ORIGINAL RESEARCH

 

Efficient trade lane selection: A total economic cost perspective on shipping lines

 

 

Alwyn Hoffman^I; Jacob van Rensburg^{II, III}; Sonja Grater^II

^ISchool of Electrical, Electronic and Computer Engineering, Faculty of Engineering, North-West University, Potchefstroom, South Africa
^IISchool of Economic Sciences, Faculty of Economic and Management Sciences, North-West University, Potchefstroom, South Africa
^IIISAAF, Johannesburg, South Africa

Correspondence

 

 

---

ABSTRACT

BACKGROUND: Efficient logistics performance is vital for global trade, yet traditional cost assessments often overlook the economic impact of time delay variability. Especially in maritime logistics, these delays can generate substantial indirect costs. This study addresses a critical gap by integrating time-related uncertainty, which contains the implicit cost aspects, into logistics cost modelling to support better decision-making in trade lane selection
OBJECTIVES: The study aims to quantify both direct and indirect logistics costs arising from time delays and variability across international shipping routes. Focusing on South Africa's import trade, it introduces a replicable total economic cost (TEC) model that enables cargo owners and freight forwarders to optimise route and shipping line choices based on holistic cost performance
METHOD: Using a dataset of 5374 import shipments (2017-2023) from a South African freight forwarder, the study segments total logistics chains into ocean, port and land legs. Time delays and their variability are analysed per segment. Direct and indirect costs - such as the cost of capital tied up in inventory, stock shrinkage and lost sales - are modelled using percentile-based TEC calculations across buffer stock strategies
RESULTS: The ocean leg was the largest contributor to time delays and cost variability. Shipping lines with lower delay variability enabled significantly lower TEC values and smaller buffer stocks. The TEC model revealed that variability-driven costs often exceeded direct logistics expenses
CONCLUSION: Minimising delay variability, and not just transport time, can significantly reduce logistics costs. The TEC model supports better strategic alignment of shipping line and trade lane choices
CONTRIBUTION: This study provides a practical, data-driven methodology for quantifying total logistics cost under uncertainty and enabling optimal choices of trade lanes and service providers, addressing a key challenge in global supply chain optimisation

Keywords: trade lane; logistics; total economic cost; time variability; statistical distributions.

---

 

 

Introduction

Ocean transport is the primary mode for global merchandise shipping and handles the largest share of international goods movement, especially in developing countries (UNCTAD 2023). Choosing the most appropriate shipping line and port combinations is therefore critical to obtaining improved logistics performance (Aaby 2012; Song 2011).

Logistics performance is typically measured through transit times, financial outlays and the consistency of service delivery. As transit time delays can be translated into cost (Haartveit, Kjøstelsen & Jacobsen 2007), cost remains the primary output measurement of logistics performance. Research on logistics costs emphasises monitoring and controlling expenditures at multiple levels (Andrejić et al. 2018). However, traditional frameworks often overlook indirect costs.

This study applies a 'Total Economic Cost' (TEC) model that converts delays in logistics processes into their equivalent economic cost. The TEC translates all time-related economic impacts into monetary terms, thus eliminating the need to account for time and cost separately. The TEC model incorporates opportunity costs, hidden logistics costs and externalities, providing a holistic measure of logistics performance (Hoffman, Mutendera & Venter 2023). The model accounts for a range of direct and indirect expenses, including financing charges tied to inventory, reductions in stock levels and the costs associated with stockouts. This study extends earlier research (Hoffman 2019; Hoffman, Lusanga & Bhero 2013; Minken & Johansen 2019 and most notably, Hoffman et al. 2023). Using a TEC model provides a comprehensive framework to quantify logistics performance costs, ensuring key metrics such as time delays and service reliability are accurately translated into economic terms, which is crucial for optimising decision-making in maritime logistics.

Because freight forwarders manage and integrate activities across the logistics chain, they are uniquely positioned to observe real-world operational processes and associated economic outcomes (see Özcan et al. 2024). Accordingly, the empirical evidence used in this analysis was sourced from a South African forwarding company and captures the conventional sequence of activities in an import process. Although this study focuses on South Africa, the methodology can be applied to any country participating in global trade.

