi-dimensional issues. This method employs statistical control which explicitly accounts for the influences of attitudinal behaviours, by measuring and including them in the behaviour equation and thereby moving them from unobserved to observed [23, 24]. Some example of these works are able to shed light on the mechanisms of behaviour change with regard to sustainability in transportation research [23, 25‐28].

The above approaches are in contrast to common transport demand modelling that would rely heavily on analysis of vehicle movement behaviour within designed transport network; where for freight, vehicle trip/cargo weight is typically modelled as dependent variable of urban freight demand [29]. This analysis is typically followed with transport operation/assignment modelling framework that address optimisation of vehicle routing problems [30]. Recent development in this type of modelling package includes creation of additional module that can model ‘pick-up’ and ‘delivery’ operation of an establishment to characterise end consumer goods movement generation [31]. The agent of analysis used on those papers is vehicle trip activity based as opposed to individual activity that relates to attitudinal behaviours that may effect changes in freight trip demand as this paper attempt to address.

Newcastle University lies at the heart of a city with a population of around 300,000 within a conurbation of around 900,000, which lies on the North side of the river Tyne in the North-East of England. At the time of this work, the University had over 144 schools/departments/institutes and over 80 buildings. The University main campus is integral to, but separated from, the main city, and there are other campuses and sites integral to the city. Steady and flat volume of purchasing (averaging around 300 purchase orders (PO) each day) was evidenced from over 2700 vendors (suppliers) throughout the academic year 2011–2012 [32]. A drop in demand at Christmas was also reported, as well as no weekend working. 70% of POs staff raised were for goods and 20% for services, the balance financial activities or large scale capital infrastructure.

Interviews were held with the procurement team, to better understand University purchasing policy and practice. Both were highly decentralised, with the purchasing team taking the Pareto rule that 80% of spend will be with 20% suppliers and that these, plus the legal regulations on the spending of public money, should be where their efforts are focussed. We cross checked with the University’s Head of Procurement the key sustainable purchasing practices drawn from the literature, to explore the extent to which this hybrid private-public institution correlates with a nominal sustainable procurement profile. Table 2 summarises that interview, examining the current purchasing practice at the University. Generally speaking, the University only partly matches the full profile of practice and policy, notably in the areas of improvement, collaboration and relationships. The use of a Code of Conduct, supplier monitoring with a compliance awards scheme, and the general endeavour towards integrating corporate social responsibility in management, were evident from the interview.

Despite the insight gained from the top-level purchasing activity data, there is still very little known about how this purchasing activity is generated. Some POs represent a single item of medical equipment that is delivered with ten others on the same vehicle; on the other hand ‘blanket’ orders that cover a month’s deliveries to multiple sites were also discovered. In other words, there is little clarity of the relationships between POs and deliveries with ‘one to one’ and ‘one to many’ relationships existing but proving unclear in the data.

Referring to Fig. 1, the urban freight activity variable is represented by purchasing order (PO) frequency; the reported response scaled from 1 ‘once a fortnight or longer’ to 4 ‘every day’. The purchasing behaviour variables measured include yes/no and multiple choices (elaborated as dummy variables in the model) on ways to raise a PO, type of goods ordered, expected delivery time, etc. Socio-demographic variables include gender, age and job role. Environmental attitude variables were measured using Likert scale statements related to built environment characteristics and purchasing activity. Since all of these variables were collected through the questionnaire survey of individuals, latent variables data are the basis for the multivariate analysis, as later discussed. Table 5 demonstrates descriptive statistics of the variables included in the final estimated model. Since the data analysis was carried using SPSS, standard default descriptive frequency analysis was produced. The mean values in Table 5 for the dummy variables are simply the percentage of the sample population. We can see that over 60% of the sample are female and ordering goods without any regular repeating patterns.

Pearson correlation analysis for all possible pair of coefficients were also estimated and showed some significant (one-tailed) correlation at p < .05 with weak to moderate level (small to medium effect) of relationships between variables as can be seen in Table 6.

