THE IMPACT OF PRODUCT CHARACTERISTICS, OPERATION AND DEMAND ON THE TYPE OF ORGANIZATION OF THE PRODUCT FLOW: FIELD RESEARCH IN SIX SECTORS OF THE RANKING EXAME BEST AND BIGGEST

The importance attributed to integrated logistics as a fundamental element for the efficient and effective management of supply chains is growing in the specialized literature on operations and services. Supply chain management normally means the management of related flows of products, information and financial resources ranging from the initial supplier to the final consumer, with financial flows as a counterpart. Integrated logistics, having as main premise the minimization of the total cost for a certain pre-established service concept, presents itself as an important process for the management of the supply chain, allowing not only the achievement and packaging of these flows, but also their characterization along the way. of the time between the different stages of the chain (BOWERSOX et al., 1996). Currently, there are several motivators that lead to a growing search for the integration of different systems of activities and value chains, within the scope of the supply chain.

1. Pressure to reduce inventory levels due to high opportunity costs of holding inventories.
2. Pressure to streamline customer service, reducing delivery time and increasing availability, in view of the growing customer demand in recent years.
3. Pressure to mass customize, that is, to offer a wide variety of customers products exclusively designed to meet their specific needs, still as a reflection of the growing demands in recent years.

For integrated logistics to assume a relevant role in the creation of a sustainable competitive advantage, its main decisions should be articulated in such a way as to mutually reinforce each other over time, allowing the creation of decision patterns consistent with the concept of the service and with the characteristics of the product and market for which this service is intended. The articulation over time of a decision-making process with the main elements that make up the external environment and the external environment, aiming at the creation and maintenance of sustainable competitive positions, is an issue widely studied by the area of ​​business strategy.

According to PORTER (1980), this strategic issue could be evaluated at the transversal and longitudinal levels. The first would deal with the connection of the company's internal and external characteristics (product, operation, market, etc.) to its performance (profitability and return) at a given moment in time. This link would help define the strategic fit between value chain/system decisions and the intrinsic characteristics of the business. The second would deal with why certain companies were able (or not) to reach positions of advantage and sustain them (or not), in addition to what allowed them to secure advantages in the future. The author points out that the analysis of the transversal level would be a priority, because without a specific understanding of what sustains a desirable position, it would be extremely complex to deal analytically with the longitudinal level.

With regard to the cross-sectional analysis of integrated logistics as a means of creating competitive positions, the specialized literature on operations and services registers, in a dispersed and diffused way, that certain decision-making patterns would be more adherent/appropriate and/or would be verified more frequently to a given set of product, operation, and demand characteristics. For example, product characteristics would include added cost, cost density, degree of obsolescence and degree of perishability. Relevant characteristics of the logistics operation would involve the cost of freight, inventory turnover, demand visibility, the ratio between deadlines and the combined time of the supply and distribution cycles. Finally, the characteristics of the demand would involve the amplitude of sales, or how much the maximum demand is superior to the minimum demand, being able to be a consistent indicator of the degree of predictability of the operations.

The field research focused its efforts on understanding how the characteristics of the product, the operation and the demand could shape the main decisions related to the organization of the flow of products in companies that manufacture consumer goods, which sell not exclusively, but necessarily, to the retail. Specifically, the following decision areas in integrated logistics were considered relevant for defining and understanding the type of product flow organization at its most generic or broad level: physical location or allocation of stocks, coordination of replenishment or product flow , and basis for triggering the flow of products (inventory management); number of stages and number of installations (dimensioning of the network of installations) and choice of modal and procedures for consolidation of shipments (transport policy).

2. THEORETICAL FRAMEWORK: HOW THE CHARACTERISTICS AFFECT THE TYPE OF ORGANIZATION OF THE PRODUCT FLOW

In this section, the theoretical-conceptual reasons are presented, in addition to empirical evidence found in the specialized literature, on how and why certain characteristics of the product, operation and demand, would favor certain types of organization of the flow of products, defined based on decisions tactical-strategic inventory management, dimensioning of the facilities network and transport policy.

