The Use of Inductive Methodology to Understand Incongruities in
the Australian Fresh Produce Supply Chain

Tamanna Ahmed Monem
School of Natural & Rural Systems Management
The University of Queensland, Australia
and
Dr. Ray Collins
Associate Professor
School of Natural & Rural Systems Management
The University of Queensland.

This study is a part of a Ph.D project at The University of Queensland, funded by RIRDC.

Abstract

During the 1990’s, critical reviews of Australian horticulture concluded that it had poor international competitiveness. Findings suggested that overseas buyers viewed Australian horticultural suppliers as unreliable, inconsistent and not supplying what consumers wanted. Critics identified a ‘perception-gap’ among Australian exporters, Asian buyers and consumers about product quality, service and value for money. They concluded that Australian suppliers did not understand what buyers needed.

Researching this problem requires a rich understanding of channel members’ perceptions and behaviours, an understanding that does not lend itself to more traditional quantitative research approaches. This study uses an inductive methodology to search for incongruities among channel member's’ perceptions of consumers’ preferences for Australian fresh produce. Inductive methodology allows the exploration of concepts from qualitative data without having pre-conceived ideas or hypotheses about the problem.

Data has been collected in two stages. Stage one documents grower-exporter-importer relationships and perceptions of what consumers want and what suppliers provide. Stage two involves looking backwards to see how consumer preferences translate to purchase behaviour. The two sets of results are compared using two relatively new analytical tools – NUD*IST (Non-numerical Unstructured Data Indexing Searching and Theorising) and SOM (Self-Organised Maps).

The objective is to examine whether consumer preferences are understood and effectively communicated back down the chain, and what communication role importers and exporters play. A case study approach was adopted. Malaysia was chosen because of its importance as one of the major export markets for Australian horticultural produce, in particular apples and broccoli. In stage one, most major exporters of apples and broccoli from four cities in Australia, and all their major importers in Malaysia were surveyed by in-depth interviews. In stage two, 1 000 consumers were surveyed face to face at point of purchase in Kuala Lumpur, Malaysia, over a three year period from 1997 to 1999.

Results demonstrate that incongruities do exist between what channel members think of consumers' preferences and what consumers' preferences actually are. For example, consumers seek long stemmed broccoli but get short stemmed broccoli. Consumers want more variety in apples importers and exporters claim they want. Consumers are more concerned about country of origin of product than importers and exporters claim. Consumers may not be unhappy with what they get, and their aggregate consumption may still be rising, but they do have unfilled needs that others in the chain do not know about, or are not responding to.

This study has implications for horticultural marketing. Efficient chains are those that can react quickly to changing customer demands and clearly identify areas for growth. Ineffective chains are associated with diminished competitiveness. In this case the chain will need to be more influenced by the ‘pull’ signals of consumers if it is to become more competitive. Here, and perhaps in many other cases, an industry-wide, whole of chain strategy focused on better consumer understanding and clearer communication is needed.

Inductive research has resulted in rich insights that may have been hidden from more traditional approaches. The use of powerful analytical tools like NUD*IST and SOMine allow deep mining of data, as well as the ability to present results visually to supply chain members.

Introduction

This paper focuses on behavioural issues related to the Australian fresh produce supply chain to an export market. Australia’s geographical position, combined with the seasonality of most horticultural products, provides scope to supply fresh produce to major markets in Asia, Europe, and North America during their off-seasons. However, throughout the-1990’s, industries related to Australian horticultural exports have been criticised for their poor performance, especially for their lack of professionalism and lack of initiative to deliver quality performance in the export market (Industry Commission 1993; Horticultural Policy Council 1992 & 1994; Science & Industry Forum 1988; Frank Small & Associates 1995). These reviews concluded by saying that buyers of Australian fresh products reported that Australia didn’t supply what consumers wanted, supply was unreliable and inconsistent, and that Australian fresh fruits and vegetables were too expensive.

Fundamental to all these criticisms is whether or not Australian suppliers understand what buyers need. In assessing buyer needs, supply chain members, from farm gate to retailers, need information on consumer preferences. Any mis-communication of these preferences has the potential to negatively impact on the supply chain and lead to diminished competitiveness of some chain members, especially suppliers.

Researching this problem requires a rich understanding of channel members’ perceptions and behaviours. This study uses an inductive methodology to search for incongruities among channel members'’ perceptions of consumers’ preferences for Australian fresh produce. The study aims to identify consumer preferences in relation to apples and broccoli purchases in a significant Asian market (Kuala Lumpur, Malaysia) and to compare these preferences against supply chain members’ perceptions of the same issues.

