Case study 2 and 3

Introduction and context

This case investigates the issues of complexity and efficiency in reverse flows of multiple-trip packaging systems and recyclable materials from retail units back to points of distribution. Academic literature around reverse logistics (RL) is abundant in theoretical and conceptual modelling identifying efficiencies of scale and scope. This case aims to identify practicalities, feasibility and readiness for multiple retailers to incorporate or adapt these principles of best practice. We augment this case by reviewing the combination of just-in time (JIT) and a lean philosophy within RL systems. We address whether implementation of JIT is feasible within a multi-trip system and what the possible performance implications of such an implementation might be, especially for information availability, flow and use.

A case study methodology provides a comparative analysis of two UK multiple retailers. The systems deployed by these retailers are described in relation to gaps that emerge between what may be theoretically or conceptually possible, according to the literature, and what was actually observed. We reason why these gaps might occur in a real-world setting, synthesizing the primary data with the extant literature to suggest improvements that could realistically be made to the systems observed. We conclude by suggesting how real-time data may be better employed for RL systems in order to enjoy improved operational effectiveness. Finally, through further synthesis with the literature, this case offers a set of options for improved lean reverse logistics systems and the introduction of an inclusive tripartite ‘walk through’ model to create action that will address system sub-optimization and drive sustainability.

Case methodology

Selection of the retailers examined was driven by a need to comprehend ‘same but different’ supply scenarios, whereby a comparative analysis could determine to what extent advantage, if any, is gained in the management of reverse flows that result from higher levels of vertical integration being present in Retailer A. The adoption of case study research enabled the comprehension of practices present within the daily operations of both organizations A and B, as well as understanding the perspectives of internal stakeholders (Cassell and Symon, 2004; Stake, 2000; Yin, 2003). This case presents four operational levels of analysis, these units of analysis being the small, medium and large retail units and distribution/logistics sub-units of both organizations.

A purposive sampling technique was utilized (Easterby-Smith, Thorpe and Jackson, 2015; Saunders, Lewis and Thornhill, 2009), commencing with a ‘bottom-up’ retail sub-unit activity, back up the supply chain to the point of distribution. The retail units studied were indicative of common store footprints, ranging from small convenience store to large supermarket. Three managers from each retailer would offer sufficient depth of perspective, and these included both logistics and warehousing managers from the distribution/logistics sub-units as they were of equal and critical importance to the case. Table 10.1 shows the themes explored during interviews.

Comparison of the case study retailers

It is common practice for multiple retailers to utilize returning vehicles (the retail backhaul) for the purpose of collecting recyclables and offloading them within a dedicated recovery area within their regional distribution centres (RDCs). Notwithstanding this, it is apparent that system differences can emerge at this distribution node and at the retail unit nodes. Industry initiatives in these areas include: improving space utilization and fuel usage, optimizing road transport, including vehicle design, engine type and fuel type, as well as developing inter-modality and the ongoing building and retrofitting of warehouse facilities to improve sustainability. In the organizational analysis between Retailers A and B, the key similarities between their distribution and fulfilment strategies include continuous 24-hour cycles to minimize stocks and maximize network efficiencies for ‘in store’ fulfilment, while managing access restrictions borne of unitary authorities’ local transport plans.

A and B differ significantly in that there is a higher degree of vertical integration within A. The key characteristics of A and B are detailed in Tables 10.2A and 10.2B.

Case findings

Process dynamics

On arrival at the RDC, the packaging systems, RPCs, full roll cages and empty roll cages are collected at a central point for deconsolidation/first stage disposition, commonly referred to across the sector as ‘De-Kit’ areas. Here, packaging systems, RPCs and recyclables are sorted and recyclables are baled, while packaging systems/reusable containers typified by roll cages and crates are sorted; some types are also cleaned before being redistributed to support forward flow operations. Differing levels of sortation and sortation quality occur depending upon the size of the retail unit, with larger units baling recyclables and smaller units storing recyclables in roll cages.

Differences of opinion emerged between DC and retail interviewees. The retail level considered the quality of baling and sortation to be of low priority, having no relevance to performance evaluation metrics, whereas at DC level it is considered a key process, although this is not reflected in the forward-facing strategy of the DCs despite the necessity of retrieving reusable systems equipment and reducing driver hours, both critical to store replenishment activities.

Low priority being given to return flows from retail units relating to sortation, quality, quantity and integrity of materials returning to the DCs is twofold in impact. First, it is difficult to plan effective deconsolidation as varying levels of sortation are present but may also change on a daily and/or route basis. Further, this lack of RL management is also evident in the analysis of interactions between the DCs and recycling companies, more so with Retailer A, where the main focus is reducing moisture penetration in card bales and identification of contaminated plastic bales. This has led to opportunistic practices by recyclers, for example recyclers declining payment for, but continuing to process, contaminated bales (plastic bales with evidence of cardboard contained within them) on an individual bale-by-bale basis. Where short-term improvements in bale quality emerged, recyclers then started to decline payment for entire loads if just one contaminated bale was found. Such opportunistic behaviour overlooks vested ‘win–win’ sustainable relationships and would be quite unthinkable in forward flows.