2.1. Inventory Management

Inventory Allocation. It would encompass the decision about the physical location of stocks in the network of facilities, being strongly associated with the physical positioning of its cycle and safety components. Centralizing inventories in a single facility would imply postponing the flow of products in space, while decentralizing them across two or more facilities would mean bringing forward the flow of products in space. The specialized literature indicates that there are three factors that would determine a greater or lesser degree of centralization of inventories in a network of facilities: the characteristics of the product, the characteristics of demand and operation, and decisions in transport policies regarding the contracting of premium transport. or exploiting economies of scale arising from the consolidation of shipments.

Corroborating the above, the characteristics of the product would cover the following dimensions: cost of goods sold, density of aggregated costs and the degree of obsolescence. In general, it could be said that the greater the COGS, the density of aggregated costs and the degree of obsolescence of the products, the greater the tendency towards centralization of inventories (SILVER et al. (1985), BALLOU (1992) and CHISTOPHER (1997)).

On the other hand, the characteristics of the demand and the operation would cover, respectively, the amplitude of the demand and the turnover. The greater these factors are, the greater the propensity to decentralize inventories, basically because the risks associated with obsolescence, loss or stranding of products are minimized. Normally, products with longer life cycles and a small number of substitutes present a more predictable demand profile (SILVER et al. (1985), MENTZER et al. (1998) and WATERS (1992)).

Other factors relevant to the decision to allocate inventories are the contracting of premium transport and the creation of economies of scale in transport. Hiring premium transport can favorably contribute to the centralization of inventories, as it breaks with the premise of local presence (BOWERSOX et al., 1996). The existence of strong economies of scale in transport, on the other hand, create a favorable environment for the decentralization of inventories, as it makes the distribution of products in the supply chain relatively less costly (BALLOU, 1992).

Basis for Activating the Flow of Products. It would involve the integration and coherent articulation of inventory management with production and distribution policies. For example, from this perspective, the decision to produce to stock (anticipating production in time based on sales forecasts) or to produce to order (postponing production in time until actual demand is met) is of fundamental importance for the design of logistic systems (BOWERSOX et al. (1996) and CLOSS et al. (1998)). Some factors normally observed in the literature regarding this decision are: the COGS, the degree of obsolescence and the fixed/variable cost structure of the production process.

The degree of flexibility of the manufacturing process could favor production to order (using actual demand as the basis for triggering the flow), to the extent that it would be economically feasible to postpone the execution of certain stages until the customer places the order. Under certain circumstances, final mixing, assembly and packaging operations would be postponed until there was a definition regarding which SKUs (stock keeping units) would be sold, thus eliminating the risks associated with the uncertainty of future demand (ZINN et al. , 1988).

The profile of costs added to the product at each stage of production would indicate when and by how much the product's added cost increases. The greater the proportion of added costs in the final stages, the greater the benefit associated with the postponement of these activities in time (PAGH et al., 1998). Finally, to the extent that it was possible to reconfigure manufacturing operations, making them more flexible (implying a greater proportion of variable costs) and less dependent on economies of scale (implying a lower proportion of fixed costs), it would become viable make-to-order policies economically.

Product Flow Coordination. It would encompass the decision about the logic of triggering the flow of products, whether pulled or pushed. The notion (perspective) of pulling or pushing the flow of products would be directly related to the stage of the chain responsible for the decision to resupply stocks, that is, for a given link, whether it would be the later stage (closer to the customer or final consumer) or if it would be the previous stage (closer to the initial supplier). A pulled product flow would start in the later stage, through the transmission of information to the previous stage pointing to the need for resupply. On the other hand, a push flow would start at the previous stage, by estimating, using forecasting techniques or other planning methods, future material consumption requirements. Empirical evidence pointed out by several authors (STALK (1988), INMAN (1999) and CHRISTOPHER (2000)) indicate two basic factors that should be observed.
ved with respect to this decision: total supply and distribution cycle times at each stage and the visibility of demand.

Shorter supply and distribution cycle times than required by the final customer would favor the coordination of the product flow to be pulled, that is, controlled by the stage closest to the final consumer. Conversely, supply and distribution cycle times longer than those required by the final customer would require that the flow of products be pushed in anticipation of demand, that is, controlled by the stage furthest from the final consumer.

