Stock Keeping Units Research Articles

Supply Chain Demand Forecast Based on SSA-XGBoost Model

Supply chain management usually faces problems such as high empty rate of transportation, unreasonable inventory management, and large material consumption caused by inaccurate market demand forecasts. To solve these problems, using artificial intelligence and big data technology to achieve market demand forecasting and intelligent decision-making is becoming a strategic technology trend of supply chain management in the future. Firstly, this paper makes a visual analysis of the historical data of the Stock Keeping Unit (SKU); Then, the characteristic factors affecting the future demand are constructed from the storage level, product level, historical usage of SKU, etc; Finally, a supply chain demand forecasting algorithm based on SSA-XGBoost model has proposed around three aspects of feature engineering, parameter optimization and model integration, and is compared with other machine learning models. The experiment shows that the forecasting result of SSA-XGBoost forecasting model is highly consistent with the actual value, so it is of practical significance to adopt this forecasting model to solve the supply chain demand forecasting problem.

A Decision Support System for Re-Slotting a Case Pick Distribution Center

This research presents a class-based re-slotting procedure that re-assigns stock-keeping units (SKUs) in a warehouse secondary storage area to new locations based on the SKU activity. The general notion of re-slotting is to place high activity SKUs in premium locations near the pickup/drop-off points of shipping and receiving. This simulation study examines the performance of an ABC class-based rule for guiding SKU swaps between the front and rear sections using case pick data provided by a large regional pharmaceutical distributor. The simulation results show this class-based case pick re-slotting heuristic provides significant savings in picker travel.

Golden zone storage assignment and picking performance: An empirical analysis of manual picker-to-parts OP systems in grocery retailing

Manual picker-to-parts order picking (OP) systems are usually laborious and cost-intensive. Nevertheless, they are still predominant in warehouse operations of traditional brick-and-mortar grocery retailers. Storage assignment strategies are highly relevant for the outcomes of OP systems. Aspiring to improve picking performance and worker health, the concept of golden zone storage assignment is frequently applied in practice. This assignment strategy includes the placing of high-demand stock-keeping units at the height between the picker's waist and shoulders. Although simulation approaches indicate a positive impact of golden zone storage on the outcomes of OP systems, empirical evidence remains scarce. Therefore, we test this assumption with an empirical setting, including 1,570,049 picks performed by 156 order pickers in one warehouse operated by a large German brick-and-mortar grocery retailer. We apply a log-logistic accelerated failure time model and incorporate human factors through a mixed-effects model allowing one regression line per order picker. Our findings offer counter-intuitive insights because we find that golden zone storage assignment decelerates the picking process by 2.55% compared to ground-level picks. From a managerial viewpoint, our findings highlight the conflicting relation of economic- and ergonomic-driven evaluations for labor-intensive OP systems. Results call for a more differentiated typology of golden zone storage, including the depth of rack picking locations.

Inventory management of vertically differentiated perishable products with stock-out based substitution

The need for optimal inventory control strategies for perishable items is of the utmost importance to reduce the large share of food products that expire before consumption and to achieve responsible food stocking policies. Our study allows for a multi-item setting with substitution between similar goods, deterministic deterioration, delivery lead times and seasonality. Namely, we model demand by a linear discrete choice model to represent a vertical differentiation between products. The verticality assumption is further applied in a novel way within product categories. Specifically, the same product typology is vertically decomposed according to the age of the single stock-keeping unit in a quality-based manner. We compare two different policies to select the daily size of the orders for each product. On the one hand, we apply one of the most classical approaches in inventory management, relying on the Order-Up-To policy, modified to deal with the seasonality. On the other hand, we operate a state-of-the-art actor-critic technique: Soft Actor-Critic (SAC). Although similar in terms of performance, the two policies show diverse replenishment patterns, handling products differently.