The paper is structured as follows: the 'Literature review' section provides an overview of logistics performance and the TEC model. The 'Objective of the total economic cost for this study' section states the objective of this article. The 'Methodology' section outlines the assessment method and data collection. The 'Total economic cost model specification' section outlines the theoretical foundation of the TEC model, which underpins this study's methodological approach. The empirical findings are presented in the 'Results and findings' section, and the 'Conclusion, limitations and recommendations' section offers concluding remarks together with recommendations for improving logistics efficiency.

 

Literature review

Logistics networks are integral to extended supply chains and must maintain predictability, sustainability, reliability and agility. The coronavirus disease 2019 (COVID-19) pandemic placed exceptional stress on these capabilities, testing system resilience and revealing several system weaknesses (Notteboom, Pallis & Rodrigue 2021:180).

Notteboom et al. (2021) analysed the timing and geographic spread of COVID-19 supply and demand shocks and compared them with those of the 2008/2009 financial crisis. The evidence suggested enhanced resilience across shipping lines, terminal operators and ports, attributable to a combination of renewed risk recognition and structural organisational modifications. Also, COVID-19 reconfirmed the market position and bargaining power of shipping lines, with an upsurge in freight rates and a positive impact on their financial results (Notteboom et al. 2021:207). These developments emphasise the value of understanding logistics costs when navigating major disruptions.

However, they warned of the potential impact over the longer term amid 'highly uncertain conditions generated by a new wave of COVID-19 cases and restrictions in countries around the world' (Notteboom et al. 2021:186). Although the acute phase of COVID-19 was relatively short lived, Notteboom et al. (2021) noted that the broader uncertainty during the period could affect logistics networks beyond the immediate crisis. Similar observations were made in later work (e.g., Ciravegna & Michailova 2022:173), and the longer-term impact was also experienced in the logistics industry in developing contexts such as South Africa (see Grater & Chasomeris 2022). The pandemic therefore serves as a useful historical example of how rapidly logistics performance can deteriorate and why a comprehensive analysis of logistics costs remains essential for improving resilience and supporting informed decision-making.

Further research assessing logistics costs (e.g. Andrejić et al. 2018) emphasised the significance of monitoring and controlling logistics expenditures at global, national and organisational levels. However, while the existing research tends to cover major cost categories like transportation, warehousing and handling, it does not fully account for the full range of direct and indirect or implicit costs. A TEC model offers the potential to accurately quantify opportunity costs, hidden logistics costs and externalities caused by variable time delays, and in this regard, the TEC model provides a holistic measure of logistics performance.

Hoffman et al. (2023) applied a TEC model to transport corridors in the SADC region, using Zambia and Lusaka as examples. They found that variability in time delays significantly impacts total logistics costs, often more than only direct transport costs. For instance, the Beira corridor, despite its low direct transport cost, had high total costs because of time delay variability, making it less favourable compared to other corridors. Within this scenario, the financial impact of time delays could represent up to 80% of the total costs associated with the corridor.

Consequently, these findings highlight the usefulness of applying a TEC model to improve understanding of logistics costs. The following sections will explain how the TEC model is applied to the South African import trade lanes analysed in this study.

Objective of the total economic cost for this study

Logistics performance is contingent upon the successful execution of multiple sequential steps, each following a logical flow within the logistics process. For international shipping, this includes purchasing, forwarding and clearing, international shipping events and final delivery. The TEC framework facilitates comparison across alternative supplier routes by examining the sources of friction that impede trade flows, including delays at sea, at borders and within ports, as well as service-related disruptions captured in the data.

The version of the TEC applied here draws on the freight forwarder's direct cost structure (covering freight, landside operations, storage, transport and similar charges) and incorporates the indirect cost factors documented in prior studies. Therefore, the TEC components explained in Table 1 are included in the model.

 

 

Based on the step-by-step timing recorded in the dataset, the analysis shows that the ocean-freight segment required an average of 35.97 days, whereas the entire physical chain amounted to approximately 40.40 days. Thus, the duration of the ocean segment and the choice of export origin and carrier emerge as the key factors in limiting total physical delays and improving logistics performance in terms of both time and cost.