One of the main problems regarding product flow coordination is limited visibility into end-consumer demand. The point at which actual demand penetrates the supply chain towards the initial supplier is known as the decoupling point, according to CHRISTOPHER (2000) or the order penetration point, according to SHARMAN (1984). ). The implicit concept in the point of decoupling or order penetration point is a change in the way of coordinating product flows. In reality, the main question is not how far from the final consumer an order is being placed, but whether the real demand (from the final consumer) is visible or not for a given stage of the chain. The non-visibility of demand would lead to pushing the flow of products, whereas the visibility of demand would allow the flow of products to be pulled.

2.2. Transport Policy

Depending on the characteristics of the chosen modal, the safety, cycle and transit stock levels could be higher or lower. Preliminary analyzes of the product's characteristics, with a view to selecting the transport mode, should transcend its weight and/or volume dimensions and the CPV. According to the literature, a determining factor for choosing the mode of transport is the cost density, that is, the ratio between the aggregate cost of the product and its weight. The implications of a product with low cost density ($/kg) in the choice of transport mode would be relevant, since in order to design the operation with the lowest total logistical cost, the transport mode whose unit cost was compatible should be chosen with the cost density of the product.

Another factor that should be observed when choosing the transport mode is the variability of the demand for the products to be transported. The implications of a product with high demand variability in choosing the most appropriate mode of transport would also not be negligible. The lowest total logistical cost operation would be achieved through a transport mode that provided enough flexibility to keep up with variations in demand, minimizing the chances of wrong decisions, such as sending the wrong quantities of the wrong product to the wrong place. In this case, faster modes of transport with less capacity, allowing shipments to be consolidated in a shorter period of time, such as road transport, would generate the necessary flexibility for operations to follow demand fluctuations.

2.3. Facilities Network Sizing

It would involve defining the number of stages in the network, the number of facilities in each of the stages, the location and mission (products and markets served) of each of the facilities. These decisions are umbilically associated with the definition of the decision pattern in stock allocation and the formalization of the transport policy.

3. PROPOSED TYPOLOGY FOR UNDERSTANDING THE ORGANIZATION OF THE PRODUCT FLOW

Once the decisions referring to the Transport Policy, Stock Management and the Dimensioning of the Facilities Network are presented as relevant for the definition and characterization of the flow of products, an additional consideration regarding its nature must be made for a better understanding of the typology for the organization of the flow of products adopted in the field research. Even though these three decisions at the strategic level are the main and primordial ones for the creation of a competitive advantage through integrated logistics; are the stock management decision areas that characterize and define the flow of products and information in its main dimensions: ratio between physical quantities and processing times, average frequency between consecutive activations, distance/time between points of origin and destination, predictability of demand/sales/operations and definition of responsibility for triggering them.

In this sense, Table 1 presents a typology for classifying product flows in their three main dimensions: coordination, allocation in space and basis for triggering in time. Although it involves a certain degree of arbitrariness as it comprises only three decisions to characterize the flow of products and disregard many other possible ones, the typology for classifying the flow of products incorporates the broader decisions found in the specialized literature, according to which a policy for the organization of the flow of products, it would be possible to define it in its broadest sense: push/pull, anticipate/postpone in space and anticipate/postpone in time. The other resources/activities or structural/infrastructural dimensions present in the transport policy or in the sizing of the network of facilities would, in principle, be defined in accordance and coherently with the chosen pattern of the flow of products and information, in turn, a reflection of the characteristics do p
product, operation and demand.

It is noticed that theoretically there would be at least eight different ways/types/patterns for a company to organize its product flow . For example, there are pull flows, triggered by demand and centralized, as in the case of Dell and the VW Modular Consortium in Resende, and push flows, triggered by sales forecast and decentralized, as in the case of the food and oil industries. . However, only six of these eight flows are subject to proof, as pushed flows can only be triggered by forecasting sales, and not against actual demand.

 

4. RESEARCH OBJECTIVES AND METHODOLOGY

Taking the theoretical framework as a starting point, and considering the perspective of a company that manufactures consumer goods, the main question that the field research proposes to answer about the static approach of creating competitive positions based on integrated logistics is:

“WHICH PRODUCT, OPERATION AND DEMAND CHARACTERISTICS ARE SIGNIFICANTLY RELATED TO THE PRODUCT FLOW ORGANIZATION TYPES?”

