SUPPLY CHAIN ​​MANAGEMENT SOFTWARE – PART 1

The end of the last decade was marked by the vertiginous growth of ERP system deployments (SAP/R3, Oracle, BAAN, etc). This movement was driven both by fears regarding the millennium bug and by the adoption by many companies of a vision of their business through processes.

After the period of ERPs, we are now experiencing a new wave of deployment of information technology packages: the Supply Chain Management Systems (SCM). According to the Mckinsey consultancy, between 1999 and 2002, more than US$ 15 billion in licenses for these types of systems were sold, not including expenses related to implementation processes and maintenance costs. Despite the large investment already made worldwide, this movement is still starting here in Brazil.

Faced with the relevance of this movement, the ILOS Institute carried out a survey to map and analyze the process of implementing these systems by Brazilian companies. This research will be presented in two articles. This article will deal with the definition of supply chain management software, its differences in relation to ERPs, as well as a presentation of the available modules. The second article will present the results obtained through interviews with companies that have already implemented this type of system.

DEFINITION

Before discussing SCM software in detail, a brief overview of ERP systems is relevant in order to clarify the differences between the two.

ERPS are transactional systems that tend to focus on the operational level and do not have much analytical capacity to help with planning and strategic decisions. They are great at letting managers know what is going on, but not at letting them know what should be going on. ERP systems can inform the current stock level of a product in a given warehouse, for example, but they are weak in determining how much stock is needed to reach a given service level.

The implementation of ERPs enables the integration of the entire company, making it more efficient. However, they do not help resolve the fundamental questions of what should be done, where, when and by whom. This is the role played by planners with the help of decision support tools, analytical systems.

In contrast to ERPs, analytics tools are not transactional systems when it comes to storing data and processing day-to-day tasks. Rather, through sophisticated algorithms and scenario analysis, they enable managers to make operations more efficient as well as better understand the impact of their strategic decisions. For example, an ERP system can provide the history of demand, inventory levels and lead time, and the analytics system can determine what the inventory level should be in order to maximize the profitability of the operation.

These analytical applications are based on sophisticated algorithms including linear programming, mixed integer programming, genetic algorithms, theory of constraints, and various types of heuristics. These algorithms are most often proprietary to the software vendor, and large investments and R&D are required to arrive at them. Due to this level of sophistication, this technology is relatively difficult to develop if the company does not have much experience in the area.

Its use not only makes it possible to make better decisions, but also allows them to be made faster than before. As an example, companies traditionally measure their production planning cycles in terms of weeks or even months, due to computational constraints and lack of information. With the help of analytical tools, the planning cycle can be planned on a daily basis.

These advantages arising from its use are a consequence of the three main characteristics of this type of system:

1. Possibility of integral planning of the entire supply chain, at least from the supplier to the customer of a single company, or even of a broader network of companies;
2. Real optimization through the correct definition of alternatives, objectives and constraints for the various planning problems and based on the use of optimizing planning methods or exact heuristics;
3. Use of a hierarchical planning system, the only structure that allows the combination of the two preceding properties: optimized planning of the entire chain is not feasible in the form of a monolithic system, which performs all planning tasks simultaneously – which would be impractical – nor by executing these tasks sequentially – which would make optimization impossible. Hierarchical planning is a balance between practicality and consideration of the interdependence between planning tasks.

Although they are often seen as competitors, analytical systems and ERPs have a strong interdependence. The full value of an ERP system cannot be achieved without the problem solving capabilities of analytical systems. Likewise, for analytical systems to be productive requires the availability of accurate data from various functions of the organization. One of the best ways to obtain this data is through an ERP system.

Classification Structure of Planning Systems

The logistics management of a company involves a wide variety of decisions, associated with different activities – transport, production, inventory, etc. In order to cover all types of decisions, SCM software has some modules, which are generally related to the type of decision to be taken, and to the logistical activities.

A good way to visualize the different existing modules is through the development of a structure proposed by Robert Anthony, a Harvard professor, during the 60s.

According to Anthony, there are no plans for eternity. That is, the validity of a planning is restricted to a predefined planning horizon. Every time this time horizon is reached, a new planning must be carried out according to the operating conditions at the time. According to the size of the planning horizon and the importance of the decision to be taken, planning tasks can be classified into three different planning levels:

• Long-term planning or strategic planning: Decisions that will define how the company will act in the horizon of a few years. This planning will structure the conditions under which the next types of decisions will be made. In the case of logistical planning, definitions of where to locate warehouses and transport terminals, the degree of automation of each installation and the sources of supply are included.
• Medium-term planning or tactical planning: Within the scope delimited by strategic decisions, tactical decisions determine, in general, how the operation will take place. In other words, it answers the following question: “Given customer demands and available resources, what can be done to maximize the company's profit?”. It includes sales and production planning decisions, and definitions regarding the characteristics of the transport fleet.
• Short-term planning or operational decisions: Once tactical planning is defined, operational decisions specify all activities for the immediate execution and control of the operation. These decisions are those that require the greatest degree of detail and accuracy of information. The planning horizon can usually be measured in days.