Automation of Best-Fit Model Selection using a Bag of Machine Learning Libraries for Sales Forecasting

Sales forecasting became crucial for industries in past decades with rapid globalization, widespread adoption of information technology towards e-business, understanding market fluctuations, meeting business plans, and avoiding loss of sales. This research precisely predicts the automotive industry sales using a bag of multiple machine learning and time series algorithms coupled with historical sales and auxiliary features. Three-year historical sales data (from 2017 till 2020) were used for the model building or training, and one-year (2020-2021) predictions were computed for 900 unique SKU's (stock-keeping units). In the present study, the SKU is a combination of sales office, core business field, and material customer group. Various data cleaning and exploratory data analysis algorithms were implemented over raw datasets before use for modeling. Mean absolute percentage error (mape) were estimated for individual predictions from time series and machine learning models. The best model was selected for unique SKU's as per the most negligible mape value.

The Price Effect of Drug Price Ceilings: Intended and Unintended Consequences

Policy makers in many developing countries use maximum price or markup policies to control pharmaceutical costs, which represent 20%–60% of their overall healthcare expenditure. We study the price effect of price ceiling policies by exploiting a major policy shift in China: the elimination of longstanding ceilings on retail drug prices. We collect weekly price and characteristics data on more than 4,500 drug stock keeping units (SKUs) from a leading pharmacy chain. By comparing the rate of discontinuous price jumps across drugs with and without price ceilings during the years before and after the policy change, we find that while price ceilings are effective in containing the prices of some drugs, they can lead to higher prices for others, particularly if the ceilings are set at the national level irrespective of local economic conditions. About 5% of nationally controlled drugs (or more than 125 drugs) had inflated prices because of price ceilings, with an average price inflation of 10%. We attribute this perverse price effect to focal point pricing and asymmetric information about production costs. Further supporting this view, we find the perverse price effect most prominent in lower-income regions where the centrally set price ceilings are arbitrarily high considering their poorer economic conditions. Moreover, drugs with highly concentrated production and less elastic demand face heightened risks of inflated prices under price ceilings. Finally, based on a sample of drugs with available price ceiling data, we find that drugs with manufacturer-specific ceilings are 100% more likely to be priced at or near their ceilings and 70% more likely to experience price drops once the ceilings are removed compared with other drugs with regular ceilings. Overall, this paper documents the unintended perverse effect of price ceilings in pharmaceutical markets and sheds lights on the ongoing debate of drug price regulation. This paper was accepted by Stefan Scholtes, healthcare management.

Product variety in local grocery stores: Differential effects on stock‐keeping unit level sales

AbstractWe empirically examine the effect of product variety on sales at the stock‐keeping unit (SKU) level. Prior empirical operations management studies generally treat product variety as a single‐dimensional measure (e.g., the number of SKUs carried by the store, distribution center, or firm) and study its impact on aggregate sales (e.g., sales at the store, distribution center, or firm). As such, SKUs from different brands are assumed to contribute equally to total sales, which is not always the case. To address this, we need to conceptualize product variety within brands and investigate its impact on sales at a more granular level. In this study, we propose and theoretically articulate two product variety dimensions (within‐brand variety [WV] and average WV of other brands [AWVO]) to investigate the mechanisms that connect product variety to SKU‐level sales. Using a detailed transaction dataset from a local grocery retailer, we first show that product variety conceptualized within brands (i.e., the mean of WVs [MWV]) does have an impact on aggregate subcategory sales, as the literature suggests. At the SKU level, we find WV increases same‐brand SKU sales, while AWVO reduces it. Our work provides an alternative explanation to the inverted U‐shaped relationship between product variety and aggregate sales observed in the literature: Increasing variety only increases sales of SKUs belonging to the brand whose variety has increased, while it decreases sales of the other SKUs. We also find that variety effects vary significantly across different profitability levels and product categories. Product variety appears to have the strongest positive effect on high‐margin products. The impact of product variety on specific product category sales varies from as high as 11% to negligible. We further discuss the practical value of our findings and show the robustness of our results through multiple tests while addressing endogeneity using the control function approach.

