Product Testing
Category: Marketing
This is the stage at which actual prototypes are tested. Deeper evaluations are now possible. In this field, the selection of an appropriate test sample is contentious. Most practitioners simply employ a judgement sample in the target segment. Midgley however argues against this by pointing out that new product success is dependent upon the purchase intentions of the innovators and that innovators may not be sufficiently represented in a judgement sample. Midgley advocates selecting a sample to contain both «innovators» and «laggards» in order to test the product against the perceptions of these two diverse groups of consumers. Ostlunds methodology for discerning innovators along dimensions of perceived risk, etc, could be used here.
Another major issue in setting up product tests is the choice between single or comparative tests. A common technique is to place a pair of products in plain packages for evaluation by a sample of consumers the «blind paired comparison». However it is possible that this exaggerates differences. Because they are given two products the respondents impute that there must be differences and so look for them. One solution to this problem is offered:
«If some kind of market acceptability rating is required, then the best strategy appears to be to make the test conditions as much like real life usage as possible, which may mean monadic (single) testing and to rely upon past experience or other research to indicate the meaning of the rating obtained».
Questions posed to respondents in product testing usually concern overall disposition towards the product and evaluations of particular attributes. Intentions to buy may also be investigated. All of these could employ scaling devices to elicit respondents’ feelings about the sample.
Product testing can appear deceptively simple. A group of consumers merely says whether or not the product is liked. Such simplicity disguises many pitfalls. For instance, do consumers make their evaluations in a laboratory setting (or in halls in various parts of the country) or do they make evaluations after using the product at home?
Realism and cost need to be balanced, but a cheap test giving the wrong result is worse than no test at all. Boyd suggests multistage testing with central testing on a larger number of product options and in home testing on a subsequently reduced number of options. It would be expected that the kind of evaluation and the circumstances in which it was made would affect the consumer’s reaction. Watkins found this to be so in his survey of company practice and says:
«A large company reported… high scores in terms of intention to purchase are obtained using monadic, in hall tests with an interviewer completed questionnaire after the respondent has tried the product first and been shown the packaging. Low scores are obtained with comparative in home tests with self-completion questionnaires».
Determining who is to do the testing can cause difficulties. Family products are obviously problematical the person who buys may actually hate the product. Pet foods can also be difficult to evaluate. Tull and Hawkins recount the case of a new dog food that was launched after extensive packaging and advertising research and which obtained good initial market penetration. Several months later sales declined heavily and it was then that it became clear that the product had not been tested on dogs: few would eat it given choice.
The figure below illustrates the concept and product testing which was conducted for the launch of the Ford Fiesta.
Concept and product testing for the Ford Fiesta
Ford’s move into the small car segment in 1976 was a major step, and it was preceded by very large scale marketing research. This research was in three phases
Mid 1972 to mid 1973 — selecting the right concept.
Mid to end 1973 — determining the specific product proposal: its precise dimensions styling interior space.
1974 to mid 1976. — detailing aspects with shorter lead time, e.g. seat design, door panel design, detailing exterior styling.
Phase 1 research methods included a survey among owners of existing small cars aimed at uncovering the strengths and weaknesses of different concepts in terms of performance, handling roominess, styling and economy. Surveys were also undertaken in specific segments, such as where a small car was a second car and among first time buyers. Two waves of ‘product clinic’ were also conducted in each of the five main European markets. Models had been built of alternative exteriors and interiors. Although these were of clay they were entirely realistic, even at close range. Samples of new car buyers were invited to these ‘clinics’ and co-operated in group discussions individual unstructured interviews and self completed questionnaires. Results from this were:
The essence of the interior design was settled.
A clear direction for styling was provided.
It demonstrated that an Escort-derived rear wheel drive small car would be a risk.
A range of possible sales volumes was derived.
Phase 2 research employed fibreglass models of four alternatives devised from the initial research. ‘Beauty contest’ clinics were used to evaluate these. There was one consistent winner among the four: this was approved and tooling for large scale production began.
Phase 3 lasted two and a half years and filled in all the final ‘detailing’. More clinic studies led to progressive refinement and allowed checks against new products from competitors. In this phase there was also research into the brand name and the launch advertising strategy.
Despite the problems of sample selection and test type selection, concept and product testing bring researchers another step closer to the market, and should furnish more opinions on the degree of success to be expected.