- EN
- ES
Supply chain resilience strategies for small and medium-sized enterprises (SMEs) during the Covid-19 pandemic
By Dr. Mustafa Çağrı Gürbüz, Professor at ZLC.
Conventional wisdom, from the Organisation for Economic and Co-operation and Development (OECD) downwards, has it that, given their relative lack of resources, finance, and ‘power’, SMEs must necessarily be at a disadvantage when it comes to coping with significant market or supply chain disruptions.
But is this necessarily true? Colleagues from the UK (Dr. Öznur Yurt, Dr. Sena Özdemir, Dr. Vania Sena, and Dr. Wantao Yu) and I explored this by conducting a case study of the strategies deployed by an SME in the textiles business (a sector notorious for a multiplicity of smaller firms) during the current pandemic. Using semi-structured, in-depth interviews with managers, amplified by corporate and public data, we identified the most significant challenges and responses and incorporated these in a more generalised simulation. Our findings strongly suggest that, at least in some circumstances and given the right choice of strategies, smaller companies may in fact be as well, or even better, placed to weather the storms than their larger competitors.
The case study
The subject of the case study was Lur Textile, a 160-employee company established since 2003 in Izmir, Turkey. The company has some production capacity of its own, but principally operates as an intermediary, orchestrating production by suppliers in Pakistan and Bangladesh with the orders of buyers large and small, most of whom are in the EU or UK.
Like almost every company, Covid-19 has had a significant impact on Lur Textile’s operations. Our first significant finding is that, unlike other exogenous shocks such as an earthquake or a tsunami, the current pandemic is not a single disruptive event: there has been a whole series of disruptions in both supply and demand markets, with different timings, locations and durations.
The most significant disruptions or risks of disruption on the supply side included reduced supplier capacity or complete shutdowns due to the direct impacts of Covid such as staff sickness and absence; shutdowns due to governmental interventions such as lockdowns; and supply disruptions due to shortages of critical raw materials from sources with their own problems further upstream; all these made potentially more critical by the geographical concentration of the supply base in just a few countries. Disruptions on the supply side led to lost sales in certain cases and/or delays is delivering the orders placed.
On the demand side, the firm faced order cancellations by first-tier customers, for a variety of reasons: flexible cancellation policies; their own operations, or those of their customers, may have been impacted by Covid; end consumer demand may have evaporated; or either supply failure, or the direct impact of the pandemic on the firm itself, could make it impossible to fulfill customer orders to time, triggering cancellation.
Second, we found that either way, early and reliable demand forecasting was rendered increasingly difficult. Again, concentration of the market on particular geographies potentially increases the risks.
Across the supply chain, disruptions could be aggravated by the fact that different countries implemented internal rules and restrictions with different notice, timings, and durations, while there were also potential risks due to border closures and other restrictions on foreign trade. Transport links were in some cases broken or delayed, and almost invariably more expensive, the extreme example being the worldwide shortage of container capacity.
Third, our findings reveal that interestingly, though all these disruptions did indeed occur, the company’s management confirmed that they had in fact over-estimated the potential risk involved. This was doubtless due in part to the range of disruption management and risk mitigation strategies that were already built into the structure of the company’s operations, and could be rapidly deployed. Some of the most significant, which we took into our simulation, included using a range of different-sized (though mostly SMEs) firms as supply chain partners; flexible sourcing through identifying and where necessary using multiple, alternative and back-up suppliers; diversifying the customer base into a wider range of sizes, characteristics and regions; seeking out alternative customers for cancelled orders (arguably made possible through postponement strategies keeping the product relevant to the widest market as long as possible); and financial support to trade through, for example, extended payment terms and discounts.
It was evident that the firm’s success in deploying these and other strategies quickly (shorter “time to respond to disruptions”) was in large measure due to their uncomplicated communications built on long lasting, trust-based and strong relations with supply chain partners, areas where it might well be conceived that smaller companies might have an advantage.
