Newsletter 01/2013 - Developing and implementing product strategies using conjoint analysis

With conjoint analysis, plant and machinery manufacturers can examine the benefits of their products for their customers and then go on to develop much more precise specifications in future product development phases. The results of such analyses also allow quantitative forecasts to be made about the market success of the product. This makes profitability calculations for product development projects much more useful and more realistic.

Read newsletter offline?

Download here

Customers demanding new machine generations, competitors putting their hard-won market position at risk, Chinese manufacturers becoming a challenge even in traditional markets – these are just some of the many reasons why plant and machinery manufacturers might want to adapt their products or even design a completely new generation of machines. In order to weigh the costs of a product development project against the benefits, you need a development budget and an idea of what sales volumes, prices and margins are realistic. However, product marketing and sales organizations often find it difficult to determine a product's future positioning in the market or to define clear specifications regarding the few really important product attributes, not to mention making any quantitative predictions about the scale of market success. Faced with this situation, conjoint analysis is the ideal tool.

A conjoint analysis can be used to determine just how much money a customer is prepared to spend on certain performance attributes, such as output measured in parts per hour (say, 4,000 parts/hour or 6,000 parts/hour), retooling times or delivery lead times for spare parts. Knowing the extent of your customers' "willingness to pay" for aspects like these enables you to describe the optimal specification of a piece of plant or machinery for a certain requirement spec and determine the target costs for individual modules and components, define the performance attributes and align the product development process accordingly.

Proven method for analyzing consumer opinions

Conjoint analysis, short for CONsidered JOINTly, was originally a concept from the field of psychology. Today it is the most frequently used analysis method for collecting consumer preferences. The decompositional approach of conjoint analysis enables you to examine the extent to which users prefer individual attributes and/or combinations of attributes within a given product.

The most important alternatives to conjoint analysis are pricing methods such as the Van Westendorp Price Sensitivity Meter or the Brand-Price Tradeoff (BPTO) method. However, both of these approaches treat each product as a whole; unlike conjoint analysis, they do not enable product attributes to be examined individually. The Maximum Difference Scaling (MaxDiff) methodology interprets the utility value of product attributes equally – conjoint analysis, on the other hand, determines the value of each individual product feature separately. The use of prices is also significantly restricted in MaxDiff compared to conjoint analysis.

A newer method, which is currently costlier than doing a conjoint analysis, is Menu-Based Choice (MBC) modeling. Although quicker to perform, MBC is still in the development stages. Conjoint analysis has proven to be the ideal method for the issues we deal with in our consulting projects with plant and machinery manufacturers.

To perform a conjoint analysis, we ask our client's customers, potential customers or even employees to answer a questionnaire online (which takes them approx. 10-15 minutes). The respondents evaluate products with different combinations of attributes that they could be offered in the real world.

The different products have both advantages and drawbacks from the point of view of the respondents. A conjoint analysis makes respondents conscious of the actual importance of the individual attributes so that they can weigh them against each other.

The combinations of attributes within the evaluated products are decomposed in the mathematical analysis and the importance of the individual attributes – as well as which attributes respondents prefer – are calculated. As such, the conjoint analysis largely mirrors the actual evaluation process that takes place in a real purchasing situation, where the consumer is likewise faced with a product in its entirety.

Identifying customer needs and drawing conclusions for product development

Conjoint analysis gives you information about which are the most important product attributes (e.g. a machine's cycle time) and which version of these attributes the customer prefers (e.g. a cycle time of 3.1 seconds; see the thick, dark red line under "Worldwide" in Fig. 1). A high preference also means that the customer's willingness to pay is highest for this attribute. In the real-life example illustrated in Fig. 1, consumers showed a positive preference for 2.8 seconds as the cycle time, but their preference – and thus their willingness to pay – was found to be even higher for a machine with a cycle time of 3.1 seconds.

This information is invaluable for technical development, because what it shows on the one hand is that it would make little sense to invest in reducing the cycle time to 2.8 seconds. And on the other hand, it tells you that the cycle time should never exceed 3.4 seconds, as customers showed a "negative preference" for that, in other words, they expressed dissatisfaction with that cycle time. In the project we conducted, these were clear signposts that the machine manufacturer was able to use to optimize their product.