These three decision levels are valid for decisions referring to any activity.

Figure 1 presents a matrix in which the decision-making levels are crossed with the main processes of a company – purchasing, production, distribution and sales.

 

In this way, all planning decisions relating to each of the processes can be classified as strategic, tactical or operational. The positioning of decisions in the matrix serves as a basis for the development of decision support systems.

Available Features

Although there are particularities between the SCM systems available on the market, it is possible to generalize the modules offered. Figure 2 shows a generic structure of an SCM system that covers all planning activities. The possible differences between the architecture shown in the figure and the one existing in commercialized systems will normally be associated with the availability of modules that encompass more than one generic module. However, from the functionality point of view, these do not present significant differences.

Thus, figure 2 represents a very realistic and, more importantly, very didactic structure for analyzing the functionalities available in SCM systems. These functionalities will be analyzed below.

Through an analysis of this figure, it is noted that while some modules are focused on only one decision-making level and a process (for example a TMS), others cover more than one decision-making level (Demand Planning) or more than one process (Master Planning). Another relevant observation is that the operational modules, which work with decisions with a high degree of detail, have a very restricted scope, and that, as decisions become more strategic, their respective modules gain greater scope, such as For example, the Network Planning module, which covers all processes in a simplified way.

Strategic Network Planning

Typically, the horizon for strategic planning of the logistics network can be considered to be two years onwards, and its decisions involve the definition of customer zones, the opening or closing of factories and distribution centers, as well as their required capacities. . The objectives of the network models are, for the most part, of a financial and aggregate type, such as, for example, profit maximization or cost minimization, always being limited by service level restrictions.

The use of these systems goes through the modeling of the network itself. This modeling covers fixed and variable operating costs, existing facilities (factories, distribution centers) and geographic segmentation of demand, among other aspects.

The responses of this type of model are generally associated with the facilities that make up the logistics network, such as their capacities, geographic allocations and orientation towards certain customers or products. Capacities and allocations defined by network planning then become constraints for the master production planning process.

demand planning

Demand planning applications help in the preparation of sales forecasts, through the use of appropriate analytical tools. These systems use as input historical sales data and any existing information that may be related to future demand, such as contracts already signed with customers or inflation rate projections.

The forecast can be calculated through both methods based on historical series and causal methods. As for methods based on historical series, there are many, each with specific characteristics. Through these it is possible to identify trends of growth or decrease in sales, calculation of seasonality, calculate daily or annual forecasts. Faced with the wide availability of different models, most systems have methodologies and algorithms that identify which method provides the best forecast for a given series of sales.

With regard to causal models, they allow sales behavior to be predicted based on variables other than historical sales. For example, estimate new tire sales as a function of car sales. These models can also be used to assess the impact of specific events, such as the impact of a new promotional campaign on sales.

The systems allow forecasts to be made and monitored through three basic dimensions, with different degrees of aggregation.

• Product dimensions: product, group, family, line
• Geographical dimensions: customer, sales region, area of ​​operation of distribution centers, national sales
• Time dimensions: day, month, year or any specific horizon required due to seasonal issues

Master Planning

The main purpose of this module is to synchronize the flow of materials along the entire chain. This supports mid-term decisions regarding production capacity, transportation availability, supply planning and inventory policies. As a consequence of this synchronization, it is possible to obtain a reduction in stock levels, mainly due to the elimination of redundant safety stocks between the activities present in the flow of materials and arising from a non-integrated planning system.

This synchronization of the flow of materials is achieved through the definition, during planning, of all the capabilities of the entities (factories, distribution centers, fleet of vehicles) that make up the supply chain in question. That is, the master production planning informs not only how much will be produced in each location, but also the operational needs for this plan to be possible.

However, this type of planning optimization is not possible, or at least not reasonable, based on information with a high degree of detail. A degree of aggregation of products is necessary as well as a simplification of capacities, such as monthly production capacity. This not only reduces the amount of necessary data, but also reduces the uncertainties of medium-term information – for example demand forecasts – as well as the complexity of the models.

Transportation Planning

This module is associated with tactical decisions regarding the planning of the transport operation. In this way, it supports, through the definition of rules and assumptions, the generation of scripts that will be used in transport scheduling. These routes are constructed observing loading rules and load consolidation opportunities, among other aspects.

In parallel to this planning, the necessary fleet dimensioning can be carried out. This sizing indicates not only the number of vehicles needed, but also the profile of this fleet, in terms of different types of vehicles or even modes, as well as their distribution on defined routes.

These planning alternatives can also be used to manage the inbound transport of companies that purchase inputs through the FOB modality.

This becomes particularly interesting when planning inbound and outbound flows are associated. When this occurs, it is possible to carry out analyzes in search of synergies between the two flows, generally associated with the use of return freight.

Inventory Planning

The inventory planning modules are responsible for defining and planning the stock policies to be used, and not for the daily inventory control, which is an ERP function.