Relocation and storage assignment strategy evaluation in a multiple-deep tier captive automated vehicle storage and retrieval system with undetermined retrieval sequence

This paper studies multiple-deep automated vehicle storage and retrieval systems (AVS/RSs) known for their high throughput performance and flexibility. Compared to a single-deep system, multiple-deep AVS/RS has a better space area utilisation. However, a relocation cycle occurs, reducing the throughput performance whenever another stock keeping unit (SKU) blocks a retrieving SKU. The SKU retrieval sequence is undetermined, meaning that the arrangement is unknown, and all SKUs have an equal probability of retrieval. In addition to the shuttle carrier, a satellite vehicle is attached to the shuttle carrier and is used to access storage locations in multiple depths. A discrete event simulation of multiple-deep AVS/RS with a tier captive shuttle carrier was developed. We focused on the dual-command cycle time assessment of nine different storage and relocation assignment strategy combinations in the simulation model. The results of a simulation study for (i) random, (ii) depth-first and (iii) nearest neighbour storage and relocation assignment strategy combinations are examined and benchmarked for five different AVS/RS case study configurations with the same number of storage locations. The results display that the fivefold- and sixfold-deep AVS/RSs outperform systems with fewer depths by utilising depth-first storage and nearest neighbour relocation assignment strategies.

Design framework for a lean warehouse – a case study-based approach

PurposeThis purpose of this study is to develop a simple framework for designing a warehouse incorporating lean principles. Multiple objectives like resource planning, material handling, storage, inventory management, including internal and external logistics, are considered.Design/methodology/approachA design procedure to incorporate lean principles for designing a warehouse for a complex multi-model production line has been proposed. The preferred standards and factors affecting warehouse design, the inputs and outputs of process flow characteristics, are incorporated into the design. Current and future state value stream mappings are drawn to bring out the challenges in the value flow.FindingsThe framework for designing a lean warehouse have been implemented and validated in a heavy machinery manufacturer. This framework will ease the work of the future lean-based warehouse designers to apply simple step-by-step processes to achieve the goal with the nearest accuracy. The steps followed can be summarized as defining the lean processes, making the lean process as the design base, collecting inputs like stock-keeping unit master, inventory and space details, and building the lean warehouse design with the step-by-step processes.Practical implicationsPractical tips on warehouse design have been explained focusing on the part volume, quantity handled, inventory and throughput. This will assist the practitioners in designing a lean warehouse and leading to an improved operational performance.Originality/valueA simplified design procedure for designing a lean warehouse, along with a real-time case study has been enumerated in detail. Effective use of space and resources with lean tools and techniques lead to better storage and picking efficiency resulting in an overall reduction in cost.

Classification-based model selection in retail demand forecasting

Retailers supply a wide range of stock keeping units (SKUs), which may differ for example in terms of demand quantity, demand frequency, demand regularity, and demand variation. Given this diversity in demand patterns, it is unlikely that any single model for demand forecasting can yield the highest forecasting accuracy across all SKUs. To save costs through improved forecasting, there is thus a need to match any given demand pattern to its most appropriate prediction model. To this end, we propose an automated model selection framework for retail demand forecasting. Specifically, we consider model selection as a classification problem, where classes correspond to the different models available for forecasting. We first build labeled training data based on the models’ performances in previous demand periods with similar demand characteristics. For future data, we then automatically select the most promising model via classification based on the labeled training data. The performance is measured by economic profitability, taking into account asymmetric shortage and inventory costs. In an exploratory case study using data from an e-grocery retailer, we compare our approach to established benchmarks. We find promising results, but also that no single approach clearly outperforms its competitors, underlying the need for case-specific solutions.

Evaluating the impact of single exit pricing (SEP) on medicine product withdrawal from the private healthcare market in South Africa.