More specifically, and for research guidance purposes, this general question unfolds into the following specific questions, subject to falsification based on hypothesis tests .

(1) Which product, operation, and demand characteristics are significantly related to decision-making regarding (a) inventory allocation, (b) product flow coordination, (c) the basis for triggering the products, (d) the transport policy and (e) the sizing of the facilities network?

(2) What characteristics of the product, operation and demand significantly influence the definition of a policy for organizing the flow of products in its broadest sense?

(3) What are the characteristics of the product, the operation and the demand and which are the decisions that present significantly different behavior, when the sample is stratified by type of sector of the economy?

The answer to these questions will make it possible to refine and propose a unification for the theoretical framework, which is currently diffuse, under the static approaches of creating competitive positions based on decisions in integrated logistics. This unification will allow/enable the presentation of a conceptual/normative instrument for evaluating the current decision-making standard in relation to the organization of the flow of products, at least within the limits that can be generalized from the sample used.

The population considered as the starting point for this research is that defined by the set of 500 companies that make up the list of the publication Exame Melhores e Maiores Edition 2000. The population in question is composed of 22 different sectors of the economy (sub-populations), covering extractive, industrial and service activities. Due to the delimitation of the scope of the research to the recognized industrial sectors of durable and non-durable consumer goods that sell necessarily, but not exclusively, to retail, sub-populations related to the primary and tertiary sectors of the economy were discarded, in addition to sub-populations of secondary sectors not covered by this restriction. In addition, time and deadline constraints imposed restrictions on the survey of the six largest sub-populations in the secondary sector of consumer goods for retail sale.

In this way, the sub-populations (sizes) addressed by the field research are composed by the Chemical and Petrochemical sector (46), Food (40), Automotive (31), Technology and Computing (26), Electro-electronic (21) and Pharmacist (17). The total size of these sub-populations amounts to 181 companies, or 36,2% of the total population of 500 companies. However, if the set of industrial sectors of durable and non-durable consumer goods that sell necessarily, but not exclusively, to retail listed in Exame magazine is considered for the purpose of defining the population, the population size is 254 companies. The 6 sub-populations addressed by the survey make up 71,3% of this total.

For each of the six sub-populations defined above, six samples were collected following a quasi-random process (quasi-random) with repetition. The process was almost random, because, despite the sample of each sub-population having been randomly generated, based on the list of companies present in Exame Magazine, some of the companies initially contacted refused to participate in the field research, leading to their replacement by others from the same sector/sub-population who were willing to participate in the research (convenience sampling). The process was repeated, as information was collected from each company regarding a class A SKU in billing and a class E SKU in billing. The determination of the sample size for each sub-population had as background the understanding and interconnection of the following questions: approximation through the normal distribution, choice of the statistical method and sample stratification, observing the following aspects.

• There is no minimum sample size necessary to confirm the validity of the normal approximation in each sub-population, for two basic reasons: in addition to sub-populations (sectors) being finite and small, historical reports indicate the asymmetric character of sectoral variables;
• The non-parametric tests do not require that the sample sizes of the six sub-populations (sectors) surveyed be the same, and should be used when the premise of the normal approximation in the sub-populations is not valid;
• As the estimation of population parameters is not the main objective of the research, but the identification of the correlation between variables and the control of spurious effects among sub-populations (or sectors of the economy), the sample fractions of each stratum do not need to be necessarily equal to the fraction of each sub-population (proportional stratification), and the sample size of the six strata surveyed may be equal for simplification and convenience purposes (disproportionate stratification). The effects of disproportionate stratification were considered when estimating population parameters, with the weight of each stratum in the total calculation corrected by the ratio between the sample fraction collected and the population fraction.

Observing these points, it was decided, for reasons of simplification, without, however, affecting the reliability of the research results and without prejudice to its objectives, to initially collect samples of size equal to 5, in each of the six sectors defined in the research . However, some companies selected initially refused to participate in the survey, configuring the final table of sample sizes and sample fractions collected in table 2.