Figure 1: Customer preference functions for machine cycle time

The preference functions can also be calculated for distinct customer groups. Our project demonstrated that all customers outside of Western Europe rated a cycle time of 3.1 seconds even more positively than customers in Western Europe. They were therefore more interested in the precise cycle time and therefore had a higher willingness to pay when the cycle time was at this point. As a result of this knowledge, the sales department was able to offer the machine with this performance attribute at a higher price outside of Western Europe – and was thus able to improve the margin without the company having to invest a single euro in technical development.

Western European customers, by contrast, were more willing to pay for a machine with a cycle time of 2.8 seconds than for a machine with a cycle time of 3.1 seconds. A more detailed sensitivity analysis was therefore necessary: we used a simulator to determine how many additional machines with a cycle time of 2.8 seconds could be sold. The result of the analysis showed that the incremental margin justified the manufacturer's investment in the further development of the machine.

Predicting the market success of a product

After analyzing customer needs, you can simulate your market and test the success of different product configurations. In doing so, you will take into account your own current products, competitors' products and new products you could develop in the future. A simulator calculates the theoretical demand for the different product configurations. For example, if a current product has a cycle time of 3.1 seconds, the effect on your sales volume can be calculated by reducing the cycle time to 2.8 seconds. After your R&D team has estimated the investment required to upgrade the machine to 2.8 seconds cycle time, this cost is offset against the additional revenue to determine whether the investment would pay off. Fig. 2 shows how the product strategy of an engineering firm can affect its market share.


Figure 2: Theoretical market shares, based on the results of a conjoint analysis [in %].

By replacing four brand I products (B, C, D and E) with two new products (a and b) and developing one additional product for brand II, a theoretical doubling of market share (from 20.8% to 43.8%) can be anticipated for the two brands. The simulator also makes it clear which competing products will lose market share (7.2% in the case of G-a, for example) and how many customers (8.5%) of those who cannot currently find an acceptable offer on the market (33.3%) the engineering firm could convert into new customers. Predicting how market share could shift after product launch allows the company to conclude which would be the most promising combinations of product parameters – and this in turn enables them to define clear product development (and brand) strategies.

The experience to enable success

Androschin & Partner regularly supports plant and machinery manufacturers as they seek the "right" product strategy using conjoint analysis. Very often, our work on developing the product strategy goes as far as developing the specifications themselves. In the last two years, for example, we have carried out projects in the fields of plastics machinery, packaging technology, machines for the pharmaceutical industry and machines for components (e.g. coolers). Some examples

  • A manufacturer of plastics machinery was in the process of developing a new machine but was not sure whether the design was in line with market requirements. A conjoint analysis with 40 respondents was carried out to determine the machine size that provided customers with the greatest benefit. The specifications, which had previously specified a machine that was much too large, were changed and the newly developed machine was a success.
  • A manufacturer of packaging machines wanted to optimize their entire product portfolio. We advised them on optimization measures for 11 products, all based on conjoint analyses with a total of more than 600 surveys collected.
  • A manufacturer of machines for the pharmaceutical industry had decided to develop a brand new machine generation. Within the product conceptualization phase we arranged a conjoint analysis to determine the optimal product specifications as a basis for the new developments.

Of course, when you seek to define the optimal product concept and the future product positioning in the market within the framework of your overall corporate strategy, you need more than just the bare results from the statistical data of a mathematical model alone. That is why we discuss the simulation results with the sales department in a workshop setting, consider the effects of different product development concepts and debate the future positioning of the products within the competitive context.

Only once this is completed are decisions made about what actual R&D work should be done. It is also important to involve the technical development team at this stage in order to be able to estimate the cost of the development projects and thus calculate the return on your investment.

Product strategy in three months

How long it takes to prepare the project, define the product attributes you want to examine, program the online survey and evaluate the answers depends heavily on the commitment of the employees within your own firm. Preparations are usually completed within one month. After the survey is conducted, another month should be planned in for the evaluation of results and the definition of product development measures.

Once you have statistical data on customer needs at your disposal, you can use it later for other purposes. This gives you market intelligence that can also be a valuable tool for your product marketing activities in the long term.

Coaching from an experienced consultant is an essential success factor in preparing a conjoint analysis efficiently, but it is also a crucial aspect in being able to carry out the "right" analyses and evaluate the results in a way that is useful for you as an engineering firm. Androschin & Partner has many years of experience with conjoint analyses and works with market-leading partners for statistical analysis methods.

Contacts for product strategies based on conjoint analysis:

Christian Androschin, Partner,
Norbert Rauh, Partner,
Alexander Ludorf, Projektmanager,