The highest significant relationship reported in Table 6 is between frequency of PO and individual behaviour who ordered goods without any repeating pattern (r = −.425), this figure effectively says that as the number of individual who ordered goods without repeating pattern goes down, frequency of PO goes up. In the next section we assess the causation of this relationship.

While significant differences are noted, descriptive statistics do not explain why these differences might exist. The next step of the analysis is to examine whether differences in Purchasing Order (PO) frequency can be related to the factors which have been described above. A cross-sectional model of PO frequency was developed, to examine the relationship between purchasing activities, type of goods purchased, socio-demographic and attitude towards private purchasing.

A Multiple Linear Regression (Ordinary Least Square) model was constructed, using PO frequency as the dependent variable and a number of factors (column 1 Table 7) identified as explanatory variables. The model initially included all variables identified as important: job types, ways to raise PO, type of goods purchased (some were dropped due to high correlation), expected delivery times, location of PO, gender, age, and private purchasing related behaviour. Insignificant variables above the 10% level were dropped, except for some variables of interest (e.g. private purchasing behaviour) and the model re-estimated.

In multiple regression, multicollinearity is always a potential problem. As well as screening for high linear correlation between the independent variables, Table 7 reports both variance inflation factors (VIF). VIF provides a measure of the effect of the inter-correlation of the regressors on the variance of the least squares parameters estimates. It has an ideal value of 1 (and hence the ideal value of the tolerance factor is also 1). VIF values in excess of 10, or with tolerance (1/VIF) of less than .2, are a cause of concern [33‐36]. The reported VIF and tolerance figures in Table 7 suggest multicollinearity is not a concern here.

The specialist, technical and professional (e.g. technician) job type significantly raised the demand via a PO, with the highest standardised coefficient (β = .245) (Table 7). Similarly, (in terms of β magnitude), having a University purchasing card significantly raised the likelihood of demand via a PO (β = .224). In contrast, purchasing made with an irregular repeating pattern is negatively associated with PO frequency (β = −.235) and purchasing made once or twice a week is positively associated with increased PO frequency (β = .124).

The factor for ‘expected delivery type: as soon as possible’ showed significant increased PO based demand, with a frequency of β = .102. In contrast, the location where the majority of administration staff are based, showed a negative association with PO frequency demand of β = −.145. Following this up with face to face discussion, we discovered that the administration staff there adopted a weekly order consolidation process, showing that one of the central administration units of the University is not a key driver of PO based demand.

The private purchasing behaviour variables are not significantly associated with PO demand. However, its negative association can be noted; thus the more private purchasing behaviour adopted, the lower the level of PO demand created. This negative association deserves further qualitative follow-up.

This paper contributes to our understanding of a University as ‘receiver’ – and therefore an inbound logistics freight generator – and the role of both purchasing policy and practice, the role of the end user as the driver of a requisitioner and, to some extent, that of the employee as generator of private purchasing delivered to work. As a large organisation with a sustainability agenda, it had a corporate interest that is not necessarily driven by bottom line cost imperatives. This research contributes to the growing body of work in the receiver-led point of view that draws potential optimisation of city logistics initiatives [37] from the inbound logistics perspective, and also informs the wider discussion about the locus of power in supply chains [38]. The purchasing data and interview with the University Head of Procurement drew insights into how the University practices sustainable purchasing, though not to the same degree as the benchmark practice within the private sector.

The staff survey carried out was intended to explore the degree to which institutional policy and actual purchasing behaviour are in alignment. This first survey and the preliminary analysis raise more questions than answers, but we can reach a few conclusions, albeit they will require further investigation.

The empirical model using PO frequency as the urban freight demand variable demonstrated that job types, ways to raise PO, type of goods purchased, expected delivery times, and location of PO are significantly influencing the PO demand. This finding has demonstrated the potential of behavioural changes that can be facilitated through the sustainability of urban freight policy. For instance, the location where the majority of staff are coordinated to adopt a weekly order consolidation process create a non-significant PO demand, this shows that implementing a weekly consolidation process in purchasing is possible (as this is already happening with the University central administration unit) and suggests that it can reduce the amount of PO.

