The new-product process stage of market testing involves which two of the following?

During the product development (PD) process, it is important to test out new iterations of the product design to verify the right fit and functionality. Later on in the process, we are then able to determine whether the design requires additional improvements and enhancements before it proceeds to production. While some companies offer virtual testing environments that analyze how products perform with regard to typical use conditions, environmental circumstances, and user error, obtaining and testing physical representations of a product (i.e., a prototype) remains a vital part of the PD process. Having a physical model allows product designers and developers to test out how the product feels and functions in real-world conditions, which can reveal issues that may not be apparent in a virtual test environment.

Optimizing the Product Design for Prototype Testing

As the product development process moves from the engineering stage to the prototyping and testing stage, it is essential to thoroughly understand the difference between a prototype and a product build. Prototyping involves creating functional models that test users can try out so they can provide feedback about usability. The insights supplied by users allows you to make necessary changes to a design before investing in the full product build.

When designing a prototype, focus on the specific characteristics and functions you want to test. Additionally, it is important to have a clear understanding of what you expect to learn from the testing stage.

Alpha vs. Beta Testing

The term “market testing” refers to the process of testing how well a product or service performs among the target audience. This phase of the product development process involves two primary testing methodologies: alpha testing and beta testing.

Alpha testing is performed on-site by internal employees of the company developing the product. It is used to identify functional and operational issues that could become a problem down the line. It does not involve an in-depth review of product reliability and security.

This testing phase can employ both white box and black box testing methods, so that test users can review a product with or without working knowledge of the system. Compared to other testing phases, this phase may take longer. However, this is necessary as alpha testing serves as a means of identifying issues that may have been missed earlier in the design process.

Beta testing is conducted by users and customers who are not associated with the product development company. Testing occurs at the end-users’ location and involves only black-box testing, so testers do not have insight into the system’s inner workings. This testing phase focuses on the product’s reliability and security, with patches released as new issues are identified. Unlike alpha testing, beta testing requires only a few months to complete.

Tips for Successful Prototype Testing Operations

When organizing prototype testing operations, there are many factors to keep in mind to ensure they proceed successfully. Some of the key considerations include:

Prototyping testing operations can be moderated or non-moderated. Both methods offer unique advantages and disadvantages.

• Moderated test sessions allow you to be directly involved during the testing process, so you can assist testers or rearrange tasks as necessary. While you have more control over these tests, you are at risk of affecting their results if you help participants too much.
• Non-moderated test sessions allow testing to be completed by testers with minimal involvement from you. This method enables you to get more accurate results but also makes it more challenging to contact participants with follow-up questions. You can, however, make changes during the testing process in some non-moderated sessions and gauge the results.

If you are new to product testing, it is a good idea to perform a pilot test so you can get a feel for whether your market testing participants will have any problems. Pilot testing involves having someone within the company go over the test. If they have questions or seem confused, it is a good sign that your testers will find the test steps difficult to follow.

• Tweet

• Post

• Share

• Save

• Get PDF

• Buy Copies

• Print

Test markets are an established part of marketing folklore and an accepted piece of the marketing armory, but do we stop often enough to ask two basic questions: Are these tests always necessary? What circumstances led the company to initiate them?

The relevance of test marketing to an industry depends essentially on the type of manufacturing process involved. The confectionery industry, for example, can frequently produce sufficient quantities of the new product for test market purposes by minor modifications to existing plants and layouts whereas, in the event of a national launch, an entirely new plant may have to be built, which would necessitate a much heavier investment.

Test markets will obviously be inappropriate in industries where the technology requires the same sort of investment for the production of one unit as for a thousand, as in the case of airplanes and cars. No doubt, the British and French governments, for example, and certainly the British and French taxpayers, would have been more enthusiastic about the national commitment to Concorde had they been in a position to test market the prototype in advance.

Consequently, in deciding whether test marketing is a relevant aspect of marketing strategy, the scale of risk and investment implied by the marketing decision must be evaluated. The most visible side of the risk tends to be the cost of plant and machinery that are needed for the new product’s manufacture.

