Says Kai Adolphs, Global Product Research Director, TNS’ Automotive Practice
Price management is a hot topic for car manufacturers today, and for good reason. Setting prices wrongly initially can waste millions over the vehicle lifecycle, for example by using major sales promotions and discounts to correct overly ambitious price and volume goals.
Ultimately, car manufacturers have a growing interest in learning how to set the price correctly, not only in highly competitive saturated markets of Europe and North America, but also in the rapid growth markets where the “gold rush” is starting to lose its luster!
It’s no wonder that market researchers and consultants nowadays offer so many tools and concepts (of varying degrees of sophistication) to support car manufacturers in their pricing decision-making. Most of these tools try to forecast market shares and sales volumes because car manufacturers increasingly demand precise answers about the volumetric potential of their vehicles. If I asked car manufacturers, market researchers and consultants what tools they really need, many would answer: “conjoint”. But is it really the answer?
Refined tools are needed for pricing research
Conjoint analysis has been a popular approach for many years and most researchers and consultants trust it almost blindly. But I strongly believe that today it can no longer be considered the ultimate answer for pricing questions. Particularly not in its traditional form for the production of realistic market shares or volumetrics for vehicle models. Let me explain why not.
- Conjoint used to be popular for its ability to cope with a large numbers of attributes. But the downside is overly complex designs and highly artificial choice scenarios. In fact, decision makers don’t follow the statistical rules of design efficiency, but use their individual and often emotion-based heuristics.
- Market differences in brand/model awareness, dealer network size, dealer performance, sales promotion effectiveness and discount levels etc. are often ignored in old-school conjoint models, mostly because preference-centered analytics did not find a way to incorporate these market factors effectively.
- Most research settings reveal the full market picture to respondents. But in reality this level of transparency does not exist at customer level. Customers have very different knowledge levels for depending on the brand, model and price. This has a major impact on the natural volume potential of the competing vehicle offers.
The black box
Market researchers and consultants with little experience in the automotive industry would solve the issue by building a ‘black box’ around their data model, replacing market intelligence with mind-numbing data crunching until something resembling market shares falls out.
Pricing strategies call for precision
The reality is that today’s pricing strategies need to be based on precision, not imprecise foundations. Conjoint traditionally concentrates on consumer preference. Key measures of business growth, like market share gains, incremental sales volume etc. are not what these methods have been designed to deliver, so don’t meet client needs. Pricing managers and volume planners want to understand the potential market opportunities with quantified premises. They want to play “what if” scenarios within the forecasted volumetric!
It is impossible to transform relative preference from a conjoint model into business growth measures purely by using mathematical algorithm. Accurate volumetric predictions call for the consideration of context, for example by a systematic cross-linkage of industry standard secondary data (e.g. Jato, NCBS, Global Insights) with a flow of proven primary research elements to complement the conjoint.
Also the use of “conjoint” itself needs rethinking. New approaches, e.g. discrete choice models and self-balancing solutions are much better suited to design interview settings that reflect real life decisions of human beings.
This way, pricing research can deliver precise and powerful insights which are far removed from a “calibrated” black box.