Consumer goods companies have to navigate through a rather heterogeneous and increasingly challenging market, with demand patterns being greatly influenced by geographic and social-demographic aspects. Tailoring a distribution strategy in accordance with all these different factors is paramount to success.
Our client operates in a very scattered market with over one million outlets. The challenge was to evaluate consumption patterns across these outlets and to optimize the product distribution on an outlet-level, aiming to identify opportunities to capture market share.
By leveraging a country-wide survey to outlets, data was collected on product-level sales. A three-step methodology was deployed to pinpoint distribution opportunities:
Outlet segmentation - Application of a clustering algorithm to characterize consumption patterns relating to price and product characteristics throughout the country;
Demand model - Development of an analytical engine to evaluate the sales potential of introducing any product in a given outlet, considering the substitution effects on similar products.
Range optimization - An optimization algorithm was developed to identify the products with the largest potential impact on total volume and margin for each outlet according to its specific consumer demand patterns.
By implementing the proposed distribution strategy, opportunities were identified leading to a 7% potential increase in sales volume through the optimization of the ranged products. Furthermore, with almost 70% of the volume increase originating from only 20% of outlets, substantial gains can be achieved with a considerably lower investment in distribution
The development of a dashboard to explore the identified opportunities enabled users to explore demand patterns and distribution gaps across the country which improved the actionability of the proposed distribution strategy and provided valuable market insights to our client.
While supply chain network design has been around for some time, organizations are now starting to acknowledge that the supply chain structure plays a too important role to be left over to chance, being able to differentiate them from the competition. Moreover, supply chains have become more complex and global due to the growing business environment volatility, as well as by virtue of the set of sophisticated technologies that have been reaching the market (and which allow the gathering of data from many different sources). So how can companies deal with these challenges and design a profitable network? A supply chain network is defined by all intervenients between the products’ origin and the final customer. The facilities’ location and their flows are of utmost importance, accounting for the majority of the supply chain associated costs and defining the service level provided to the customers. The main goal of network design is to balance these two levers and align them towards the companies’ strategic positioning, supported by robust analytical modelling. On another perspective, the complexity of the supply chain is directly linked to both internal and external factors that coexist around the business environment. The supply chain should be able not only to adjust to externalities such as changes in demand patterns or government regulations but also to accommodate the companies’ growth plans or new revenue streams (new products and/or new markets). In this insight, we explore two practical applications of network design that tackle distinct viewpoints on the topic: customer-orientation and efficiency-orientation. An efficiency-orientated approach on network design for a consumer electronics retailer Our client, a consumer electronics retailer, was facing the need of reviewing its supply chain to reduce the distribution costs. The challenge was tackled using a two-step optimization approach. Firstly, given the intended number of cross-docking hubs, the best locations to set up the operation were found. Secondly, for each configuration, operational costs were calculated using a state-of-the-art routing algorithm and the optimal number of hubs was determined. A simulator was then developed to assess the feasibility of the solutions since there were other operational details to be considered, as the variability in travelling or installation times. This validation was essential to ensure the intended service level. The optimization of the distribution network allowed the retailer to reduce its operational costs by 3%. Tackling network design in an online fashion retailer from a customer-centric standpoint The online fashion retailer aimed to assess all the impacts of opening a new warehouse with a focus on the client experience (i.e., delivery lead time, predictability) while not jeopardizing the operational distribution efficiency (i.e., shipping fees, warehousing costs). The first step of the methodology consisted on the development of a customized predictive model with the goal of identifying how the experience may impact client retention. Then a supply chain analytical model was applied to forecast shipping fees, lead times and anticipate warehousing costs. Several scenarios were outlined and the best location for the new warehouse was found. This study allowed to conclude that the opening of the new unit would increase the margin by 8%. Leverage the future performance of your supply chain As we can learn from the previous examples, irrespectively of the initial motivation behind the supply chain optimization, any network design problem needs to balance the client experience or service level against the operational efficiency. From a methodological perspective, an effective network design requires: a holistic perspective of the supply chain, with the identification of the costs, capacity levels and lead times in each point of the network; a forecasting approach, capable of foreseeing changes in the volume or customers’ expectations during a period of several years; a solid optimization or simulation approach, to evaluate different network design possibilities; a detailed roadmap towards the sustainable implementation of the new configuration.