But new products also involve other substantial costs: the advertising expenditure; the amount of sales force time occupied by selling the new product; the goodwill and reputation of the company at the wholesale and retail level; the store shelf space required by the new product; and the production and rework problems that always accompany any new line start-up. If test markets enable large parts of these important costs to be deferred until a reliable estimate of national sales can be made, based on actual experience of the product in the marketplace, then test marketing is desirable.

The purpose of this article is to evaluate the uses and limitations of test markets and to describe the role of test marketing within its real context—the field of market research. My premise is that test markets are often employed to discover facts that could just as easily have been established by other cheaper and faster research techniques.

To Test or Not to Test?

Test marketing offers the marketing company two important benefits. First, it provides an opportunity to test a product under typical market conditions in order to obtain a measure of its sales performance. As well as enabling top management to make an accurate prediction of its potential national turnover, this information often forms the basis of the decision whether to extend the product nationally. So the importance of its accuracy is self-evident.

Second, it provides an opportunity, while the product is on limited sale, for management to identify and correct any weaknesses in either the product or its marketing plan before making the commitment to a national sales launch, by which time it will normally be too late—and certainly very expensive—to incorporate product modifications and improvements.

Decision factors

Despite these benefits, however, the decision to test market should never be routine. It is a costly, laborious method of collecting information about reactions to new products, and therefore a test market should be used only as a last resort. The more profitable route to follow with a new product—provided the risk is acceptable and the research sufficiently reassuring—is to launch nationally and avoid the costs and delays of a test market. Test marketing enables the company to minimize losses but not to maximize profits.

There are four major factors that should be considered in determining the efficacy of test marketing:

1. It is necessary to weigh the cost and risk of product failure against the profit and probability of success. For example, at Cadbury Typhoo Limited, though we have test marketed 24 products during the past three years, during this period we have also successfully launched 4 products nationally, but without utilization of a test market phase. In each case of launching nationally, I should stress that the costs and risks of product failure were low.

2. The difference in the scale of investment involved in the test versus national launch route has an important bearing on deciding whether to test. Of the products we have launched directly into national market, very little difference in manufacturing investment was called for whether we opted for a test or national launch. On the one hand, where plant investment for a national launch is considerable, but only slight for a test market, the investment risk favors the test launch approach.

At Cadbury, we have found this to be particularly relevant where, in association with foreign companies, we have been able to import sufficient product to cover test market requirements and so avoid any investment in manufacturing facilities. In the past three years we have test marketed four products that were made by foreign companies in existing plants abroad and that we tested at a price which assumed local production. Now, two of these test markets have provided us with sufficiently encouraging sales performances to justify the planning of local production.

On the other hand, by restricting a product to the test market area during the considerable period required for performance to be predicted accurately, there may be a high opportunity cost. This opportunity cost may amount to the foregone turnover and profit of one year’s national sales, depending on how long the product is kept in test.

3. Another factor to be considered is the likelihood and speed with which the competition will be able to copy your product and preempt part of your national market or overseas markets, should the test be successful. Competitors will be monitoring your test market, and where they have the technology, they will be developing their own versions of your product—and marketing it if you leave the opportunity open for them to do so. Within two years of the start of Cadbury’s successful test market of a children’s chocolate line (Curly Wurly) in the United Kingdom, we have seen identical competitive versions of the product appear on Canadian, Japanese, West German, and U.S. markets.

4. Apart from the investment in plant and machinery that may be involved, every new product launch is accompanied by a substantial marketing investment that varies with the scale of the launch. New product launches call for heavy advertising and promotional expenditure; they require sales force time, attention, and effort; and they need shelf space in wholesale and retail outlets, which is sometimes obtained only at the expense of the space already given to the company’s existing products.

Moreover, if a new product fails, the costs of rebating and reclaiming unwanted stocks from customers have to be faced, along with those costs of writing off unwanted and unusable materials and packaging. Top management should also take into account the possible damage that a new product’s failure can inflict on the company: its reputation in the eyes of consumers and customers may be blemished, which is a real if not quantifiable danger.