Methodology

Since this research aims to understand market competitiveness of an Australian fresh produce supply chain, it is in effect studying a phenomenon, thus it is an applied research. This research uses an inductive methodology to identify areas of concern related to this issue. Inductive methodology allows the exploration of concepts from qualitative data without having pre-conceived ideas or hypotheses about the problem (Glaser and Strauss, 1967).

Malaysia was chosen as a case country because of its importance as one of the major export markets for Australian horticultural products. Critics also have identified this market as one of the dissatisfied markets in relation to suppliers’ behaviour. In order to gain insights and understandings into how consumer preferences, export marketing functions and suppliers are perceived, understood and managed by chain members, an iterative style of research was required. Thus this research adopted an inductive, exploratory approach, where the problem becomes more and more clearly defined from understandings drawn from the data. This paper describes the first phase, where the objective was to gain insights into consumer related issues and how this information can flowed back to the supply chain.

Methods

The study was entered into with no prior hypotheses on supply chain members’ perceptions of consumer preferences in Kuala Lumpur, Malaysia. Data has been collected in two stages. Stage one documents grower-exporter-importer relationships and perceptions of what consumers want and what suppliers provide. Data in stage one were collected from depth interviews with these supply chain partners. Stage two involves looking backwards to see how consumer preferences translate to purchase behaviour. Data were collected from point of purchase interviews with consumers. Comparing supply chain members’ perceptions against consumer preferences will reveal whether or not consumer preferences are understood and effectively communicated from the demand end of the supply chain to the supply end.

Data collection and analysis

Malaysia was chosen as the case country because of its importance as one of the major export markets for Australian horticultural products, in particular apples and broccoli. Apple and broccoli were chosen as the case study products because they represent significant fruits and vegetables, and also for their importance in that market.

In stage one, seven exporters, four grower representatives, and seven importers were selected using a ‘snowball-chain’ sampling technique (Patton, 1990), which began by asking related people about export and import of fresh produce from Australia to Malaysia. In-depth interviews were conducted with participants, using a standardised open-ended interview approach (Patton, 1990) to produce descriptive explanations (Yin, 1994).

Supplier interviews were conducted in four major cities in Australia (Sydney, Melbourne, Brisbane and Hobart). Importer interviews were conducted in Kuala Lumpur, Malaysia, during 1997, 1998, and 1999. All interviews were taped, transcribed and analysed using NUD*ISTÓ (Non-numerical Unstructured Data Indexing Searching and Theorising). a comparative analytical tool to understand patterns or issues from qualitative responses (Qualitative Solutions and Research 1997).

In stage two, a total of nine hundred and ninety one consumers were surveyed face to face at point of purchase at five major retail stores in Kuala Lumpur, Malaysia, over a three-year period from 1997 to 1999. The objective of point of purchase interviews was to gather spontaneous and accurate responses from shoppers who were in the act of purchasing the case study products.

Consumers were asked questions regarding their preferences while purchasing fruits and vegetables, specific to apples and broccoli. The objective was to identify critical factors influencing their purchase decisions in relation to fruits and vegetables.

Consumer data was taped or noted, then transcribed and analysed using Kohonen’s SOMine, (Self-Organised Maps) an artificial neural network system, which helps interpret quantitative data in qualitative manner. SOMine produces maps, which are termed SOM (Self-Organising Maps) that allow visualisation of an abstraction of the original data. A component plane representation of a SOM provides information about correlations between data, the division of data in the input space, and the relative distributions of the components (Schmidt and Deboeck 1998).

Stage one: Suppliers and Importers Perceptions

Table 1 illustrates supply chain members’ perceptions of consumers’ preferences for apple and broccoli, generated using NUD*ISTÓ software.

Table 1: Perceptions of Participants about Consumer Preferences

Stage two: Consumer Preferences

A total of 991 responses were analysed using SOMine. Of these, 606 cases related to apples and 385 to broccoli preferences. SOMine produced a main map along with component planes which represent influences of different variables, for each case.

Determinants of fresh produce purchase - scenario apple

SOMine produced six clusters based on demographic data. Figures 1 and 2 show the main map along with the component planes indicating the respective contribution of each variable to the SOM. The scale at the bottom of each plane goes from the lowest value to highest value for that component.