The introduction of medicine pricing policies in South Africa (SA) in the form of single exit pricing (SEP) provided a mechanism to improve medicine price transparency and reduce the medicine price and inflation. However, regulation of medicine prices may have further unforeseen effects on the availability of medicine. This research presents the impact of SEP on discontinuation of medicine products on the private healthcare market in SA. To evaluate the impact of SEP legislation on the availability of medicines in the SA private health sector in terms of withdrawal of medicines from the market. A descriptive, quantitative analysis of all registered medicines on the SA market by stock-keeping units (SKUs) was done to establish medicine products that were withdrawn from the market by SKUs during a 14-year period (2001 - 2014). A total of 152 manufacturers discontinued 3 691 SKUs between 2001 and 2014. The mean number of discontinuations per generic manufacturer was 22.34 (standard deviation (SD) 58.11), while innovator manufacturers discontinued a mean of 27.61 (41.89). The largest number of SKUs were commercially withdrawn in 2002 (n=603), followed by discontinuations in 2003 (n=463) and 2004 (n=407). There was a negative correlation between number of discontinued SKUs per year and SEP increase (Pearson's correlation coefficient r ‒0.414; p=0.14). The results showed that SEP and a transparent pricing policy may have had an impact on SKU withdrawal from the market prior to SEP implementation. The result of reduced product availability on the market and its impact on the cost and quality of healthcare to the patient need to be regularly monitored and evaluated to ascertain if direct price regulations achieve the intended outcomes. Other intended or unintended effects on pharmaceutical market dynamics should also be evaluated.

Modeling Single-Picker Routing Problems in Classical and Modern Warehouses

The standard single-picker routing problem (SPRP) seeks the cost-minimal tour to collect a set of given articles in a rectangular single-block warehouse with parallel picking aisles and a dedicated storage policy, that is, each stock-keeping unit is only available from one storage location in the warehouse. We present a compact formulation that forgoes classical subtour elimination constraints by directly exploiting two of the properties of an optimal picking tour used in the dynamic programming algorithm published in the seminal paper of Ratliff and Rosenthal. We extend the formulation to three important settings prevalent in modern e-commerce warehouses: scattered storage, decoupling of picker and cart, and multiple end depots. In numerical studies, our formulation outperforms existing standard SPRP formulations from the literature and proves able to solve large instances within short runtimes. Realistically sized instances of the three problem extensions can also be solved with low computational effort. For scattered storage, we note a rough tendency that runtimes increase with longer pick lists or a higher degree of duplication. In addition, we find that decoupling of picker and cart can lead to substantial cost savings depending on the speed and capacity of the picker when traveling alone, whereas additional end depots have rather limited benefits in a single-block warehouse. Summary of Contribution: Efficiently routing order pickers is of great practical interest because picking costs make up a substantial part of operational warehouse costs. For the prevalent case of a rectangular warehouse with parallel picking aisles, we present a highly effective modeling approach that covers—in addition to the standard setting—several important storage and order-picking strategies employed in modern e-commerce warehouses: scattered storage, decoupling of picker and cart, and multiple end depots. In this way, we provide practitioners as well as scientists with an easy and quick way of implementing a high-quality solution approach for routing pickers in the described settings. In addition, we shed some light on the cost benefits of the different storage and picking strategies in numerical experiments.

Data Set: 187 Weeks of Customer Forecasts and Orders for Microprocessors from Intel Corporation

Problem definition: This data set contains 187 consecutive weeks of Intel microprocessor demand information for all five distribution centers in one of its five sales geographies. For every stock keeping unit (SKU) at every location, the weekly forecasted demand and actual customer orders are provided as well as the SKU’s average selling price category. These data are provided by week and by distribution center, producing 26,114 records in total. Academic/practical relevance: The 86 SKUs in the data set span five product generations. It provides years of product evolution across generations and price points. Methodology: As a data set paper, its purpose is to provide interesting and rich real-world data for researchers developing forecasting, inventory, pricing, and product assortment models. Results: The data set demonstrates the presence of significant forecast bias, heterogeneity of forecast errors between distribution centers, generational differences, product life cycles, and pricing dynamics. Managerial implications: This data set provides access to a rich pricing and sales setting from a major corporation that has not been made available before.