Staff who are administrative, technical, specialist, or operational, or with a University purchasing card, have significantly higher demand expressed though PO. This finding could have policy implications towards behavioural change, in that the frequency of PO could be reduced by decreasing the number of purchasing cards, or by reducing and centralising the number of staff able to initiate purchasing, with a view to consequently consolidating goods or services and improving the University’s purchasing sustainability.

The manner of goods delivery demonstrated in Table 8 would suggest logistics delivery in small parcel vans and light goods vehicles, and indeed previous surveys have shown 82% of all delivery vehicles to the campus to be light goods vans [5]. The main University campus site at this time attracted or generated more than 50% of city traffic In Newcastle, primarily in light good vans, which while less polluting, contribute more to network congestion [27].

From a qualitative analysis we can see that this ‘buying’ population feel that they should not buy private goods for workplace delivery, but that it would be very helpful and welcomed. We also found that this belief had no basis in University policy but appeared entirely cultural, it would be worth comparing this to the student purchasing behaviour reported at the University of Southampton, a different cultural group [39]. Those that have ordered goods for private use, did so because ‘there is no-one at home’ and because ‘travelling to collect failed delivery is inconvenient’. The respondents also were well informed and made positive suggestions about improving freight deliveries, which shows that ideas for intervention in this area are widespread and not restricted to experts [32]. We see this as confirmation that bottom-up stakeholder led interventions have a strong body of innovative ideas to call upon (for further discussion, please see [40‐42]). We cannot show that the nature of the respondents, who were primarily administrative staff, made them more compliant with the perceived view that private purchasing is not allowed. This, however, would merit further investigation among academic and non-clerical staff, not least because there is no formal prohibition on private purchasing at the University. Indeed one of the immediate impacts of this research was that it informed the University Executive’s decision to support private purchasing at work, and plans to introduce locker banks on campus for such private deliveries.

We asked a very broad question: is the changing behaviour of purchasing towards freight efficiency a feasible option for a sustainable institutional organisation? Here we can only move ahead slightly, in regard to that question. We would draw out that the qualitative surveys show willingness amongst the buying population to both suggest and embrace alternative ideas. On the other hand, the quantitative model demonstrated key determinants of PO demand. The most significant staff roles, such as ‘technical’ ‘specialist’, ‘administrative staff’, and ‘operational services with a University purchasing card’ could be trained to engage in a city logistics initiative, such as Delivery and Servicing Plan (please see [3, 43] for references), so that urban freight delivery can well be managed. We can see that a very small core of people raise most of the orders and therefore it should be possible to influence the majority of orders through them. We should be concerned that the high frequency core orders may generate the most efficient inbound logistics and it may be the low repetition, low efficiency orders that should be addressed - a much wider pool of people, for whom this is not a core activity. We can also show that the University has the basis for sustainable procurement policy and practice, but there is an opportunity for development, and that policy development could be aligned with a sustainable inbound logistics practice. Lastly, demonstrating to the University’s Executive that employee views on private purchasing directly influence the practice, has convinced the board that changing purchasing behaviour towards freight efficiency is a feasible option for a sustainable institutional organisation.

The key contribution of this paper is demonstration of the important role of logistics receiver can make in delivering sustainable city logistics. This is especially true for large organisation with multi-sited and multi-level management (central vs local) that require multi-type of logistics in a city-centred bound historic built environment University. This paper shed light on identifying the key determinants of freight demand at University that can be managed and act as catalyst for accommodating urban freight in city planning.

As a postscript we wish to note that in 2016, partially in response to this and associated research, the University centralised all purchasing into a new hub, with the intention of reaping the benefits of consolidated purchasing power. The policy on private purchasing delivered to the workplace has not been changed, indeed several senior managers have explored the addition of locker banks to the campus to facilitate it.