The questionnaire constructed to meet the objectives of the field research is composed entirely of factual questions administered by two other researchers, in addition to the author of this work, during the second half of 2001. Questions of a factual nature considerably simplify the questionnaire project, as aspects such as the elaboration of different alternative ways for the same question are minimized, while in aspects such as the possibility of the interviewer interacting with the interviewee to ensure the correct understanding and answer to the question, there is greater flexibility than in opinion questions (MOSER et al., 1971). In addition to these aspects, it is worth mentioning the fact that the questionnaire was designed to be answered in an average time of one hour and a maximum time of one and a half hours.

It should be noted that the factual questions contained in the questionnaire presented at the end of this section refer to current events and are familiar to respondents. Managers or supervisors of medium hierarchical level, responsible for the logistics area (storage, order processing, inventory and distribution) answered for each company. In some cases, the interviewed managers also had within their reach of control decisions related to the purchase of supplies, production scheduling and export/import, a fact verified exclusively in the automobile, pharmaceutical and technology and computing sectors. Furthermore, in circumstances in which respondents were unaware of the variables surveyed, employees from other sectors with specific expertise on the topic responded.

5. ANALYSIS OF RESULTS AND CONCLUSIONS

The researched sample, due to its size and heterogeneity, did not adhere to the conditions of symmetry and normality required for the analysis of variance (ANOVA) techniques and calculation of Pearson's Linear Correlation Coefficient to be satisfactorily applied. Thus, Non-Parametric tests were used to test the hypotheses of mean differences between sectors and the significance of the correlation between strategy decisions and product, demand and operation characteristics. Specifically, the Kruskal-Wallis test was used to identify differences between sectoral means, both for decisions and for characteristics, and Kendall's Tau-b Correlation Coefficient (G) to identify significant relationships between variables ( CONOVER, 1971). All the analyzes that follow had a significance level of 95% (p=0.05) as a criterion to test the null hypotheses of equality between sectoral means and that there is no correlation between the variables.

Even though the results found did not fully adhere to the literature review on which product, demand and operation characteristics would shape decisions in logistics services strategy, it was possible to confirm several statistically significant relationships in Brazilian companies, in addition to identifying other that had not been considered. In summary, the following statements could be made, considering a static perspective of the competitive environment that permeates the different sectors surveyed.

Product flow coordination did not present significant correlations with demand visibility and supply and distribution cycle time, as pointed out in the literature review. Other variables such as the degree of obsolescence, degree of perishability, COGS and freight value, in addition to the ratio between deadlines, which is a variant of the supply and distribution cycle time, had significant impacts. A possible interpretation for these results would be to state that the decision to coordinate the flow of products can be shaped by four main factors, as illustrated in figure 1.

(a) The Intrinsic Flexibility of System Response, indicated by the ratio between the delivery time of the finished product to the customer and the delivery time of the longest input (RP). The higher this ratio, the more flexible and cost-effective the company's ability to respond to the customer will be, leading to the organization of pulled flows.

(b) The Nature of the Aging Process of Stocks, indicated by the degree of obsolescence (GO) and perishability of products (GP). In the researched sample, these variables were negatively correlated (G=-0,678; p<0,001), suggesting that most of the times the element that directs the aging of products is exclusively obsolescence (life cycle) or perishability (expiration period). validity). The greater the degree of obsolescence, the greater the propensity for flows of pulled products, while the greater the degree of perishability, the greater the propensity for flows of pushed products. This is probably explained by the need for manufacturers of non-durable (perishable) consumer goods to articulate the pace of production and distribution with the shelf life or useful life of the product on the shelf of the retailer or distribution channel. On the other hand, the degree of perishability and COGS are negatively correlated (G=-0,372; p=0,001) implying a lower commitment of working capital to place one more product unit in stock; contrary to the degree of obsolescence, positively correlated with COGS (G=0,58; p=0,001).

(c) The Working Capital Need to finance one more unit of product in stock, represented by the COGS. The lower this need, the greater the propensity for pushed product flows.

(d) Expenses per kg with Distribution, represented by the value of the distribution freight per kg. The greater these expenses, the greater the propensity for product flows pulled.