Designing efficient assortments: A branch-and-bound method to optimise volume and satisfaction

Category assortments are a key driver of retailer success. When a customer’s preferred product is unavailable, retailers must provide viable alternatives or risk losing the customer. This paper presents an analytical method that uses a survey-based approach where consumers make hypothetical (albeit not conjoint) choices. The objective is to identify how delisting a stock-keeping unit (SKU) will affect overall sales; the extent to which customers will abandon the store; and overall satisfaction with the store. Such knowledge will help retailers make the best delisting decisions and help both retailer and brands with their category management negotiations. It can also identify possible gaps for the brand, such as which alternatives in the competition’s portfolio have low substitutability and hence should perhaps be added to the brand’s product portfolio. This approach also allows brands to test new concepts and evaluate how adding a new (concept) SKU will affect sales of other SKUs and overall sales and profitability. The data are then modelled using a branch-and-bound algorithm. This approach can easily be implemented at the store level and allows for the optimisation of multiple objectives.

A Heuristic Storage Location Assignment Based on Frequent Itemset Classes to Improve Order Picking Operations

Most large distribution centers’ order picking processes are highly labor-intensive. Increasing the efficiency of order picking allows these facilities to move higher volumes of products. The application of data mining in distribution centers has the capability of generating efficiency improvements, mainly if these techniques are used to analyze the large amount of data generated by orders received by distribution centers and determine correlations in ordering patterns. This paper proposes a heuristic method to optimize the order picking distance based on frequent itemset grouping and nonuniform product weights. The proposed heuristic uses association rule mining (ARM) to create families of products based on the similarities between the stock keeping units (SKUs). SKUs with higher similarities are located near the rest of the members of the family. This heuristic is applied to a numerical case using data obtained from a real distribution center in the food retail industry. The experiment results show that data mining-driven developed layouts can reduce the traveling distance required to pick orders.

Order Picking and E-Commerce: Introducing Non-Parametric Efficiency Measurement for Sustainable Retail Logistics

Order picking is a crucial but labor- and cost-intensive activity in the retail logistics and e-commerce domain. Comprehensive changes are implemented in this field due to new technologies like AI and automation. Nevertheless, human worker’s activities will be required for quite some time in the future. This fosters the necessity of evaluating manual picker-to-part operations. We apply the non-parametric Data Envelopment Analysis (DEA) to evaluate the efficiency of n = 23 order pickers processing 6109 batches with 865,410 stock keeping units (SKUs). We use distance per location, picks per location, as well as volume per SKU as inputs and picks per hour as output. As the convexity axiom of standard DEA models cannot be fully satisfied when using ratio measures with different denominators, we apply the Free Disposal Hull (FDH) approach that does not assume convexity. Validating the efficiency scores with the company’s efficiency assessment, operationalized by premium payments shows a 93% goodness=of-fit for the proposed model. The formulated non-parametric approach and its empirical application are promising ways forward in implementing empirical efficiency measurements for order picking operations within e-commerce operations.

High-frequency forecasting for grocery point-of-sales: intervention in practice and theoretical implications for operational design

Food waste in grocery supply chains may exceed one third of the total volume, depending on the category. To address this problem effectively, grocery retailers are introducing automated systems for more efficient store replenishment and dynamic pricing. The stock keeping unit (SKU) and store level forecast is pivotal in these operations management solutions, but operationally challenging. Large grocery retailers have millions of SKU-store combinations that depending on the operational application would need to be forecasted on a weekly, daily, hourly, or even 15-min frequency. However, in grocery it is challenging to account for demand variation at high frequencies without introducing manual decisions into the process of forecast model configuration. To investigate the limits of current practice and explore opportunities of technology-enabled change, we explore how an advanced forecasting method for electricity demand, called TBATS, can automate daily forecasting for grocery store replenishment. Adopting an interventionist approach, we explore the implications for the design of the operational process in the operational setting provided by the case company. We find that TBATS can produce high frequency base forecasts for the SKU-store level accurately for a period exceeding 3 months. This finding points to the opportunity of shifting operational focus from recalculating forecasts to monitoring forecast errors. Introducing variable, even indefinite re-training frequencies for forecasting models is a significant change of the forecasting process for situations where monitoring requires less computation than retraining, potentially reducing the time and cost associated with increasing the forecast frequency.

