PRODUCT DEVELOPMENT AND PROMOTION REFERENCES

Contact Us

Modified From Article: The developments in computer technology throughout the late 1980's and early 1990's have put the computing power of main frames on to the desks of nearly everone who wants it. But perhaps more importantly, the arrival of truly user friendly software operating through the WINDOWS format has meant that, for the first time relatively inexperienced computer users can run a wide range of very powerful packages. In addition, the relative cost of automated data collection systems has dropped sharply.

These three factors have produced something of a revolution in many areas of the food industry, particularly in the control of food processing and in the management of food distribution and retailing. However in the area of new product development (NPD), the introduction of novel uses of computers has been relatively slow. One reason for this is that NPD is not a "routine process" and does not offer itself up for automation very easily. Another is that NPD involves a large number of distinct sections of a company, who are often doing a number of other activies, and it is rare that the NPD process per se comes under scrutiny for efficiency or innovation.

Failure Costs

The need for innovation in NPD is evident from the large number of new products that are withdrawn a few months after launch. The overall costs are difficult to gather together, but one estimate puts the annual cost to the food, beverage and fragrance industry of new product failure in the USA to be $12 billion dollars! (Humphries, 1994). One immediate result of comptuerized records and costings is that the true profit of a particular product, taking into account all development and launch costs, can now be calculated fairly precisely. This is perhaps why some companies have reduced their innovative NPD during the recent recession and concentrated on introducing brands that are successful in other countries.

The approach in the article was to examine the product development itself and then to examine the scope for innovation in each of its phases and to describe briefly some recent developments that are available at the current time.

Product Development Cycle

In the United Kingdom the big retailing chains have been great engines of NPD over the last 15 years, the figure below indicates a critical path for the NPD process at TESCO, one of the U.K.'s most successful and innovative retailers. (Cross, 1994).

The Seven Phases of NPD Concept Generation Concept Screening Product Development Product Testing Package Development - including advertising material First Production Run Launch

The average life of an own label (O/L) product in the, UK is four years, but the development cycle may take from 16 weeks to a year. Sixteen weeks would be for a range extension with an existing supplier, whereas a totally new product concept might take a year. However the average target development time is 28 weeks and products that overrun their cirtical path timescales during development are often canceled because the cost outlay will not be returned in scales.

A Critical Path for New Product Development at UK retaielr Tesco

These timescales may be vastly extended when very novel products are being produced. The concept to launch period for the mycoprotein product "Quorn", including safety testing and legislative approval was over 10 years. Here we will look at the scope for computers in each of the phases of product development. Whatever the timescale, the 7 phases of NPD are: concept generation, concept screening, product development, product testing, package development (including advertising material), first production run and the launch.

Concept Generation

This may take a variety of forms. Brainstorming by the product team is the cheapest, but not always the most effective. Competitor ananlysis is a good starting point although "me too" products usually contain some innovations to ensure that the O/L product is in some way superior to the compeitor. Lifestyle trends analysis may be used and large scale media tracking and market trend analysis has been available from the big market research companies for some time.

Focus groups (Casey and Krueger, 1994) use small groups of consumers to discuss their needs and preferences for aspects of products. These are valuable for generation and screening (phase 2). A recurring problem has been how to get an immediate assessment of an individual's private views during the discussion. This has recently been solved using hand held radio transmitters connected to a computerized data collection facility similar to those used in television show audience voting systems.

Market gap analysis is a useful technique and is often done by making a new combination of qualitative attributes-e.g. a British soft cheese, a diet country-style lemon-ade, fish fingers with ketchup already added.

Recent innovations do enable product developers to do their own quantitative market gap analysis fairly cheaply using their sensory panels and multivariate mapping techniques. Thus descriptive profiles of the products in the market place can be obtained and mapped and sensory gaps in the market mix dentified.

With the introduction of preference mapping methods (Greenhoff and MacFie, 1994) on PC's it is now possible for the product range in the market place to be mapped using preferences testing of a number of products by consumers in central location tests.

Without doubt the future contribution of computers in concept generation will be in the automatic provision of up-to-date information on the market place and a faciity to test and revise concept ideas through mathematical methods that will identify gaps or enable theoretical "bundles of attributes" to be projected into the current market map.

Concept Screening

The market opportunity for a new concept can also be more accurately assessed using preference mapping because the various consumer segments that are likely to buy the new product can be identified at the very start of the process. If these segments are currently buying a competitor's product then further NPD will be initiated. However if the new concept will damage sales of a sister product, it will be dropped. The importance of the special functional or sensory attributes of the concept can be tested using conjoint analysis.(Cattin and Wittink, 1982). Using this technique consumers are presented with cards containing a range of attributes of a theoretical or real life product and asked to rate it for acceptability, likelihood of purchase, convenience etc. ƒ.The potential to use the consupt of consumer segments in connection with compuyterized consumer interviewing techniques to obtain a fully automated concept screening exercise, is clear.

The feasibility of production will clearly involve detailed consultation with the production engineers. However, there are now good opportunities for product developers and production to engate in taguchi style experimentation (Stone and Veevers, 1994) to estimate the feasibility of producing specific functional and sensory combinations prior to the main development phase.

Estimate Cost and Profit Now

The future will see computers bringing large-scale databases together to integrate these different phases of concept testing. Thus production engineers will sit alongside marketing, new product developers and financial managers and simultaneously test the effects of revising a concept on its production, consumer acceptance and profitability.