Characteristics of materials acquired for upstream disposition

Multi-trip systems and packaging waste differ from unit to unit; waste and returns were influenced exclusively by replenishment activities. Common to both retailers is the apparent limited or lack of control at the retail level with regard to deciding the amount of packaging systems and recyclable materials to be returned, as this is dependent on available vehicle capacity.

Capacity

There are two main capacity issues common to both A and B. Storage capacity by design is intentionally limited at retail units in order to drive stock replenishment and inventory throughput. This naturally leads to limited storage space for returns, resulting in reverse flow materials commonly being stored outside and thus impacting paper and board quality. The other capacity issue common to both is related to excess capacity at the DC baler, which is not optimized as flow velocity and volumes are dictated by driver availability, dock availability and staff availability. Consequently, this also leads to inefficiencies, such as excessive ‘slack’ in the system at certain times of day, queues at other times, and a degree of empty running across the network. Holistically this creates an unbalanced network, whereby DC capacity is both greater than the total demand placed upon it, and significantly impacted by the peculiarities of forward- flow optimization.

Dynamics of ownership and capabilities

Prevailing process conditions at both A and B’s retail units and DCs are characterized by task/time orientation; this stifles dynamic task orientation in support of RL and can be clearly linked to the low priority of reverse flows at the retail level. At DC level, lack of flexibility to support dynamic responses to RL peaks can be clearly evidenced through regular usage of agency staff within ‘De-Kit’ areas.

Inhibiting factors and potential system improvements

The principal issue that emerges for enhanced RL efficiencies in a leaner system is related to cost control in relation to duplication of vehicle movements. Other issues recorded as potential barriers included:

· internal focus on the day-to-day activities;

· lack of support for existing projects;

· upstream RL not seen as a priority in servicing customer-facing flows;

· limited internal collaborations, lack of holistic insight and poor legacy management/swift ‘post-project death’ where RL initiatives have been trialled;

· reactions to reverse flows rather than generation of simple, extensible data to allow anticipation to influence demand planning and collaborations;

· non-sustainable and opportunistic behaviours by recyclers.

Table 10.3 summarizes inhibiting factors and system potential.

Case discussion and actions

Measuring success factors

As with forward flows, the basic requirements to have the right materials in the right places at the right time are compelling factors for network efficiency. Any RL system improvement allowing for efficacious development needs to commence with the retail unit as supplier. This issue can be considered as a compelling need for supplier development and supplier relationship management. The overriding need to ensure customer satisfaction means that there remains a primarily forward focus on movement, whereby RL remains as non-value-adding from the retail perspective. Nevertheless, this attitudinally driven behaviour and mindset at the retail unit end could be considered as ‘low-hanging fruit’ and an easily attained win within the paradigm of the theory of constraints (Mabin and Balderstone, 2003).

Imbalances between staff availability and dedicated tasks at DCs hinder system change, with teams needing to react rather than proactively manage both forward and reverse flows. This is a shorter-term attitudinal/operational constraint rather than a longer-term operational reality. To this end, this case introduces a conceptual road map, based upon the prioritization of issues emerging from the primary data analysis to drive collaboration between supply partners from retail, logistics and DC operations sub-units, presented in Table 10.4.

Networks

Access to retail units for planned retrieval beyond simultaneous drop-off and pick-up is currently restricted by both local authority time and access restrictions imposed on some of the retail units. This issue is only relevant to smaller retail units (in urban and residential areas) and makes assumptions around current fleet type/fleet availability. Nonetheless, it negatively affects both the retail unit as supplier in its ability to feed the system as well as other critical factors such as network design and supplier delivery timing, as local conditions create levels of unconditional inflexibility within an existing network and fleet. While there are differences in provision of logistics service ownership between A and B, there are prevailing characteristics that can act as facilitators in each of the current networks; these include current empty running and excess capacity at certain times of the day.

The networks and systems investigated in this case are internal, part of the same organizations, enabling collaboration between retail units and DCs driving system viability. This leaves DCs in an operating environment where they must anticipate retail unit RL demand and manage that demand by behaving with a degree of flexibility, albeit in a closed network system.

The role of supporting elements

Centralized communications prevalent within both systems work against direct articulation between retail units and DCs. However, this indicates that the essential IT infrastructure is present in both organizations but is not currently utilized for RL. In considering the possibility to ‘do more with less’, current information systems can support both RL development in the short term and JIT-RL practices in the longer term, providing efficiency benefits without an incremental escalation in current node and network complexity.