Supply chain optimization is a topic with increasing importance, not only because efficiency is key in a progressively more competitive environment, but also because consumers are continuously more demanding about assortment diversity, lead time and price. Using analytics for supply chain optimization might just be the key to get the edge over the competition. From understanding to decision-making To run successful supply chain analytics projects, five guidelines should be followed. 1. Understanding the operation and mapping indicators before using analytics Despite seeming an effortless task, disregarding the significance of a decision on adjacent areas is excessively common, mainly because managers are focused on their own operational scope and tend to neglect what surrounds it. Besides undermining the project’s expected results, ignoring side effects often becomes a pain point for change management in later phases. 2. Modelling the as-is operation and fully grasping its intrinsic characteristics The initial analysis frequently aims at modelling the as-is situation and comparing it with real costs and KPIs, guaranteeing that the model is a reliable representation of the reality. It is important to discuss all assumptions and outcomes with the involved stakeholders to ensure some comfort about what is and what is not being considered. 3. Mapping and exploring alternatives for further testing and evaluating While taking other players moves into consideration is obviously valuable, bringing operational and mid-level teams together to brainstorm about new possibilities is often game-changing and might result in a competitive advantage. 4. Validating the developed models and retrieving final indicators and results Considering the defined scenarios, it is time to adjust the models created to predict how each indicator would react to new supply chain configurations. Merging analytics’ prowess and business expertise is pivotal to successfully accomplish this mission. Validating the model and verifying if each lever leads to the expected result is a critical step to guarantee confidence in the results. 5. Gathering relevant stakeholders and making the final decisions Since decision-makers always have the last call, independently of the quality of the analytics project, we reinforce the importance of involving every internal and, if possible, external stakeholders. The success of the change management stage largely depends on everyone's confidence in the benefits of the project. Tailor-made Supply Chain optimization: a case study Take the work we have recently done with a large electronics retailer, in which a challenge emerged to reshape the entire Iberian supply chain and to redesign the retailer's global network. Let us go through the five guidelines. 1. Understanding the operation and mapping indicators before using analytics In this context, the first phase of the project consisted of understanding the problem and analyzing the involved business requisites. At the same time, the team concentrated on mapping all relevant indicators (e.g., service level, lead times, etc.) and costs (rental cost, handling and stock costs for both warehouses and stores, transportation to stores and home-delivery and structure costs). 2. Modelling the as-is operation and fully grasping its intrinsic characteristics The following step of the project was to explore the available data and build an analytical model that represented the operation of the company at the time. Through statistical analysis and optimization modelling, the team was able to replicate the logistics process with close-to-reality operational indicators such as distance travelled during transportation, warehouse productivity and stock levels. 2. Mapping and exploring alternatives for further testing and evaluating Together with the retailer and according to the company’s strategic goals, a wide variety of configuration hypothesis was established for testing and evaluation. The different available solutions included having storage hubs with or without stock, keeping stock in one country or choosing a hybrid approach and outsourcing or internalizing the warehousing process. 3. Validating the developed models and retrieving final indicators and results The ensuing phase consisted in developing a holistic optimization and simulation model to test the different scenarios regarding warehouse future locations, logistic flows and warehouse-to-store allocations, aiming to minimize overall costs while ensuring a high service level. In this step, and for each scenario, all previously defined indicators and costs were predicted. 4. Gathering relevant stakeholders and making the final decisions Towards the end, all involved stakeholders were gathered, the different options were discussed and compared, and a final decision was made. The most relevant decisions of the project involved the centralization of part of the distribution process in the Portuguese warehouse, the relocation of the Spanish warehouse to a more cost-effective area, and a significant boost of the supplier-store direct deliveries. In this step, it was critical to ensure alignment between stakeholders and to reach a conclusion in which everyone agreed on, even though it wouldn’t necessarily bring gains to all parties involved. It is crucial to bear in mind that to maximize the overall benefits, some departments and stakeholders may temporarily have losses.
Most companies have now realized the importance of seeking and maintaining data, not only from their own operations but also from suppliers and customers. However, not many of them fully understand how to use that data to boost their supply chain competitiveness, by either decreasing costs or increasing customer satisfaction. As analytics becomes a popular subject, managers are broadening its range of applications within the existing corporate departments. Nevertheless, many supply chain decisions are still uniquely based on qualitative insights or competitor moves. To run successful supply chain analytics projects, there are a few guidelines that should be followed: Mapping and exploring alternatives for further testing and evaluating Understanding the operation and mapping indicators before using analytics Modelling the as-is operation and fully grasping its intrinsic characteristics Validating the developed models and retrieving final indicators and results Gathering relevant stakeholders and making the final decisions Transforming supply chain decision-making As stated above, when performing analytical projects, it is common to make the mistake of planning modifications to organizations without considering a holistic view of its impacts. Having this thought in mind, a project was developed together with a food retailer aiming to simulate its supply chain from end-to-end, encompassing operations from the main warehouses all the way to the store shelves. The ultimate goal of this project was not to find a better supply chain configuration on itself but to give the company a tool to continuously and independently do so in the future. Therefore, the main project steps were revisited. Mapping and exploring alternatives for further testing and evaluating The first phase of the project consisted on exploring and listing all the use cases to be covered, detailing all possible strategic decisions, tactical levers, external factors and operational variables. The defined scenarios range from high-level decisions, such as network configuration or modifying the promotional activity’s magnitude, to operational factors, such as workers’ productivity improvement. Understanding the operation and mapping indicators before using analytics In the following step, a significant mapping effort was made to get a thorough understanding of all supply chain processes – cargo unloading, warehouse picking, store delivery, shelf replenishment, etc. –, levers – transportation delivery windows, workers and machines productivity, etc. – and variables, considering both material and informational flows. Modelling the as-is operation and fully grasping its intrinsic characteristics The subsequent phase consisted of modelling and developing the simulator while, in parallel, a vast range of scenarios were outlined, in order to further test some of the retailer’s initiatives and action streams. This phase was particularly lengthy as model validation is a highly demanding process, implying an iterative process of both comparing model results with the reality of the operations and making further adjustments for model enhancement. Validating the developed models and retrieving final indicators and results For each simulated scenario, several comparisons can be made with the retailer current situation, ranging from the value of store and warehouse stocks, the costs of spoilage and shrinkage to the transportation and workforce costs. The developed project and resulting tool enabled a more aggregate and holistic view of the value chain, allowing the testing of distinct scenarios and hypothesis, thus empowering a more conscious decision making and sustaining the retailer’s competitive advantage. Gathering relevant stakeholders and making the final decisions The delivery of a tool enables the continuity of the project and leverages the developed work since it allows for the teams to carry on exploring new hypothesis in an independent manner. In this way, each team or department will be able to find, filter and decide on its own improvement opportunities, leveraging the decision-making process all through the supply chain. In summary, this project brought together business expertise and analytical knowledge, to give managers a state-of-the-art tool that allows them to automatically and continuously assess different what-if scenarios and choose the most beneficial for the company's future.