The calculation of descriptive statistics for each of the factors considered above is extremely relevant to quantify under what circumstances the product flow should be pulled or pushed. In other words, knowing the mean, median, minimum, maximum and standard deviation of variables such as the COGS, the degree of obsolescence, the degree of perishability, the ratio between deadlines and the cost of freight, in relation to sets of researched cases in which pull and push flows were verified, can help managerial decision-making with regard to the coordination of the flow of products. In this way, it would be possible to answer the following questions: what is the typical magnitude of a high COGS? From what level could a product be considered with a low degree of perishability? What is the order of magnitude of a high freight value per kg?

Figure 1 - Factors that Influence the Product Flow Coordination Decision

Table 3 is an instrument to support decision-making, showing the median of the factors significantly correlated with the decision to coordinate the flow of products. According to this table, products with a COGS of less than $6,50, without a degree of obsolescence, with a degree of perishability greater than 0,04 (equivalent to a validity period of 25 months), with a response time required by the customer of less than equivalent to 3% of the delivery time for the most critical input and a freight cost of less than $0,07/kg. On the other hand, products with a COGS greater than $680, with a degree of obsolescence greater than 0,04 (equivalent to a life cycle of 25 months), without a degree of perishability, with a response time required by the customer greater than equivalent to 11% of the delivery time for the most critical input and a freight cost greater than $0,14/kg.

The allocation of products is significantly influenced by the density of costs, the range of sales, the turnover of products and the contracting of premium transport. Variables such as COGS and the degree of obsolescence, contrary to what the literature review pointed out, did not present significant correlations with the allocation decision. A possible interpretation for these results would be to state that the allocation of stocks can be shaped, in general, by four factors described below, as illustrated in figure 2.

(a) The Risk of Maintaining Safety Stocks, represented by the Sales Range (AV). The greater the range of sales, the greater its variability, normally calculated by the standard deviation of sales, implying centralized inventories.

(b) The Need to Create and Exploit Economies of Scale in Transport, represented by the density of added costs (DC). Products with low density of added costs per kg or m3 normally imply transport policies of low unit cost per kg or m3, as a way to ensure competitiveness. Nevertheless, the cost density and the freight value were positively correlated (G=0,547; p<0,001).

Figure 2 - Factors that Influence the Inventory Allocation Decision

(c) Adherence of the Inventory Replacement Policy to sales levels, represented by inventory turnover (G). High product turnover normally results from inventory replenishment policies that systematically monitor sales levels or sales levels that are relatively higher than average batch sizes for production and distribution of products. These factors favor the physical decentralization of stocks, since the risk associated with their dispersion is lower. It is interesting to note that inventory turnover was negatively correlated with the product's cost density in the surveyed sample (G=-0,329; p=0,005). This corroborates the thesis raised in the previous items: if, on the one hand, products with low density of added costs would need to generate and exploit economies of scale in transport to remain competitive, probably through the consolidation of shipments over long distances, on the other hand , the average sales level between two consecutive replenishments should be high enough to ensure a reasonable inventory turnover.

(d) Transport Response Time and Variability, represented by contracting premium transport (TP). The faster and more reliable the transport response time, the greater the tendency towards centralization, overcoming the paradigm of local presence.

Table 4 presents the medians of the significantly relevant variables for the decision to allocate stocks, considering the centralization and decentralization decisions. This table also constitutes an instrument for decision-making regarding the allocation of products, allowing the identification of questions such as: what is the typical magnitude of a product with high turnover? From what level could a product be considered of low cost density? What order of magnitude of a high range of sales?

The adoption of the trigger base, as expected by the literature, whether based on sales forecasts or in response to actual demand, is significantly influenced by the COGS and the degree of obsolescence. In addition to these variables, they also influence the degree of perishability, the cost of freight and the ratio between deadlines. The four factors that influence the determination of the basis for planning are the same as those that affect the product flow coordination decision. Nevertheless, these two decision areas were strongly correlated in the surveyed sample (G=-0,822; p<0,001), indicating that pushed product flows necessarily depend on sales forecasts for their triggering, and pulled product flows are in great majority stimulated by real demand.