 

Methodology

This study builds on previous work (Hoffman 2019; Hoffman et al. 2013, 2023; Minken & Johansen 2019) and uses a TEC model to quantify logistics performance costs. The TEC framework incorporates a combination of direct and indirect cost elements, including capital charges tied to inventory, shrinkage-related losses and the financial impact of stockouts.

Although the shipping environment is dynamic and conditions evolve continuously, the TEC model retains practical relevance because it functions as a flexible framework rather than a fixed prediction tool. Freight forwarders routinely capture the operational data required to apply the model, allowing TEC estimates to be updated as conditions change. Importantly, while many implicit costs - such as stock-in-transit financing, buffer-stock requirements and disruption risks - are recognised by practitioners, they are seldom evaluated collectively. The TEC consolidates these components into a single cost measure, enabling cargo owners and forwarders to compare trade lanes and carrier options systematically and to quantify the economic impact of variability in delivery performance. In this way, the model complements existing decision-making rather than serving as a purely theoretical construct.

The empirical evidence for this study is drawn from transaction-level records compiled by a South African freight forwarder for the period July 2017 to August 2023. The dataset covers around 5374 individual shipments and integrates information from technical service providers, including documentation process timelines, container-tracking data reflecting the physical movement of cargo and associated financial records.

Table 2 summarises the dataset that was obtained for this study.

 

 

The TEC model developed by Hoffman et al. (2023) was adapted for this study into the steps explained in Table 3.

 

 

As shown in Table 3, the TEC framework was adapted for this study to translate the detailed shipment-level data into a unified measure of total logistics cost across different trade lanes. The adaptation integrates direct charges, operational cost components and the time variability captured in the dataset to evaluate how delays and route-specific performance influence overall logistics efficiency. A more detailed description of the framework is published in Van Rensburg (2024). By aligning the model with the structure of the available freight-forwarder data, the TEC output provides a comparable cost measure for each origin-shipping line combination, allowing the identification of the factors that most strongly affect end-to-end trade lane performance.

Total economic cost model specification

This section outlines how the aggregated cost inputs were combined to calculate the different components that constitute the TEC.

Direct costs

The cost parameters applied for direct cost calculations are indicated in Table 4. The majority of the cost parameters were obtained from the datasets indicated earlier, and other costs were also incorporated based on the work by Hoffman et al. (2023):

 

 

Direct costs were calculated using the Equation 1:

where RTD_i = denotes the round-trip delay (in days) for trade lane i, and TD_i represents the one-way delay from origin to destination.

where DC_trip,i is the driver cost per trip for trade lane i, and DC_monthly is the monthly driver employment cost (Equation 2).

where FC_i is the cost of fuel lane i, Dist_i is the origin to destination distance for trade lane i, FuelCost is the fuel cost per litre and FuelEcon is fuel economy in km/litre (Equation 3).

where DTC_i is the cost of a direct trip, N3T is the N3 toll fees, whereas OC represent other costs (e.g. subsistence payments for the driver) (Equation 4).

where NumTripspm_i is the amount of monthly trips for the i-th trade lane (Equation 5).

where TCpm_i is the total monthly truck cost for the i-th trade lane, and Instpm is the monthly instalment for each truck (Equation 6).

where TranspCost_i is the total trip cost for the i-th trade lane (Equation 7).

Transport costs for each trade lane were expressed as a proportion of the cargo value, consistent with the retail and tyre-import categories represented in the sample.