The simulation
Informed by this research, we built a simulation model to explore disruptive effects in four areas: supply side disruptions; unpredictabilities and cancellations in demand; transport disruptions and delays; and of course increased costs across the board from materials to transport.
The response strategies we modelled fell into two categories. We looked at strategies such as discounted pricing which are ‘reactive’ as they are offered in response to customer requesting to cancel an order.
We also looked at ‘proactive’ strategies: different supply chain configurations that can at least in theory be built in to mitigate against all manner of disruption risk. As a base line we took a simple ‘triadic’ structure: supply from one source through our company to a single buyer. (That may sound unrealistic but in fact, for any individual product line, it may well be realistic: an apparently complex business may in reality be a bundle of simple triads).
In a second model, we split sourcing over two small suppliers, and introduced a second customer. Because these additional players were still in the same geographical regions, we called this model ‘concentrated’. In the real-life case, Lur Textile did have examples of successfully diversifying its customers, but since the additional customers were in the same area, the UK, as the original buyers, this only went some way to mitigating risk. So in a third model, we placed these new actors in different geographies in a ‘dispersed’ model.
We tested our models against both localised or individual, and global or catastrophic disruptions, with a range of probabilities and cost impacts.
The results of the simulation
Of course, the simulation also recognizes that none of these strategies come for free. Discounting reduces margin, extended payment hits cashflow. More complex supply chains may require greater resource in procurement, on the one hand, and sales/marketing on the other, while spreading orders over more suppliers and customers may reduce transport efficiency (less than full container shipments, for example). Additionally, a more diverse supplier or customer market may reduce the firm’s ‘power’ and position in these markets. If good strategies were free, everyone would use them?
So as might be expected, the simulations showed that as long as economies of scale or scope (market position) are not major determinants of profits, working with more suppliers and buyers, with a wider geographical dispersal helps firms continue serving customers and maintain target profit levels even when major/minor disruptions occur frequently.
The costs of proactive and reactive mitigation strategies must be traded off with the increased profits (relative to the ‘do nothing’ scenario) stemming from better service levels (higher “fill rates”) and better profit margins (eg fewer markdowns).
The simulation revealed that a combination of “proactive” and “reactive” strategies seems to outperform other alternatives. Reactive measures are easier to implement, require less investment, and the impact is immediate. Reactive strategies such as discounting must be used with caution though as buyers might take this as the ‘new normal’ for pricing, or perhaps use threats of order cancellation to secure discounts even if they have no immediate crisis. Proactive strategies involving the network design on the other hand, seem to be more effective in improving profits, but if not already in place require time and investment which may not be available in the emergency. The added benefit of these strategies become even more pronounced with the increased likelihood of catastrophic events.
In conclusion we note the remarks of the management at Lur Textile, who told us that their combination of proactive and reactive strategies “may involve more people, but we know the people, and can get a quick response”. The numbers show that Lur has not only survived, but grown, through the pandemic.
Further research
Collaborative approaches aimed at improving the “individual reliability” of suppliers appear to pay off, especially if catastrophic events are to happen much more frequently (because of climate change, perhaps). We think the impact of improved visibility through advances in big data analytics in particular (for more accurate estimates of demand changes and disruption frequency/duration, early detection of potential problems, supplier disruption detection/sensors, and data-driven decision making) is an interesting future research area, especially as these techniques are likely to become increasingly affordable for smaller companies.
We deliberately restricted our simulation to the first tier up and down stream of the focus firm. However, there is an increasing interest in the ‘ripple’ effect: what happens as disruptions propagate through a supply chain’s connections (the impact doesn’t necessarily diminish just because of distance from the original event), and how one disruptor can trigger another. It would be interesting to apply the parameters we have studied to models of different and longer supply chains to explore how effective proactive and reactive mitigation strategies of the sort we have been studying may be in more complex networks.
This article is based on a work-in-progress research paper written by Mustafa Çağrı Gürbüz, Öznur Yurt, Sena Özdemir, Vania Sena, and Wantao Yu. For further information about the program contact Dr. Mustafa Çağrı Gürbüz at mgurbuz@zlc.edu.es