These systems help not only in deciding which policy to adopt, but also play a fundamental role in calculating the parameters of the chosen policies. For this purpose, information is used about inventory maintenance and transportation costs, the necessary service levels, as well as operational parameters, such as supply and manufacturing times and demand projections. With all this information in hand, the algorithms are able to determine policies that obtain the best balance between the cost of inventory and the cost of lost sales due to lack of product.

Among the calculated parameters, what confers greater differentiation to these systems is the safety stock. While in transactional systems, the safety stock is just a field to be filled in by the user, in supply chain software the calculation is carried out considering the operational parameters already mentioned above, the desired service level, and the uncertainties associated with the material flow (sales forecast accuracy, supply reliability).

production scheduling

Given a master production plan, it must generate detailed production plans for each production center. This is the function of the scheduling module, that is, to generate detailed production schedules, in relatively small time intervals. The production schedule indicates, for each order within the planning interval, its start and end times, as well as the resources required for its processing. In this way, the production schedule determines the order in which all orders will be processed. It is exactly at this point that SCM software adds the most value.

The production schedule carried out by these systems is based on production models. The models are structured according to the characteristics of the production system in question – matrix of set-ups, manufacturing times, priority rules, batch sizes and costs involved – as well as information about what must be produced – quantity of each product and deadline for delivery.

Once the production model is specified, the systems seek the best production programming, through optimizing algorithms, depending on some objective. This objective is usually expressed in the minimization or maximization of some aspect of production such as: number of set-ups, total backorders and variable production costs.

The systems treated below are essentially operational, not having great differentiation from the analytical point of view. They can be found in many transactional systems, or be provided by small companies that sell simple systems. However, as they are found in the diagram of available applications, and as they act in logistical functions, they will be addressed quickly.

Transportation Management System – TMS

The main activities of a TMS can be divided into three groups: monitoring and control, execution and freight audit. These groups are discussed briefly below.

Monitoring costs and services through information from the operation itself. In this way, the most suitable indicators for each operation can be measured, such as: performance of carriers, modes of transport, use of premium freight, return freight, delivery performance, breakdowns, etc.

The functionalities associated with the execution consist of determining the routes and modes to be used, sequencing the vehicle stops and the estimated time of each one of them, preparing the necessary documents for dispatching the vehicles and checking their availability.

Finally, with regard to freight auditing, these systems maintain a database of freight rates used to remunerate the service provided and for the auditing process. The systems are able to compare the amount charged by the transport service provider against what was calculated and point out any differences.

Warehouse Management System – WMS

These systems are responsible for managing the day-to-day operation of a warehouse. Despite having some algorithms, their use is restricted to fully operational decisions such as:

• Definition of collection routes, with the aim of minimizing the average distance traveled in order picking.
• Definition of product addresses, based on logic that uses criteria that once again seek to minimize the average movement distance, considering the number of shipments of each item, its volume in stock and the complementarity between the items (i.e., keep the products that are usually shipped together nearby).

Procurement

Procurement applications focus on the relationship between the company and its suppliers as well as the process that exists in relation to this relationship. Its basic objectives are to allow an efficient and streamlined purchasing process, and to manage specifications, prices, purchase orders, and the suppliers themselves.

These systems allow for analytical comparisons between suppliers and between products to help decision makers with regard to what to buy and from whom.

Order Fulfillment

The fulfillment process, or demand fulfillment, determines the promised delivery date for the orders and, therefore, strongly influences the lead-time of the orders, as well as the punctuality indicators for delivery. In the current competitive environment, it is very important to generate delivery dates quickly and reliably, a procedure that helps in providing a quality logistics service. The traditional approach to calculating these dates is to check the stock level to assess whether a given order can be filled immediately. If there is not enough stock, delivery of the order is promised for a date shifted in the future according to the required manufacturing time. This procedure may result in unfeasible orders, since other constraints that were not taken into account may be being violated, such as the availability of capacity or the supply of the necessary raw material.

Existing demand fulfillment solutions in supply chain systems use more sophisticated date determination procedures, with the aim of:

• Increase the timeliness of order deliveries by generating more viable orders
• Decrease the amount of lost sales

These are, in general, the functionalities available in SCM systems. Based on the large number of possibilities offered by these systems, a very pertinent question is how they are being implemented and used.

This will be the focus of the next article, who is implementing these systems, what are the problems normally encountered during this process, and what is the degree of satisfaction with the tools.

BIBLIOGRAPHY

KAHL, Steven J., 1999, ”What's the Value of Supply Chain Software?”, Supply Chain Management Review.

KANAKAMEDALA, Kishore, RAMSDELL, Glenn, SRIVATSAN, Vats, 2003, ”Getting Supply Chain Software Rigrht”, The Mckinsey Quartely, Number 1

SIMCHI-LEVI, David, KAMINSKY, Philip, SIMCHI-LEVI, Edith, 1999, Designing and Managing the Supply Chain: Concepts, Strategies, and Cases, Irwin/McGraw-Hill

STADLER, Hartmut, KILGER, Christoph, 2000, Supply Chain Management and Advanced Planning, Concepts, Models, Software and Case Studies, Springer, Berlin

Anthony, RN (1965) Planning and control systems: A framework for analysis, Cambridge/Mass