Variable time delay costs

To minimise overall costs, cargo owners implement a specific buffer stock policy. Above-average delivery delays can still cause stock-outs and economic losses despite cost-minimising buffer stock levels. Therefore, calculating the TEC requires assessing the costs across all possible time delays. This involves determining the percentiles of time delays for each trade lane and calculating the expected TEC for each percentile. By combining these components, the model yields the total expected cost of cargo deliveries based on the observed distribution of time delays from the cargo and truck datasets. Beyond the direct transport costs, the following indirect cost elements were identified for import operations, following the approach of Hoffman (2019):

Time delay impact: Because time-dependent costs rise as stock remains longer in transit, the total cost is calculated by integrating the cost incurred at each possible delay duration with the probability of the delay occurring:

where Cost(t) is the cost associated with a delay of length t, and p(t) is the probability distribution of transit times (Equation 8). As the true distribution is not known, an empirical approximation is used by averaging across all observed delay percentiles (Equation 9):

where Cost_i denotes the cost corresponding to the i-th percentile of observed delays. This method is applied to all time-dependent cost components that follow; for brevity, the summation form is not repeated for each category:

Interest on stock-in-transit: The importer carries the financing cost of goods from the point at which ownership transfers - typically when the cargo leaves the origin facility or is loaded on board (in approximately 88% of cases in this dataset, under EXW and FOB terms¹) - until final delivery. The interest cost for stock-in-transit is therefore expressed as a fraction of the cargo value over the duration between shipment and receipt (Equation 10, Equation 11, Equation 12).

Therefore:

where:

• CI = Cost of interest p.a.

• IR = Interest rate p.a.

• VGIT = Value of goods in transit

• VAC = Value of annual consumption

• TD_i = Transit delays (calculated in days) for the i-th percentile

Shrinkage in transit: Because shrinkage losses rise with longer transit times, total shrinkage was estimated by averaging the shrinkage cost across all observed delay percentiles (Equation 13 and Equation 14):

where TotShrinkage denotes the total shrinkage across all possible transit times, Shrinkage_i is the shrinkage associated with the i-th percentile, Shrinkagepd is the daily shrinkage rate and TDPercCorr_i represents the transit time (in days) for the i-th percentile. The formulation reflects that, as each day passes, a smaller quantity of stock remains exposed to further shrinkage.

Out-of-stock losses: Losses in sales or production may occur when delivery delays exceed the expected arrival window ('delivery expected' to 'cargo delivered'). If buffer stock is maintained, losses only arise when an unexpected delay surpasses the duration that the buffer can cover (Hoffman 2019):

Retail: For retail operations - such as those represented in this model - it is assumed that sales losses occur once the buffer stock is exhausted and the next delivery has not yet arrived (Equation 15, Equation 16, Equation 17, Equation 18).

where

where:

• LRI = Loss in retail income

• FSL = Fraction of sales lost

• GM = Gross margin

• OTL = Operational time loss

• ADT = Actual delivery time

• SDT = Standard delivery time

• BSP = Buffer stock period

• BSS = Buffer stock size

• UR = Usage rate

Storage costs incurred in maintaining buffer stock, calculated as a share of the cargo value (Equation 19; Equation 20; Equation 21):

Therefore:

where:

• SC = Storage cost p.a.

• SR = 1 year storage rate per unit

• SRFrac = 1 year storage cost as fraction of unit value

• MIS = Max inventory size in units

• MDT = Min delivery time

Equations 1-21 provide a framework for expressing the total cost arising from transport and logistics delays as a fraction of the overall value of goods purchased.

where in Equation 22:

TEI = Total economic impact

As noted earlier, the TEC is evaluated at the buffer-stock period (BSP) that minimises total cost. Because TEC depends on the full distribution of actual delivery times - and no closed-form expression exists for this distribution - the optimal BSP cannot be derived analytically. The BSP is therefore identified through a numerical procedure that evaluates TEC across observed delay percentiles for different BSP values. The BSP that yields the lowest TEC is then selected as the optimum.

Total economic cost sensitivity concerning cost parameters

The preceding equations show that the TEC, from the cargo owner's perspective, is dependent on various cost parameters. To assess this sensitivity, TEC was recalculated across a range of parameter values, as outlined in Table 5.

 

 

Results

This section discusses the results generated by the TEC model and presents the key findings derived from the analysis.

Direct transport and logistics cost

Table 6 presents a summary of direct transport and logistics costs for each origin-destination combination, together with the average distance from the origin port to Durban, the corresponding model-based transport cost and the charges observed in the dataset.