The foregoing marketing costs—or risks—are reduced by limiting the new product launch to a test market. The cost of concentrating sales force priorities on an unsuccessful new product, and of allowing profitable existing products to lose some share of market as a result, can be greater than the more visible cost of a piece of unwanted machinery.

Our company recently undertook a successful preemptive launch in the soup market. We had identified an opportunity for a powdered snack soup product in the United Kingdom after having observed the success of this product category in other markets, particularly in the United States. Because we had no franchise in the conventional soup market, we faced potential competition from at least four aggressive and highly respected soup manufacturers with strongly established product ranges. Since one of these manufacturers had already test launched a snack soup in the United Kingdom, we realized that the only chance to successfully launch our brand would be to preempt the traditional manufacturers by introducing, distributing, and promoting our brand nationally and being first on the market.

We are currently brand leaders in this segment of the soup market, and an important factor in our success was our recognition that as a nonsoup house we could afford neither the time to test market nor the risk of being swamped by the established competition.

Obviously, there must be some compelling reasons to persuade management of the wisdom of bypassing a test market and moving directly to a national launch. In the case of our entry into the soup market, we felt justified because (a) speed was essential, (b) research had shown the product to be highly acceptable, (c) we judged the package design and advertising to be suitable for the product, and (d) we were involved in only limited manufacturing investment.

Nevertheless, we continue to believe that test marketing is indispensable in order to minimize the major risks involved in most new product launches.

New Product Research

During the past 10 years our Foods Group has test marketed 13 products, of which 7 have been extended nationally and now account for 33% of that group’s turnover. During the past 4 years our Confectionery Group has test marketed 27 products, of which 17 have gone national and will comprise approximately 15% of the company’s 1975 confectionery turnover.

I am not suggesting that these success/failure rates are either good or bad or that comparisons with other manufacturers’ experience will be meaningful. New product activity and test market performance can only be judged within the total marketing policy and strategy of a particular company and will vary widely between markets and product categories.

To illustrate, confectionery is an indulgent, fun, lighthearted product category, and new product launches provide novelty and variety to sustain consumer interest and create awareness. By contrast, the grocery food market tends to be more conservative, since a nation’s eating habits change slowly. This combination probably explains why, for example, the major innovation in the U.K. bread market in the past 20 years has been the sliced variety.

I would only say in assessing our own performance that had we spent more time researching new product candidates, using all the research techniques we now employ, before we made the decision to test market, we could have identified and eliminated a number of failures before the test market stage and so saved ourselves considerable expense and management time.

Before the decision to test market a new product is even considered, we believe that to properly evaluate it, it is important to research all aspects of the product in the form in which it will ultimately be presented to the public. In other words, we want to research not only the product itself but also the name, package design, advertising, and price. The competitive environment that we envisage for the new product is also important in the research.

Using the following total package of research techniques, we are confident that we can expose any major negatives in the package, and either correct them where feasible or abandon the project prior to embarking on the costly test market venture. Here are the steps we follow:

• Select a sample of respondents according to their claimed buying habits.

• Show respondents the introductory advertising that is planned for the new product and get their reactions to it.

• Show the respondents the product and get their pretrial reactions to it.

• Give the respondents five samples of the product to test at home.

• Conduct an in-home interview to ascertain respondents’ (a) posttrial opinions of the product, (b) recall of the advertisement, and (c) claimed intended frequency of purchase.

In addition to these steps, another useful research tool can help to further reduce uncertainty: the model test market. This is a stage of research interposed between pretest research and a full test market. Model test markets simulate the appropriate marketplace in an attempt to reduce both the time necessary to properly evaluate a full test market and the expense involved.

Model test markets are particularly useful in situations where research results have been inconclusive, where considerable capital investment is required to set up even a test market, or where a speedy evaluation of a product’s potential is necessary. Model test markets are currently a fashionable area of research development and are becoming increasingly useful as their accuracy and sophistication progress.