There are obvious lessons to be learnt from the boom in online retail and the challenges successfully overcome by these organizations and third-party logistics service providers. An example of this is given anecdotally by online fashion retailer ASOS, who report a return rate of 56 per cent of downstream being returned by customers, with all returns requiring central collation, 100 per cent inspection, repacking, relabelling and reassigning to a central stock control system. Our investigation revealed that each retail unit has the ability to ‘write on’ and ‘write off’ stock at store level, thus it would seem remiss not to simply identify empty packaging systems and recyclable wastes as types of stock with a zero value stock keeping unit (SKU) descriptor but a variable SKU component value. Each retail unit would thus create dynamic visibility for DCs and vertically integrated suppliers within systems, allowing for dynamic planning to supplement stable material flows back to DCs as well as providing valuable information, informing disposition strategy and execution further on in the RL supply chain.

Case consequences, creating actions and options

Separate JIT systems flowing forwards and backwards utilizing the same vehicle and driver pool would appear to contradict two of the seven classic muda wastes: unnecessary movement and waiting. Running two JIT systems can largely remove pick-up requirements in forward flows, significantly shortening delivery times and increasing dock availability while reducing turnaround/route fulfilment times and increasing route planning options in an integrative system, allowing for increases in RL efficiency pursuant to forward flows. Further, the use of real-time data on types and volumes of recyclables at retail units would in turn allow for load maximization and levelling of process capacity for RL flows into the DCs, ensuring more efficient employment of DC and driver staff. Medium- to longer-term benefits potentially extend to re-profiling vehicle fleets (size and type) and a reduction in DC operational running costs (ie agency staff), while increasing throughput (Table 10.5).

Conclusions

This case assesses the feasibility of increasing RL efficiencies to benefit positioning of packaging systems and recyclable materials. Our analysis of the RL systems prevailing in retailers A and B reveal commonality in ‘same but different’ network systems. While no definitive conclusions are stated for either A or B in relation to their current systems, it is beyond doubt that there are sufficient indicators of RL system sub-optimization to act as a barometer for change through RL system adaptation.

The case does not suggest that wholesale investment is required, since there is evidence of excess capacity in the networks. RL systems are borne of historical development of store replenishment, in which retailers’ reuse and recycling practices are present, formalized but informally managed. Both A’s and B’s systems lend themselves to increased RL efficiency through facilitating IT infrastructure elements already embedded across operations to generate and mine extensible RL-relevant data from forward flows. Increased accountability of materials flowing back up the supply chain can also address opportunistic behaviours; for example, it is possible for A and B to develop a more sustainable, integrated RL system where queue management begins at retail units, not at often congested DCs.

This case reveals a low priority of packaging system/recyclable materials/RPC ownership regardless of retail unit size or concerns over ‘dead hauling’ common across both retailers. Notwithstanding these issues, and critical to the case, there is a consistent throughput and a great degree of commonality in DC fulfilment practices, indicating the predictable volume and low combinatorial complexity typical of ‘lean’ flows and JIT. What is missing from both is locally generated and extensible exported/mined data which would reposition queue management, occurring earlier and more efficiently (Figure 10.1).

Under traditional lean paradigms, demand generates processing, which provokes supply to meet demand. This stands counter to the RL systems revealed here, in which capacity imbalances exist and outstrip supply, amassing multi-trip packaging systems leading to stockpiles in the wrong place at the wrong time.

Revealing RL sub-optimization, barriers to increasing RL efficiency and potential benefits of lean/JIT-type RL system adaptation, this case indicates that extensive and sweeping changes to current network structures are not always required. Alternative overarching systems and strategies can be developed to allow better decisions utilizing existing infrastructures.

Alternatively, an independent third-party system provides a potentially complementary strategy (refer to Table 10.5) if the ‘low-hanging fruit’ prevail, or when there is already a degree of third-party integration in logistics and DC activity. This would allow either retailer to continue to focus on forward flows, while inviting third-party logistics providers to build value into RL systems. Forward distribution’s transport and warehousing operations are highly visible parts of the multiple retailer supply chain, key to delivering system efficiencies for multiple retailers. Therefore, in this regard it is lackadaisical not to investigate areas of RL optimization to extract further value from these networks.

We acknowledge that this case is not beyond circumspection in its relatively narrow analysis of two multiple retailers in a geographically defined area of the UK. However, the method of analysis lends itself favourably to creating improvement actions and generalizable applicability to other sectors where RL flows are relatively deterministic and consistent. Such possible future research applications could be reasonably tested among other retailers, food service organizations and third-party logistics providers managing forward and reverse flows. We believe that this novel comparative analysis case advances RL system comprehension beyond conceptual mathematical models, serving to stimulate further actions and investigations into the operational management of RL systems, to the point where RL is an efficient, reciprocal phase in a circular logistics system where last-mile logistics criteria can prompt mimicking in first-mile recovery decisions.

Answer the questions below. Keep questions with answer

1) Of all of the design factors taken into consideration, which was the most significant from a business perspective?

2) It would seem that there were multiples of databases/sets included in the exercise, rather than, say, the inventory systems. In your opinion, why was that?