 

 

Total economic cost model costs

For each origin-shipping-line combination, delay percentiles were calculated for the ocean, port and land segments and for the total trade lane. The cost contribution of each of the segments to the overall TEC was then quantified. When isolating the TEC impact of a specific segment (e.g. the ocean leg), the model applies average values for the remaining segments (e.g. port and land). By isolating the contributions to time variability of each segment of the total trade lane, it was possible to separately calculate the TEC component contributed by each trade lane segment. This approach highlighted the significant impact of the ocean segment on delivery time variations.

Figure 1 illustrates delay percentiles across all segments. The ocean leg clearly dominates total transit time, highlighting the need to select the best-performing supplier-shipping-line route. The ocean segment averaged 35.97 days compared with 40.40 days for the entire physical chain.

 

 

The second graph in Figure 1 indicates that delays on the waterside and landside occur at similar levels. This implies that terminal performance affects total delay equally, regardless of whether inefficiencies arise from ship-to-shore crane operations or landside handling activities.

The bottom graph in Figure 1 indicates that inbound road transport has the smallest overall impact on delays, although notable disruptions occur from the 96th percentile onward. The pattern also suggests that the prevailing inbound approach (short haul to a depot, then long haul to destination) may merit reassessment, given the prolonged dwell times at depots or warehouses that increase storage costs and influence buffer-stock decisions.

The next results illustrate a clear pattern in how TEC changes with the buffer-stock period. We initially examine how buffer stock levels influence each cost component (interest costs, stock holding costs, lost sales costs) and their impact on total cost, explaining why a cost-reducing stock level minimises costs. We then calculated TEC contributions by each trade lane segment (ocean, port and road) as a function of buffer stock level to quantify their relative contributions.

Figure 2 illustrates the relationship between logistics costs and the buffer-stock period.

The variables considered include interest on inventory, storage costs, lost-sales costs and delay-related costs. Increasing the buffer-stock period raises interest and storage costs but reduces lost-sales risk. Interest and storage costs therefore rise directly with longer buffer periods (see top left and right panels of Figure 2). The ocean leg contributes most to delays, with the loss of sales eliminated at around 70 days of buffer stock. The graph shows an initial decline in total costs with increasing buffer stocks, reaching a minimum at the ideal buffer stock level and then rising again as interest and stockholding costs further increase with further buffer stock increases.

As the performance of individual trade-lane components differs substantially, the calculations were repeated by isolating the delay and variability associated with each specific segment. In this approach, one segment was allowed to vary while all others were held at their average delay values. This made it possible to quantify the separate contribution of each segment to the overall TEC.

Figure 3 illustrates how the cost contributions of the ocean, port and road segments change with different buffer-stock levels. Each curve shows a clear minimum point - the ideal buffer-stock period. For the total trade lane, this optimal level is approximately 34 days, closely matching the average duration of the ocean leg. When segment-level variability is considered individually, the minimum cost point aligns roughly with that segment's average transit time. As with earlier findings, the ocean leg remains the dominant contributor to time-related logistics costs. Importantly, lower delivery-time uncertainty allows firms to maintain smaller buffer stocks without incurring significant sales losses.

When the cost impacts associated with delay variability are disaggregated, the ocean leg is the most significant contributor to time-related logistics costs for imports into South Africa. Table 7 indicates the transport and logistics costs arising from delay variability, expressed as a fraction of cargo value for each trade lane. The analysis covers the top 14 origin countries and the top 10 shipping lines (see the tables in Appendix 1). A dash (-) indicates that a particular country-shipping-line combination is not serviced and is therefore not applicable.

The table yields several insights. For France, Cosco achieves the lowest TEC from time delays (3.9% of cargo value) and Mediterranean Shipping Company (MSC) the highest (12.1%), a notable gap given that direct logistics costs average about 9.2% of product value. This difference reflects Cosco's shorter average ocean transit time (22.6 days) compared with MSC (38 days). Such differences may arise from several operational factors, including vessel size and deployment, route structure, service frequency and the number of intermediate port calls. While the dataset does not permit isolation of these individual drivers, the observed transit times capture their combined effect as experienced by cargo owners. From a decision-making perspective, these differences are economically meaningful, as longer and more variable transit times increase buffer-stock requirements and implicit costs, thereby influencing freight forwarders' carrier-selection decisions.