The model or mini test market that we use involves a representative panel of housewives who receive a weekly visit at home from a van salesman with a mobile shop. The visit is preceded by the mailing of a sales catalogue and an order form that features the products being tested, along with all the leading brands and any promotional support that is either current or being tested.

There are other types of model test markets but their differences lie in the way the actual marketplace is simulated rather than in what is measured (usually trial, repeat, and frequency of purchase). Since it is the purchasing situation that is simulated, the alternative formats are limited and usually center around different ways to restrict the purchasing of the samples to a medium that the company is able to monitor on an ongoing basis.

Limited capabilities

It would be wrong to suggest that new product research can supply all the answers, or that sales estimates based on the research findings are accurate enough to make the test market stage redundant. A number of factors differentiate research from real-life situations and therefore act as constraints on the degree of predictive accuracy possible prior to gaining test market experience. These factors have to be considered when using the results of research to predict a product’s performance.

In research (a) the market in which the new product operates has to be artificially defined and limited, (b) the respondents are acting under enforced 100% trial and repeat trial, and (c) the time factor is artificial, so that management either selects only heavy users of the product or enforces artificially heavy consumption upon lighter users.

Although each new product and test market has its own features, our experience has led us to draw some general conclusions about which product characteristics can be evaluated in research and those which cannot. Specifically, research can indicate the acceptability of recipe, package design, name, price, size, and advertising. But research cannot take the place of a test market when trying to assess the frequency of purchase, novelty value, and extent to which the new product will substitute for sales of existing company products. Also, it cannot indicate the role, character, and position that the product will develop in the minds and purchasing patterns of consumers.

The more unusual or novel the new product, the more difficult it is to assess its future prospects at the research stage. The two most successful products developed by a Cadbury company during the past 20 years have been Smash instant potato and Marvel instant dried skimmed milk. The introduction of these products created quite new markets and caused our competitors to launch similar products.

Yet, when presented with a totally new product concept, respondents cannot predict at the research stage the precise role and position the product will later assume. Smash and Marvel are now regarded by U.K. consumers as universally acceptable and desirable convenience foods. But when they were in the research stage, instant mashed potatoes and skimmed milk evoked memories of wartime substitute foods and images of slothful housewives looking for the easy way out, not taking the time and trouble to prepare the real thing for their families. The marketing success and brand acceptance later achieved by these products could never have been forecast from research alone. Research could not indicate that the two brands would totally alter consumers’ perceptions of the product categories.

Forecasting the degree of substitution a new product will have from pretest research presupposes a level of knowledge of consumer behavior in the marketplace that marketing managers do not yet possess. The only accurate way of assessing substitution is through a test market. This need was an important factor in our determining recently to launch a second premium-priced liquer-flavored chocolate bar. We had developed the market ourselves with a rum and raisin chocolate bar two years before. Thus there was no benefit to Cadbury in launching the second product unless it successfully expanded the total market, not simply leaving the company with the same market shared by two products.

In a test market we were able to show that the degree of substitution could be kept to an acceptable level. This was subsequently verified when more than 50% of the new product’s sales came from market expansion.

In a highly developed and saturated market like confectionery, new products seldom increase overall consumer consumption, so to be successful they have to reduce sales of existing products. It follows then that a confectionery manufacturer needs to satisfy himself as far as possible that his proposed launch will substitute for his competitors’ products rather than for his own.

Measuring Test Performance

To assess performance in the test market, management needs to know not just the volume of sales, but also the nature of those sales: where they are coming from, and what levels of retail distribution are being achieved.

It is necessary to agree on standards or targets for all these factors before commencement of the test market. The real difficulty of this is the way these standards vary between products and markets. Nevertheless, a manufacturer ought to know his markets, and through the establishment of norms he ought eventually to be able to determine satisfactory criteria with which to measure his test market performance. At Cadbury, we have well-defined objectives for test markets that relate to volume of sales, trial, repeat, and frequency of purchase, distribution, and brand share (where appropriate).