The challenge to the food industry will be to resource the experimentation and investment needed to fill these databases and develop the techniques to produce integrated prediction models.

Product Development

It should be clear that if the preceding phases have base done properly then this phase should be simply a case of finalizing the formulation and processing parameters to achieve thenecessary product performance. However, it is often the case that no previous experimentation has been done before this phase and so there is often scope to bring the computer-assisted technology into action.

Formulation studies requiring the accurate assessment of specific combinations of constituents and the use of mixture designs (Cornell, 1981) generated by computer (Snee and Marquett, 1974) are now becoming widespread. The most cost effective design algorithms follow the so-called evolutionary operations (Box, 1957) and are interactive. Thus one or more experimental samples are formulated and tested and the results are fed into the program which generates a new set of samples for testing. The program APO(Marte, 1986) has been used this to optimize baking operations. (Hanneforth et al., 1994).

Another recent innovation is the use of automated on-line formulation of a beverage in which sensory properties are optimized using the simplex algorithm, (Bardot et al., 1992) .

Sensory Analysis

Sensory analysis will be a pivotal part of the development process. Traditionally this testing would be done informally by the product development team, but increasingly a scientific approach is being taken and full descriptive profiling by trained sensory panels is carried out.

Besides the automation of data collection, analysis, graphing and presentation of results, these systems [not included here] offer means of data collection that have not been available before. Free choice profiling (Williams and Langron, 1984), in which assessors can use their own individual vocabularies, or time-intensity (Lynch, Liu, Mela and MacFie, 1993) analysis, in which assessors use a continuously moving facility such as a mouse or a joystick to record their time perception of an attribute such as sweetness or bitterness, are just two examples. The systems have also been linked up to staistical design programs so than samples automatically randomized and balanced for carry-over effects (MacFie, Bratchell, Greenhoff and Vallis, 1989).

The future of this phase of product development will, in the longer term, involve less practical work, and more computing as formulation databases and mathematical models are built. However in the short term it is again true that these facilities only come into place through extended experimentation and statistical analysis. The use of neural network algorithms (Bardot, Bochereau, Martin and Palagos, 1993) to model complex manufacturing processes and almost any other complex process will certainly increase.

Product Testing

Sensory assessment It is to be expected that a full sneosry profile of the product and all its variants will be available using the computerized systems mentioned above.

Consumer assessment. There are three possible phases of consumer testing. The simplest and cheapest is to use members of staff who taste it in-house or take it home - the so-called overnight panel. The second layer is the central location test; this has already been discussed above, but it is worth nothing that in the UK, Tesco have initiated the use of full time in-sore consjmer testing facilities, (Cross, 1994). The third, and most expensive, phase is the home test. There would appear to be considerable opporutnities to use computers and the information highways to automate and enhance the speed and quality of data being collected directly from consuemrs in their own homes.

Instrumental testing. The physical, chemical, functional and microbiological aspects of the product will be assessed at this stage. The scope for enhanced data collection and analysis through increased computerization is well understood and does not need to be rehearsed here. However, the recent development of predictive models for for microbial growth (Robers et al., 1993) does have implications for NPD and needs further research.

Package Development and Advertising

This operation has usually been udnertaken quite separately to the formal product development process and there has been a growing awareness that package design should be more intimately integrated with functional and sensory properties that the physical product delivers to improve perceived quality and potential added value. This is particularly true for non-retail brands which must establish or maintain themselves by these two aspects during periods of low advertising budgets. At the Institute of Food Research we have recently developed an integrated system in which packages can be manipulated using sophisticated desktop design software and then displayed for controlled time periods in the sensory booths before or during sensory assessment. (Buckley and Shepherd, 1993). This methodology needs to be further developed to include new statistical designs and testing protocols.

In anotehr exciting inovation Moskowitz and Gandler (1993) have used conjoint ananlysis to combine visual and audial images into an optimum cognitive presentation of the desired project. Since this takes consumer segmentation into account, it is potentially a very useful tool at the product concept as well.

First Production Run

Once a product has been passed by the technologists and the marketing department, there is usually a great race to get it on to the shelves or, if it is a branded product, to launch. The first production run is the time when the gap between concept and reality often becomes clear. It is important that the technologists are on hand to resolve processing problems. Samples should be sent for detailed sensory testing. If proper preliminary profiling and elucidation of the critical processing and formulation variables has been done, the reasons for derivation from target will be understood and immediate corrective action can be taken. If not there will be an embarrassing gap while the reasons for the faults are found, or more diastrously the product will be simply launched as it stands because the launch timetable cannot be changed. The moral here is to use all the computing and technological expertise to learn all about the product prior to the production run.

Launch

There is a crucial need to follow up the sensory quality and the consumer acceptability of the product for several weeks after launch. First it is important to know whether the production process has settled down and second to see how consumers are "conditioning" to the product. If expectations are not being satisfied, early reformulation may be signaled. In commuting terms, there are no new techniques that have not already been covered in this article, although routine monitoring of consumer acceptability of products through communication networks would appear to be a very useful tool in this context.

There is tremendous scope to enhan ce current new product development through imaginative use of computers to optimize processing and to gain added insight into consumer response. Perhaps the only barrier to the adoption of these new techniques will be the unwillingness of companies to switch investment from high technology labor saving production apparatus to aspects of the NPD process.

References

Updated: Thursday, September 6, 2007.