Across shipping lines, considerable variation is observed in both average ocean transit times and their variability. Evergreen shows the longest average delays, although with relatively low variability, suggesting consistently slow performance. In contrast, MSC and ONE exhibit the largest standard deviations, indicating highly inconsistent transit times. Cosco, Maersk and Gold Star Line display shorter and more stable transit times, while CMA-CGM and Orient Overseas Container Line (OOCL) rank among the best performers in terms of both average duration and consistency. These differences in reliability directly influence the TEC outcomes for each trade lane.

The shipping lines with the highest variability in ocean transit times are ONE and MSC, with standard deviations of 28.63 and 24.68 days, respectively. At the opposite end, OOCL and Gold Star Line show the lowest variability, with standard deviations of 3.79 and 4.43 days. This indicates that ONE and MSC exhibit far more dispersed transit-time outcomes - leading, on average, to higher time-delay costs - whereas OOCL and Gold Star Line display more consistent performance and therefore lower associated delay costs.

The results also illustrate the extent of market coverage provided by each shipping line. MSC serves nearly all listed markets, Maersk connects to 11, and ONE to 10. All three lines operate services to Thailand, which emerges as the best-served origin and also the one with the widest range of delay-related TEC outcomes (from 4.2% to 24.5%).

The minimum buffer-stock period varies considerably across origin-shipping-line combinations. For example, shipments from Thailand require anywhere from fewer than five days to more than 33 days of buffer stock, depending on the carrier; a similar range is observed for shipments from the Netherlands. A clear pattern emerges: shipping lines with lower variability in transit times consistently enable shorter buffer-stock requirements, thereby reducing total logistics costs. This illustrates how freight forwarders can improve overall performance by selecting carriers whose delay variability aligns best with the specific origin market.²

 

Conclusion, limitations and recommendations

This study developed a TEC model to support more informed logistics decision-making in international trade. The TEC model quantifies the impact of time delay variability on direct and indirect logistics costs across different trade lanes. It built on previous work and added factors like time delay variability, shrinkage of stock and out-of-stock losses. Although the shipping environment is dynamic, the TEC framework remains operationally relevant because it can be recalculated as new data becomes available, enabling ongoing comparison of trade lanes and carrier performance. While the results show that shipping-line choice plays a particularly influential role given its strong impact on transit-time variability, the value of the TEC model extends beyond carrier selection alone. The framework also informs buffer-stock policies, highlights the cost implications of reliability differences across logistics segments and enables systematic comparison of trade-lane configurations as operating conditions evolve.

The findings reveal substantial differences in transit times across consignments, with delays ranging from only a few days to well over 100 days in certain cases. This underscores the need to quantify how delivery-time variability affects TEC from the cargo owner's perspective. It also highlights the importance for freight forwarders to carefully evaluate trade-lane and shipping-line choices, as differences in delay performance can materially influence the overall economic cost of imported goods.

The ocean segment, averaging 35.97 days, had the strongest influence on total delays. Waterside and landside port operations contributed similarly, while inbound road transport had the smallest effect except for occasional large disruptions past the 96th percentile. The findings also suggest a review of the short-haul-long-haul inbound model, given the extended time many consignments spend in depots or warehouses.

Investigation of different buffer stock scenarios revealed that for each shipping line-trade lane combination, an ideal buffer stock period can be found. This results from the fact that interest and storage costs rise, while the cost of lost sales or production decreases, with increased buffer stock periods. For the available data set, an ideal buffer stock level of around 34 days minimised total logistics costs. For those shipping lines with low time delay variability, the buffer stock period could, however, be reduced to below 10 days. The findings emphasise the importance of aligning shipping lines with specific origin countries, as lower optimal buffer-stock requirements signal more reliable services and reduced TEC.