Figure 1: The main map shows six clusters; the component planes on the right show the influences of different variables eg. prepacked, health/nutrition, snack, and variety of fruits. Blue areas indicate lower values of the planes; red areas indicate higher values of the planes. 

Figures 1 and 2 explain visual interpretations of each variable on purchase decisions. For example, price and colour are important to Cluster 1 (Figure 2), but prepacked apples may not be a major preference (Figure 1). Price and colour are important to Clusters 2 and 6.

Figure 2: The main map shows six clusters; the component planes on the right show the influence of purpose of eating – nutrition, snack, variety, and preference for prepacked apples.

More detailed information can be obtained from the statistical table produced by SOMine. Table 2 shows six clusters, four of them matched 30% to 9% of the matching records, two of them matched less than 8% of the respondents. From this table information about consumer clusters can be generated. |

Table 2: Statistical summary of clusters of determinants of purchase decisions for apples

• Cluster 1 Cluster 1 represents mostly young (20-30 years of age) and older (above 50 years of age) Malay-Chinese and Malay consumer groups. Price and colour are the most important factors in their purchase decisions followed by taste. Nutrition is important. Red apples are popular to this cluster.
• Cluster 2 Cluster 2 represents middle-aged (30-50 years of age) Malay-Chinese female consumers whose purchase decisions are influenced by price, then eating quality of apples followed by colour of the fruit. Red apples are very popular with this group. This group likes to eat apples for health/nutrition. Prepacked apples are popular with this group.
• Cluster 3 Cluster 3 represents middle-aged Malay female consumers. Purchase decisions are influenced by colour followed by price. They prefer hard, crunchy, juicy and a combination of sweet and sour or sweet red and yellow textured apples. Health/nutrition is their number one reason for eating apples.
• Cluster 4 Cluster 4 represents middle-aged Malaysian-Indian male consumers. Price and country of origin influence their purchase decisions. They prefer Washington apples, then New Zealand and Australian apples. Royal Gala is popular with this group.

An overall understanding of these clusters can be gained from Figures 3 to 6.

Figure 3: Three-dimensional plot of factors affecting purchase decisions for each cluster for apples.

In Figure 3, horizontal bars represent demography and purchase factors for apples. The vertical axis represents relative value of each factor. This graph shows that colour plays a very important role in cluster C3’s purchase decision. Price is highly important to cluster 1 and cluster 2. Country of origin is considered to be an important factor in cluster 4. Nutrition and snack purposes of apples are also important factors for this cluster.

Figure 4: Preferences of colour for each cluster

Figure 4 presents the importance of red colour apples to Cluster 2. Yellow apples are popular among Clusters 3 and 4.

Figure 5: Preferences for taste among clusters

Figure 5 represents the importance of taste for Cluster 2, followed by Clusters 3 and 1. Cluster 2 is very specific about defining a good quality apple. To them a good quality apple means ‘very juicy, crunchy, hard and to a lesser extent sweet apples’.

Figure 6: Country of origin factors and cluster

Figure 6 illustrates the importance of country-of-origin to Clusters 4 and 1. Cluster 2 prefers Australia, but Cluster 1 prefers apples from the USA.

Apples

Table 3 represents an overall picture of cluster preferences. For example, colour is important among middle-aged Malay females, and young and older Malay-Chinese and Malay consumer groups. However, red apples are very popular among middle-aged Malay-Chinese consumers, and to a lesser extent among young and older Malay-Chinese and Malay consumers.

Determinants of fresh produce purchase - scenario broccoli

Figures 7 and 8 show the SOM results regarding factors influencing consumer purchase decisions for broccoli in Kuala Lumpur, Malaysia.

Figure 7: The main map shows six clusters; the component planes on the right show the influencing factors eg., country of origin, stem and head size, price and overall size of broccoli.

Figure 8: Influence of prepacked broccoli and purpose of eating on consumers' purchase decisions.

Figures 7 and 8 show visual images of factors influencing purchase decisions of broccoli. For example price, country of origin, stems length and head size influence broccoli purchase decisions of Clusters 1- 4 and 6. Clusters 4 and 5 consume broccoli for nutrition.

Table 5 provides detail information regarding SOM-produced clusters, C1 to C6, from 385 consumers. Four main clusters are between 61% to 6% of the records and two small clusters are below 4%.

Table 5: Statistical summary of clusters of determinants of purchase decisions for broccoli