Important criteria

The sales budget or target in the experimental region will be directly related to the level of national turnover that would justify a national extension of the new product. Performance against the sales budget can be measured simply and regularly from the weekly sales receipts. There is always the danger that the sales total in the experimental region will be inflated by exporting to other regions. While this cannot be prevented altogether, it should be kept to a minimum. Otherwise an inflated assessment of test market performance will occur.

The source of sale is important in indicating the future level of settled-down sale. To assess the source, we audit the experimental region every two months after launch to check consumer awareness, trial, and repeat purchase of the product. This audit reveals whether the advertising is motivating people to try the product, and—most important—whether having tried the product, they buy it again.

The repeat purchase figure is the key to the level of settled-down sale. If the product does not match the consumer expectation created by the advertising and presentation, repeat sales will be disappointing even though early test market results have indicated success.

For this reason there should not be too much dependence on the results observed during the early days of a test market. High sales reflect high trial purchases, while a low initial takeoff may simply signal that the advertising lacks impact, something which can usually be corrected during the test. Satisfactory settled-down sale will only be achieved if a large enough core of regular purchasers can be gained for the product, and naturally, this will not be apparent during the early stages of the test market.

It is important to establish what share of the market a new product is acquiring so as to assess its relative performance and to confirm the size of the market being competed for. With a product named Cheers, we overestimated the market for cold milk additives in the United Kingdom by assuming that the growth observed before our launch would continue. In fact, it did not. The market leveled off, and our volume expectation for the product was, with the benefit of hindsight, unrealistic. Our mistake had been to overestimate the future size of the market sector into which we were launching.

Currently, through subscribing to a retail audit every two months, we are able to monitor market share performance, market sizes and trends, and—within that framework—the performance of our own product in terms of share, shop and sterling distribution, purchasing distribution, sales per point of sterling distribution, retailer purchases, consumer sales, and stock cover. “Shop” distribution, expressed as a percentage of the total number of potential stockists, is the number of shops stocking a product; “sterling” distribution measures the share of total retail sales held by those retailers stocking the product. So, if Brand X is in 50% of grocery outlets and those outlets have 75% of total grocery sales, Brand X will be described as having 50% “shop” and 75% “sterling” distribution. For the United States, simply read “dollar” for “sterling.”

The foregoing combination of data can be extremely useful in tracking the development of the new brand and in identifying problem areas associated with it. In addition, special analyses can be helpful in pinpointing marketing problems associated with, for example, age of stocks, cumulative distribution, pricing, display, and merchandising.

Incidentally, the action standard we set for new grocery products in terms of sterling distribution is 65% nationally, but this standard will vary among product categories.

As I have already indicated, our leading share of the confectionery market means that we are vitally concerned with the substitution or cannibalization factor any proposed new confectionery product will have on our existing range. The objective of our new product program is to win share from our competitors. Thus to judge the value of a new product, we need to measure the performance of existing products (our own and competitors’) in order to assess whether the new product sales are coming from ourselves or the competition. The results here can have an important bearing on management’s final judgment of the test market.

Distribution targets for major categories of potential stocklists are agreed on in advance with sales force management. These targets must be realistic if they are to have sales force commitment. And so we set them to peak after 12 months. The time it takes to achieve full distribution will depend on the size and effectiveness of the company sales force, and on the nature of the particular market’s wholesale and retail distribution pipelines.

Selection of Test Region

Since more than 90% of Cadbury’s media expenditure is on television, the eight TV boundaries in the United Kingdom determine our test regions. None of these regions conforms to the national pattern of food and confectionery consumption and expenditure. Because we cannot pick a representative region, the important thing for us to know is by how much each region varies from the national pattern. With this knowledge we can develop a behavioral factor for each region that we can consider when extrapolating national sales. Our experience shows that when differences in purchasing patterns between regions are neglected, an error of more than 50% can result—certainly sufficient to present a quite inaccurate picture of subsequent national outcome.