These results offer practical guidance for freight forwarders and cargo owners seeking to enhance logistics performance by selecting the most suitable shipping lines for specific routes, where sufficient operational and historical data are available. They also support the identification of appropriately calibrated buffer-stock levels by highlighting origin-carrier combinations that minimise TEC.

In addition, the variation in optimal buffer-stock requirements sheds light on how inventory decisions interact with key cost drivers affected by time delays - such as interest on stock in transit, storage costs and the risk of lost sales.

Despite the substantial dataset that was available for this research, several limitations remain.

Firstly, the dataset used, although substantial, represents only a sample of 5374 shipments, which may not fully capture the broader variability in global trade patterns. Secondly, the analysis focuses on limited combinations of country of origin and shipping line schedules, restricting the generalisability of the findings to other trade routes and logistical conditions. Thirdly, the TEC model relies on available data for specific elements, and gaps in comprehensive information, particularly indirect costs, could affect the model's accuracy.

Lastly, the analysis is constrained by historical data and may not account for evolving trends in shipping practices or regulatory changes impacting logistics performance. These limitations suggest areas for further research to enhance the model's robustness and applicability.

Regardless of these limitations, the study demonstrated that faster and more predictable movement of goods through the supply chain improves trade volumes and reduces logistics costs, highlighting the need for the industry to apply more robust data analytics in order to improve logistics performance. More specifically, where sufficient historical and operational data are available, the application of the TEC can significantly enhance the freight forwarder's ability to make informed shipping-line selection decisions for specific trade routes.

 

Acknowledgements

This article includes content that overlaps with research originally conducted as part of Jacob (JE) van Rensburg's doctoral thesis titled 'Investigating the use of data analytics towards improved logistics performance for South African imports', submitted to the Faculty of Economic Sciences, North-West University, in 2024. The thesis was supervised by Sonja Grater and Alwyn Hoffman. Portions of the data, analysis and discussion have been revised, updated and adapted for publication as a journal article. The original thesis is publicly available at: https://repository.nwu.ac.za/server/api/core/bitstreams/7b054d9d-df5c-46d2-8cee-ab6f947a4c96/content. The author affirms that this article complies with ethical standards for secondary publication, and appropriate acknowledgement has been made of the original work.

Competing interests

The authors declare that they have no financial or personal relationships that may have inappropriately influenced them in writing this article.

CRediT authorship contribution

Alwyn Hoffman: Conceptualisation, Methodology, Formal analysis, Software, Resources, Writing - review & editing, Supervision. Jacob van Rensburg: Methodology, Formal analysis, Investigation, Data curation, Resources, Writing - review & editing. Sonja Grater: Writing - original draft, Validation, Resources, Writing - review & editing, Supervision. All authors reviewed the article, contributed to the discussion of results, approved the final version for submission and publication and take responsibility for the integrity of its findings.

Ethical considerations

This article followed all ethical standards for research without direct contact with human or animal subjects.

Funding information

This research received no specific grant from any funding agency in the public, commercial or not-for-profit sectors.

Data availability

The majority of the data were obtained from a logistics company in South Africa (a permission letter for the use of the data is available upon request). The data were further supplemented with secondary, open-access data from Transnet National Port Authorities

Disclaimer

The views and opinions expressed in this article are those of the authors and are the product of professional research. They do not necessarily reflect the official policy or position of any affiliated institution, funder, agency or that of the publisher.

 

References

Aaby, B.C., 2012, 'An analysis of shipping lines' selection criteria when choosing European container terminals', Master's thesis, Høgskolen i Molde - Vitenskapelig høgskole i logistikk.         [ Links ]

Andrejić, M., Bojović, N., Kilibarda, M. & Nikoličić, S., 2018, 'A framework for assessing logistics costs', International Journal of Logistics Management 27, 770-794.