 

Figure 3 - Factors that Influence the Definition of the Basis for Activating the Product Flow

 

Table 5 presents the median of the variables that are significantly relevant to the definition of the trigger base, and is also a support tool for decision-making.

In addition to these conclusions, analyzes carried out among these six surveyed sectors revealed that there are significant differences in their decision-making patterns of product flow coordination, stock allocation, definition of production policy and choice of transport mode. Significant differences were verified not only in these decisions, but also in relation to some characteristics of the product, demand and operation.

• Food, electro-electronics, pharmaceuticals and chemicals and petrochemicals are sectors where pushed flows of products predominate, while automotive and technology and computing are sectors where pushed flows of products predominate.
• Pharmaceuticals and technology and computing are sectors where centralized stocks in a single facility predominate, while food, automotive, electro-electronics and chemicals and petrochemicals are sectors where decentralized stocks predominate.
• Automotive and technology and computing are sectors where actual demand predominates as the basis for activating the flow of products, while food, electro-electronics, pharmaceuticals and chemicals and petrochemicals are sectors where sales forecasts predominate as the basis for activation. Table 6 presents these results.

With regard to the types of product flow organization, field research points to the following conclusions, as shown in Table 7. (1) 69,2% of the cases surveyed have their product flows pushed based on sales forecasts, that is, anticipate future sales (Types 1 and 3); (2) 23,1% of the cases surveyed have their flows pulled in response to real demand, that is, they are postponed until real sales materialize (Types 2b and 4b); (3) the remaining cases (7,7%) have their flows pulled by stages closer to the final consumer, but even so they use sales forecasts to trigger them (Type 2a and 4a); (4) decentralizing inventories across multiple facilities is a dominant practice (61,6% of surveyed cases) in various Brazilian sectors; (5) pushing product flow means always using sales forecasts to anticipate future events in production and distribution and (6) pushing product flow means, most of the time (75% of cases), using actual demand to postpone production and distribution operations until demand is met.

Below are the conclusions regarding the type of organization of the flow of products for each of the six surveyed sectors.

(a) In the food sector, 70% of surveyed product flows are push, decentralized, and forecast-planned (Type 1), while the other 30% are pull, decentralized, and forecast-planned (Type 2a). The food sector is characterized by products with low COGS, low cost density, a certain degree of perishability, low freight costs, visibility of demand only in ECR programs and a small range of sales; with non-perishable foods being included in Type 1 and perishable foods in Type 2a.

(b) In the pharmaceutical industry, 80% of product flows are push, centralized, and forecast-planned (Type 3), while the other 20% are pushed, decentralized, and forecast-planned (Type 1). The pharmaceutical sector is characterized, above all, by a high cost density, very low obsolescence, as life cycles are long (proprietary and patented processes) and low perishability. Type 1 includes low CPV drugs and Type 3 high CPV drugs.

(c) In the chemical and petrochemical sector, 40% of the flows are push, decentralized and forecast planned (Type 1), 40% of the flows are pushed, centralized and forecast planned (Type 3) and only 20% are pulled, decentralized and forecast planned by demand (Type 2b). Type 1 includes fuels, Type 3 lubricants and Type 2b some chemical inputs with a higher CPV and low turnover.

(d) In the electronics sector, 50% of product flows are push, decentralized and forecast-planned (Type 1), 25% are push, centralized and forecast-planned (Type 3) and 25% are pull, decentralized and forecast-planned planned by demand (Type 2b). Type 1 includes imports with low CPV, Type 3 includes imported products with high CPV and Type 2b domestically manufactured electro-electronic products with low CPV.

(e) Finally, in the automobile sector 33 1/3% of the flows are push, decentralized and forecast planned (Type 1), 33% are pulled, decentralized and demand planned (Type 2b) and the last 33% are pulled, centralized and planned by demand (Type 4b). The technology and computing sector, due to the high percentage of imported items and the Brazilian customs barriers, does not have a defined pattern of organization of the flow of products.

6. NOTES

It can be approximated by the COGS – Cost of Goods Sold – financial-accounting indicator.

Ratio between the added cost and the weight or volume of the product, aiming to answer how much the product would cost per kg or m3.

