Predicting the purchase of innovations

Abstract

The main objective of this research is to examine how purchase intention can improve the prediction of actual purchase on an aggregated level. On macro level the adoption of products result in a product diffusion curve that can influence the relationship between purchase intention and actual purchase. According to former research on micro level, attitude, subjective norm and perceived behavioural control are expected to influence purchase intention. Using an Australian database provided by Roy Morgan Research, this research formulated three hypotheses:

H1: Purchase intention of a population at t-1 is positively related to actual purchase at t. H2: Product diffusion positively moderates the relation between a consumer’s purchase intention at t-1 and the actual purchase at t. H3: Purchase intention of a population is positively influenced by the consumer’s attitude, subjective norm and perceived behavioural control.

A sample that contains 22 entertainment products and purchase behaviour data of Australian households is used to test the expected relationships. The hypotheses are tested along the lines of the diffusion theory and the theory of planned behaviour. When a product enters the market, consumers adopt the product and the diffusion of the product starts. According to the widely used Bass model, the adoption of products occurs based on the imitation of precursor adopters. In other words, social pressure results in purchase of products. The Bass model is a S-shaped curve that displays the product adoption by consumers over time until a certain saturation point is reached. In this research, a difference is made between the first, A, and second, B part of this diffusion curve. This is tested by comparing results of the data in these two product diffusion parts. Furthermore, the theory of planned behaviour is used to approach purchase intention. This theory originates in the theory of reasoned action assuming that attitude and subjective norm influence the consumers’ intention. Later, research shows the added value of perceived behavioural control in understanding purchase intention. Therefore, all three variables are included in this research. The database of Roy Morgan Research contains factors that directly measure purchase intention, actual purchase and product diffusion. Attitude is measured by four statements regarding interest in technology that are derived from the dataset. Moreover, since product diffusion and subjective norm are both mainly based on social pressure, product diffusion is used to measure the influence of subjective norm on purchase intention. In addition, perceived behavioural control is measured by income and the ownership of complementary goods. All used data is a percentage of the total population that are transferred from the database to SPSS. Subsequently, ANCOVA models are used, because this test is able to include continuous variables, it enables to correct for the differences between the included entertainment products and it tests the relation between variables. All models that compare data in diffusion part A and B when compared to models analyzing all data, better explain the dependent variable based on the included independent variables. The result of interpreting the output of ANCOVA models is a confirmation of the first hypothesis regarding the main relation. Purchase intention positively affects actual purchase, as expected. The second hypothesis is not confirmed. With regard to the third hypothesis a significant positive relation was found between subjective norm and purchase intention. No significant relation was found between attitude or perceived behavioural control and purchase intention. Therefore, the third hypothesis concerning this relationship, except for the part concerning subjective norm, was rejected. In conclusion, subjective norm positively influences purchase intention and purchase intention is positively related to actual purchase. Firms can use this model to make predictions for innovations and more diffused products. First, when firms have developed several innovations and aim to determine which of these products will perform best on the market, the management team can use purchase intention to predict the future sales. Second, when a firm, for instance, lacks in having a precise sales history, purchase intention can be used to determine actual purchase of existing products as well. At least one year upfront a firm needs to survey consumers on the market about their purchase intention. Furthermore, the social network of consumers will influence the purchase intention and should be taken into account while measuring the purchase intention. In diffusion part A the actual purchase will be approximately the same population fraction as stated to have a purchase intention, while in part B the actual purchase will be approximately three times the measured purchase intention. To conclude, firms can base their future sales predictions on the purchase intention. After conducting this research, some recommendations regarding further research are formulated. First, more research on aggregated level can focus on the theory of planned behaviour, because in this research only subjective norm is significantly related to purchase intention, while also a relation was expected between both attitude and perceived behavioural control and purchase intention. Second, the hypotheses were tested with trend data, whereas panel data can provide more insights into the individual decisions. Third, further research can be done with other product categories and data obtained from more than one country to make it easier to generalize the results. Finally, future research can focus on the reason of purchase. Not enough data was available concerning replacement purchases, which can give a more complete view of the nature of the purchase intention and the reason for actual purchase.