Ciravegna, L. & Michailova, S., 2022, 'Why does the world economy need, but will not get, more globalisation in the post-COVID-19 decade?', Journal of International Business Studies 53(1), 172-186. https://doi.org/10.1057/s41267-021-00467-6        [ Links ]

Grater, S. & Chasomeris, M.G., 2022, 'Analysing the impact of COVID-19 trade disruptions on port authority pricing and container shipping in South Africa', Journal of Transport and Supply Chain Management 16, 772. https://doi.org/10.4102/jtscm.v16i0.772        [ Links ]

Haartveit, E.Y., Kjøstelsen, L. & Jacobsen, B.S., 2007, Time is money - Quantifying logistics costs by measuring time, The Norwegian Forest and Landscape Institute, Norway, viewed 04 September 2024, from https://nibio.no/en/publications?year%5B%5D=2007-2007&filters=1&q=&page=4.

Hoffman, A., 2019, Quantifying the impact of freight transport performance on the total economic cost of cargo importers, IEEE Intelligent Transportation Systems Conference, Auckland.

Hoffman, A.J., Lusanga, K. & Bhero, E., 2013, 'A combined GPS/RFID system for improved cross-border management of freight consignments', in 2013 AFRICON, Pointe aux Piments, Mauritius, September 09-12, 2013, pp. 1-8, IEEE. https://doi.org/10.1109/AFRCON.2013.6757841

Hoffman, A.J., Mutendera, C. & Venter, W.C., 2023, 'Comparing transport corridors based on total economic cost', Journal of Advanced Transportation 2023, Article ID 6336630, 1-15. https://doi.org/10.1155/2023/6336630

Minken, H. & Johansen, B.G., 2019, 'A logistics cost function with explicit transport costs', Economics of Transportation 19, 100116. https://doi.org/10.1016/j.ecotra.2019.04.001        [ Links ]

Notteboom, T., Pallis, T. & Rodrigue, J.P., 2021, 'Disruptions and resilience in global container shipping and ports: The COVID-19 pandemic versus the 2008-2009 financial crisis', Maritime Economics & Logistics 23(2), 179-210. https://doi.org/10.1057/s41278-020-00180-5        [ Links ]

Özcan, S., Oflaç, B.S., Tokcaer, S. & Özpeynirci, Ö., 2024, 'Mastering timely deliveries using dynamic capabilities: Perspectives from logistics service providers and shippers', The International Journal of Logistics Management 35(5), 1653-1677. https://doi.org/10.1108/IJLM-03-2023-0089        [ Links ]

Song, S.Y., 2011, 'A study on the factors of choosing the liner shipping companies using AHP method by international freight forwarder', International Commerce and Information Review 13(2), 95-117. https://doi.org/10.15798/kaici.13.2.201106.95        [ Links ]

Transnet Port Authorities, 2024, Port statistics, viewed 04 July 2025, from https://www.transnetnationalportsauthority.net/Commercial%20and%20Marketing/Pages/Port-Statistics.aspx.

Van Rensburg, J.E., 2024, 'Investigating the use of data analytics towards improved logistics performance for South African imports', PhD thesis, North-West University, viewed 04 September 2024, from https://repository.nwu.ac.za/server/api/core/bitstreams/7b054d9d-df5c-46d2-8cee-ab6f947a4c96/content.         [ Links ]

United Nations Conference on Trade and Development (UNCTAD), 2023, Review of maritime transport, viewed 04 September 2024, from https://unctad.org/topic/transport-and-trade-logistics/review-of-maritime-transport.

 

 

Correspondence:
Sonja Grater
sonja.grater@nwu.ac.za

Received: 16 Oct. 2025
Accepted: 16 Jan. 2026
Published: 04 Mar. 2026

 

 

1. EXW (Ex Works) and FOB (Free On Board) are Incoterms indicating when responsibility and risk transfer from seller to buyer: at the seller's premises under EXW and once goods are loaded on board the vessel under FOB.
2. For the purposes of this article, all numerical results were not included but are available at a detailed level from the authors.

 

 

Appendix 1

 

 

Table 2-A1 summarises the country of export in terms of total time delays from SOB to cargo delivered.

 

 

Table 3-A1 shows the products (42 910 different lines - each representing a product imported) shipped via the respective shipping lines.

 

 

Table 4-A1 reports descriptive statistics for the ocean transport segment (Shipped on Board and Actual Time of Arrival [SOB to ATA]for each shipping line represented in the dataset).