Inverse of the lifecycle duration; products with longer life cycles would have a lower degree of obsolescence.

Inverse of the duration of the shelf life; products with a longer shelf life would have a lower degree of perishability.

Ratio of sales to average inventory level.

Indicates whether or not the company would access final consumer demand in real time.

Measures how much longer the deadline required by the customer is than the delivery deadline offered by the company; the higher this ratio, the more flexible and cost-effective the company's ability to respond to the customer will be.

Calculated as the sum of the delivery time for receiving the most critical input with the delivery time offered by the company.

Since they range from the commitment of resources and physical structures associated with the feasibility, achievement and packaging of the flow of products in space and time, to their characterization, through the establishment of routines and procedures for their activation.

Resultant of all possible combinations regarding push/pull decisions; bring forward/postpone in time and bring forward/postpone in space 2 x 2 x 2 = 8.

Conventionally, hypothesis tests test the theory-fact congruence in the direction of empirical data validating the theory (CASTRO, 1978). An example in this sense would be the centralization of inventories of products with high added cost (fact). The theoretical justification for such a decision would be the benefit obtained with the reduction of safety stock levels in the logistics network, and the consequent demobilization of tied up working capital. If there is a significant correlation (hypothesis test) between the variables added cost (independent or explanatory) and inventory centralization (dependent), it could be deduced based on the theoretical framework that this decision is related to the desire to reduce the opportunity costs of holding inventories, materialized by the level of added cost.

Defined, based on the theoretical framework, as a consequence of the articulation of allocation decisions, the basis for triggering the flow of products and coordination of the flow of products.

7. BIBLIOGRAPHY

BALLOU, RH, 1992, Business Logistics Management, 4th ed, Prentice Hall.
BOWERSOX, DJ, CLOSS, DJ 1996, Logistical Management – ​​The Integrated Supply Chain Process, 1st ed, McGraw-Hill.
CASTRO, CM, 1978, The Practice of Research, 2nd Edition, McGrawHill do Brasil, São Paulo.
CHRISTOPHER, M., 1997, Logistics and Supply Chain Management – ​​Strategies for Cost Reduction and Service Improvement, 1 ed. São Paulo, Pioneira Publishing House.
CHRISTOPHER, M., 2000, “The Agile Supply Chain – Competing in Volatile Markets”, Industrial Marketing Management, Vol. 29, pp.37-44.
CLOSS, DJ, ROATH, AS, 1998, “An Empirical Comparison of Anticipatory and Response Based Supply Chain Strategies”, International Journal of Logistics Management, pp. 21-34.
CONOVER, WJ, 1971, Practical Nonparametric Statistics, 1st Edition, New York, Wiley & Sons.
INMAN, R., 1999, “Are You Implementing a Pull System by Putting the Cart Before the Horse?”, Production and Inventory Management Journal, Second Quarter, pp.67-71.
MENTZER, JT, BIENSTOCK, C., 1998, Sales Forecasting Management, 1st ed, New York, Sage.
MOSER, C., KALTON, G., 1971, Survey Methods in Social Investigation, 2nd Edition, London, Heinemann Educational Books.
PAGH, JD, COOPER, MC, 1998. “Supply Chain Postponement and Speculation Strategies: How to Choose the Right Strategy”, Journal of Business Logistics, Vol. 19, No. 2, pp.13-33.
PORTER, M., 1980, Competitive Strategy: Techniques for Analyzing Industries and Competitors, New York, The Free Press.
SHARMAN, G., 1984, “The Rediscovery of Logistics”, Harvard Business Review, September/October.
SILVER, EA, PETERSON, R., 1985, Decision Systems for Inventory Management and Production Planning, 2nd ed., New York, Wiley & Sons.
STALK, G., 1988, “Time – The Next Source of Competitive Advantage”, Harvard Business Review, July-August, pp.41-51.
WATERS, CDJ, 1992, Inventory Control and Management, 1st ed., New York, Wiley & Sons.
ZINN, W., BOWERSOX, D., 1988, “Planning Physical Distribution with the Principle of Postponement”, Journal of Business Logistics, Vol.9, No.2, pp.117-136.