While we have not used two test regions for the same product, such a course would provide more evidence on which to base final predictions, though it would involve increased marketing, and possibly production, investment. The use of two test regions could be a particularly appropriate route were it intended to test two quite different marketing strategies for the same product (e.g., aiming the product exclusively at children in one region and at mothers for their children in another).

In summary, our choice of test region is governed by logistics rather than the individual characteristics of any region. The determination of our final decision is based on factors such as our own existing marketing activity, that of competitors, availability of stock, seasonality, the size of the test area in relation to the budget set for the product, the structure of the trade in the selected area, and the degree of anticipated retailer cooperation.

Control of experimental market

Care must be taken to avoid distortions in test market results. The conditions in the experimental region, for example, should resemble as nearly as possible those under which the product would be launched nationally. Even though any new product should receive priority from the sales force and intensive marketing support during the launch period, the support in the test area must be of the same weight as could be afforded on a national basis. In spite of the natural temptation for marketing managers to seek disproportionate attention to sales and marketing support for their test launches, such temptation has to be resisted: the research or forecasting manager must make an objective prediction that assumes normal support.

While top management can control the company’s activity in the experimental area, competitive activity unfortunately cannot be so regulated. The important thing is to allow for any unusual competitive activity when assessing test market results.

Other distortions may be caused during the test market by stock shortages of the test or competing products, or by modifications made during the test to the product or its marketing plan. It is consequently important that the test market run long enough so that the effects of these distortions and changes will either have disappeared or made their full impact on the level of settled-down sale.

Length of test run

One question frequently asked is: How long should a test market run? The answer will depend on the length of time it takes to judge the product’s performance against the company’s objectives for it. Time must be allowed for sales to settle down from their initial honeymoon level; in addition, the share and sales levels must be allowed to stabilize. After the introduction of a product, peaks and troughs will inevitably stem from initial consumer interest and curiosity—as well as from competitive product retaliation.

Sufficient time should be given to iron out any deficiencies in either the product or the marketing program. If the advertising is not communicating, then new advertising will have to be developed and implemented if the product is to be given a proper trial. If the package design lacks impact, this too can probably be corrected during the test.

Time must also be allowed for planned levels of distribution to be achieved. To illustrate, Exhibit I gives the distribution figures reached during test market by two new Cadbury confectionery products. One was subsequently judged successful and was therefore extended; the other failed to meet the criteria laid down and was later withdrawn.

Exhibit I Levels of Distribution Achieved for Two New Confectionery Products

The reader can see that at least for the first six months there is no appreciable difference in the relative levels of distribution achieved, but thereafter distribution of the unsuccessful product begins to tail off. While some clear success and failure can be identified sooner, we place little weight on test market results achieved during the first six months. After that, the pattern of repeat sales begins to establish itself and accurate predictions become possible.

Why Products Fail

The reasons for new product failure in a test market are normally complex and not easy to identify. A product does not usually fail for one reason alone but rather through a combination of factors. It is nevertheless true that the majority fail because the product does not contain a demonstrable consumer advantage over competitive brands.

An analysis of 18 of our own new product failures reveals the following major reasons for lack of success, together with their relative incidence:

In the confectionery market as in other food markets, recipe is a key variable and, not surprisingly, it is the most important factor in achieving success or failure. Conceptual shortcomings refer to the overall positioning and advertising of the product. I have not identified advertising as a separate category, because the best advertising execution in the world will not sell a product if it does not meet and satisfy an existing or potential consumer need.

Price determines the value the product offers to consumers, and it is never easy to judge the correct balance between profitability and the need to be competitive. Nevertheless, there is no purpose in test marketing new products at a price that will not meet the company’s profit objectives. Thus the argument that a new product must be launched at a price below the competition is a dangerous one, because it simply delays the time when the product must be priced to produce satisfactory profits. If a product meets a need, consumers will pay for it.

The package (i.e., its size, composition, and design) is another criterion that plays a key part in determining the acceptability, visibility, and impact of a new product in the marketplace.

