Last week we addressed part one of trade-off analysis and explored several use cases for MaxDiff. In this post we will explore a variation of trade-off analysis – choice-based conjoint.
For somewhat more complex business objectives such as product optimization, price sensitivity or willingness to pay analysis, choice-based conjoint is flexible yet robust enough to answer these objectives.
Problem: Product Optimization & Price Sensitivity
Solution: Choice Based Conjoint (CBC)
Product optimization can be a very complex task and understanding cannibalization and incrementality when the new product is launched is critical. Choice-based conjoint (CBC) is most commonly used and is generally preferred for product optimization, and when paired with ENGINE’s custom online simulator, the analysis can provide valuable configuration recommendations while accounting for competitive context. CBC is utilized primarily to determine what combination of a limited number of attributes is most influential on respondent choice or decision making.
Quantitative interviewing is conducted to determine the most critical product/service “features” (e.g. price) and then “levels” (e.g. $50, $80, etc.) within a feature to test in the conjoint. The exercise involves a series of trade-off screens where respondents are asked to select their most preferred configuration. The configurations will be presented to respondents and represent all possible combinations of attribute levels. In this way, responses to a relatively small number of scenarios can be used to predict the effects of many configurations. By showing a controlled set of product configurations in a simplified choice-based exercise, we achieve several advantages:
- Mitigated scale biases
- Greater discrimination
- Easy exercises for respondents as humans are much better at judging items at extremes rather than discriminating among items of middling importance or preference
- Suitable for use across a wide range of respondents – differing ages, education or culture
- Determine the implicit valuations of the individual elements that make up the holistic product configuration, which can be used to simulate endless what-if scenarios. It can also be used to understand price perceptions at the brand level.
CBC methodologies are not only used for product optimization objectives and simulating market situations, but are flexible to accommodate a wide variety of designs from complex optimization scenarios to a simple price sensitivity analysis. Providing a “None” option for respondents if the products displayed do not meet their needs offers a more true to market situation that respondents can choose to not buy at all based on what has been presented.
Testing price sensitivity within a choice modelling exercise such as CBC is ideal because of the forced choice nature of analysis. Respondents are simply choosing between the configurations presented rather than providing an evaluation on a traditional scale. Within CBC we are also simulating a shelf set where other branded products are options as well as a “None” if none of the options are suitable to the respondent. Backend simulations also provide valuable insights for optimizing price across the shelf set, determining cannibalization or incrementality across various tiers of products and determining willingness to pay for premium features.
Problem: Product Optimization & Price Sensitivity on Products in Certain Categories or Industries Where There is Reason to Believe a “None” Weight Could Be High
Solution: Choice Based Conjoint (CBC) + Dual Response None Methodology
Typically, we include a “None” choice for respondents to tell us they would NOT choose any concepts in the choice tasks which allows respondents to avoid making uncomfortable choices. However, there are drawbacks to including the “None” option on the same screen as the other configurations:
- Respondents sometimes feel they are being helpful by picking a configuration rather than the “None” even though they would never consider that option in the market
- Can promote “click through” responses to complete the exercise
A Dual Response None methodology moves the “None” option to a follow-up screen after each choice task to ask whether the respondent would purchase the respective configuration on the previous screen. The benefits of moving the “None” to a separate screen allow us to:
- Learn about preferences (even if all alternatives are poor) with greater precision
- Increases the utility of the “None” (“None” threshold is higher), resulting in larger share of preference for “None” in the market simulator, which could potentially be more realistic than the traditional “None” weight depending on industry and the specific exercise/product being evaluated
- Medical Insurance or Necessary Products/Services – Likely more realistic None weights
- Impulse or “Nice To Have” Products/Services – may benefit from DRN but traditional “None” methods have worked successfully
However, there are some disadvantages to implementing a dual response “None” methodology within a CBC exercise:
- It does take longer – twice as many questions to answer
- Still doesn’t necessarily lead to accurate “take rates”
With the advantages and disadvantages explained, the ultimate decision of which CBC “None” methodology to implement is – it depends. It is best to consult with ENGINE and their Data & Analytics team on a case by case basis to determine the best fit for your business objectives.
In conclusion, trade-off analysis has many advantages to mitigate scale biases and handle unique market situations. Trade-off analysis can be very flexible and designed to accommodate many business objectives as well as budgets.