How common is new product failure and when does it vary?

This study aims to explore how common new product failure is in consumer packaged goods (CPG) categories and investigate the conditions in which the new product failure rate varies. This study analyzes 83,719 new stock-keeping units (SKUs), which were introduced over 8 years (2002–2009) across 31 CPG categories in the USA. Failure is the permanent cessation of sales, which is measured in consumer panel data. We find that one in four (25%) new SKUs are no longer bought 1 year later—a rate that increases to approximately 40% 2 years post-launch. New SKU failure was more likely for new launches that were introduced into higher revenue categories and by smaller share parent brands. Our findings can be used as a resource to aid marketing practitioners’ understanding regarding how common failure is for new CPGs and when there is greater associated risk.

Shrinking the Upper Confidence Bound: A Dynamic Product Selection Problem for Urban Warehouses

The recent rising popularity of ultrafast delivery services on retail platforms fuels the increasing use of urban warehouses, whose proximity to customers makes fast deliveries viable. The space limit in urban warehouses poses a problem for such online retailers: the number of stock keeping units (SKUs) they carry is no longer “the more, the better,” yet it can still be significantly large, reaching hundreds or thousands in a product category. In this paper, we study algorithms for dynamically selecting a large number of products (i.e., SKUs) with top customer purchase probabilities on the fly, from an ocean of potential products to offer on retailers’ ultrafast delivery platforms. We distill the product selection problem into a semibandit model with linear generalization. There are in total N arms corresponding to N products, each with a feature vector of dimension d. The player pulls K arms in each period and observes the bandit feedback from each of the pulled arms. We focus on the setting where K is much greater than the number of total time periods T or the dimension of product features d. We first analyze a standard Upper Confidence Bound (UCB) algorithm and show its regret bound can be expressed as the sum of a T-independent part and a T-dependent part, which we refer to as “fixed cost” and “variable cost,” respectively. To reduce the fixed cost for large K values, we propose a novel online learning algorithm, which iteratively shrinks the upper confidence bounds within each period, and show its fixed cost is reduced by a factor of d. Moreover, we test the algorithms on an industrial data set from Alibaba Group. Experimental results show that our new algorithm reduces the total regret of the standard UCB algorithm by at least 10%. This paper was accepted by J. George Shanthikumar, big data analytics.

An Exponential Factorization Machine with Percentage Error Minimization to Retail Sales Forecasting

This article proposes a new approach to sales forecasting for new products (stock-keeping units [SKUs]) with long lead time but short product life cycle. These SKUs are usually sold for one season only, without any replenishments. An exponential factorization machine (EFM) sales forecast model is developed to solve this problem which not only takes into account SKU attributes, but also pairwise interactions. The EFM model is significantly different from the original Factorization Machines (FM) from two fold: (1) the attribute-level formulation for explanatory/input variables; and (2) exponential formulation for the positive response/output/target variable. The attribute-level formation excludes infeasible intra-attribute interactions and results in more efficient feature engineering comparing with the conventional one-hot encoding, while the exponential formulation is demonstrated more effective than the log-transformation for the positive but not skewed distributed responses. In order to estimate the parameters, percentage error squares (PES) and error squares (ES) are minimized by a proposed adaptive batch gradient descent method over the training set. To overcome the over-fitting problem, a greedy forward stepwise feature selection method is proposed to select the most useful attributes and interactions. Real-world data provided by a footwear retailer in Singapore are used for testing the proposed approach. The forecasting performance in terms of both mean absolute percentage error (MAPE) and mean absolute error (MAE) compares favorably with not only off-the-shelf models but also results reported by extant sales and demand forecasting studies. The effectiveness of the proposed approach is also demonstrated by two external public datasets. Moreover, we prove the theoretical relationships between PES and ES minimization, and present an important property of the PES minimization for regression models; that it trains models to underestimate data. This property fits the situation of sales forecasting where unit-holding cost is much greater than the unit-shortage cost (e.g., perishable products).