Illustrative failures

Here are four examples of mistakes we made that resulted in favorable research findings but led nevertheless to new product failure in the test market:

1. In the case of a new biscuit product, we were never able to match in commercial production the product samples which had been made up in the R&D laboratory for the research tests. The product we test marketed was neither the same nor as good as the product we initially developed, and the sales in the test market never came up to expectation. The lesson is a simple one: be sure you are researching with samples identical to the product you plan to market.

2. As a result of our success with Smash instant potato, we developed an instant croquette potato under the brand name Smasher. Our home economics unit had prepared the samples in shallow pans during research sessions, and housewives had not given any indication that they would have difficulty making up the finished product.

However, it turned out that the majority of housewives did not possess the appropriate shallow pans that our well-equipped home economics unit had been using, and the product was very difficult to prepare in the deeper saucepans found in most British kitchens. Therefore, despite the good rating the product itself had received, difficulty of preparation torpedoed it.

3. We developed a fresh cream product called Swiss Dessert, which received excellent ratings from consumers in research. The product was tetra packaged, which preserved the potentially unstable dessert safely, but the new package was impossible to stack and difficult to display. Thus it proved to be unpopular with the grocery trade and never received adequate support in the test market area.

4. We developed an instant tea under the brand name Fine Brew, which—prior to the advent of the tea bag—had important convenience advantages over the conventional pot-brewing ritual. The product researched well and we proceeded into test market, where the product proved singularly unsuccessful. The concentration of the tea was so strong that half a teaspoon was sufficient per cup, and so small a measure made preparation difficult. It is also possible that we underestimated the challenge of weaning the British public from the traditional teapot.

Future Developments

Marketing companies will only improve their test market strike rate and predictive accuracy when they develop ways of more accurately identifying and recognizing the characteristics of consumer behavior and patterns of consumer expenditure which shape our markets month by month. Marketing knowledge in this area is inadequate.

Thus the need for test marketing in the future will be primarily governed by developments in marketing knowledge and understanding of the mechanics of consumer behavior. The better top management understands its markets, the less frequently will test markets fail, and the sooner it will be able to risk dispensing with test markets altogether.

At Cadbury we are learning much from the correlation of pretest market research data with the launch sales patterns which subsequently develop. This information is revealing identifiable differences between the brands that succeed and those that fail in test market. If successful and unsuccessful patterns can be established and recognized at the pretest stage, we will not only be able to abandon at an earlier stage the projects likely to fail, but perhaps we can also reach the point where the test marketing of new products becomes redundant.

However, that day is a long way off, and our research techniques—whether they consist of total package tests, extended in-home placement tests, mobile shops, or controlled warehouses—will all have to be developed considerably before an acceptable substitute for test marketing can be claimed.

Of equal significance to future developments is the use and analysis of consumer panel data. These data give us a record of consumer purchasing and consumption behavior, and they can be enormously valuable in providing information from which we can learn how consumers do in fact behave. Most definitions of markets are manufacturer-oriented and relate to sales of the product rather than to the consumer who makes the buying decision. For example, we talk of the boxed chocolate assortment market, when boxes of chocolate are also competing in the consumer’s mind with flowers, records, books, and other semicasual gifts.

Market definition can have an important bearing on the way consumer research of a new product is structured. We need to know which consumers are potential purchasers of our brands, how many of them there are, how frequently they buy and switch brands, which other brands they buy, and how best we can communicate to them. If we knew more about our consumers and their buying behavior, we could make our test market, and marketing activity generally, more accurate and less wasteful.

Meanwhile, there is a knowledge gap between pretest market research and the national launch, and for now it can be bridged only by the test market. Though existing research techniques are useful, none can replicate the market situation, and this is the unique contribution of the test market. For this reason test marketing will continue to be a most important marketing tool and gatherer of information for marketing decision making.

A version of this article appeared in the May 1975 issue of Harvard Business Review.

What are the steps in marketing testing?

What is the role of market testing for a new product?

What are the two broad product categories?

What are two key characteristics of the introduction stage of the product life cycle?