Conjoint Capabilities 2015
- 2. Conjoint Analysis What is a “conjoint” and what can it do? • Conjoint methodology models human selection behavior and mirrors how individuals consider jointly-grouped attributes of a product or a service – the process they go through when they purchase • With a conjoint, we can determine how much value individuals place on each attribute or element inside of a theoretical bundle • Better/nearly-perfect products or services can be built by combining the most highly-valued attributes • Conjoint exercises also show what features consumers might be willing to trade away in order to get the features they really want • Study data can be used to create market simulations that predict the impact of a new product offering or price changes on existing market share www.vernonresearch.com 2
- 3. Conjoint Analysis www.vernonresearch.com Excellent for: • Determining the best combinations of product or service features • Setting the best pricing for products or services • Gauging brand equity • Understanding a new product’s impact on the current market, including potential cannibalization issues • Finding utility values for all features tested, even those which may be hypothetical or merely being planned • Forecasting sales in future markets 3
- 4. Conjoint Analysis www.vernonresearch.com Setup for conjoint: • “Factors” and “Levels” are determined – Factors are general areas for examination – Levels add detail to the factors and should reflect the range of possibilities for the product or service • Different brands and price levels are often included • Factors should reflect the most important parts of the product or service 4
- 5. Factor/Level Example www.vernonresearch.com Factors Levels Manufacturer None- Generic Whirlpool Frigidaire Kenmore GE Kitchen Aid Style Freezer on top Freezer on bottom Side-by-side French doors Available Colors White White and off-white White, off-white and black White, off-white, black and stainless Additional Features None Icemaker in door Filtered water dispenser TV monitor in door Warranty None 1 year 3 years 5 years Price $750 $1000 $1250 $1500 $1750 $2000 $2250 5
- 6. Choice-based Conjoint (CBC) www.vernonresearch.com • Respondents choose which bundle of random features they would purchase, all else being equal. • Respondents typically complete 8-12 sets of choices, with 3-4 product bundles per choice. • The maximum number of factors is generally 6. • The maximum number of levels per factor is determined by how large the sample will be and the number of interactions necessary between levels. 6
- 7. Adaptive Conjoint Analysis (ACA) www.vernonresearch.com • Instead of choice between product bundles, Adaptive Conjoint measures respondent ratings between individual levels and groups of levels. • This method is good for exercises with a large number of attributes/features. • This method is also good for examining new product categories. • This method is not suited to examine price sensitivity. 7
- 8. Adaptive CBC (ACBC) www.vernonresearch.com • ACBC leverages the best aspects of CBC and ACA. • ACBC is an interactive experience, customized to the preferences and opinions of each individual. • This method is based on solid behavioral theory (first consider, then choose) which directly incorporates non-compensatory decision-making to obtain strong individual-level estimates and can even work with small sample sizes. • This method also provides more accurate market simulation data than CBC. 8
- 9. Menu-based Conjoint (MBC) www.vernonresearch.com • MBC is an advanced modeling methodology. • MBC mimics the real-world experience of choosing among a variety of options when configuring a preferred product. • Studies with MBCs are useful in situations where end-users may select pre-designed bundles as well as items a la carte. • Respondents generally enjoy these exercises due to the freedom provided them during the customization process. • Examples of products and services when MBC might be preferred: – choosing options to put on an automobile – selections from a restaurant menu – banking options – configuring an insurance policy – mobile phone plans – home phone, Internet and/or cable bundles with options 9
- 10. Output www.vernonresearch.com Importance (of factors) •The relative weight respondents give each factor when making purchasing decisions. •Importance is not consciously assigned; instead, weights are calculated based on how each choice decision was made. •Importance is also additive, so the sum total of two factors of lesser importance may exceed a factor of higher importance. 10
- 11. Output www.vernonresearch.com Utility (of levels) •This is a measure of each level’s usefulness on an integer scale. •Utility values are created through effects coding, where the sum of all values equals zero. •Negative scores do not necessarily translate into an unattractive evaluation. •Another way of thinking about utility is as level preference within the same factor. 11
- 12. Output www.vernonresearch.com Market simulations •This is the most powerful tool derived from conjoint data. •Market simulations can predict the impact on market share of new products or new level combinations of current products. •By changing one level of a single factor, market share gains and losses can be measured. •The reliability of simulations depends on how accurate the exercise mirrored reality. 12
- 13. Market Simulation Example www.vernonresearch.com PRODUCTS Brand Style Colors Features Warranty Price #1 – FreezeMaster Kitchen Aid French Door All Filtered water dispenser VARIABLE VARIABLE #2 – Penguin’s Choice Whirlpool French Door White, off-white or black Icemaker in door 1 year $1,449 #3 – Snowbank Pro GE Freezer on top All TV monitor in door 3 years $1,699 #4 – Siberian Huskee Kenmore Side by side White or off- white Icemaker in door 1 year $1,599 #5 – Dr. Brrrr Frigidaire Freezer on top White, off-white or black None 1 year $1,149 #6 – Just a Fridge None-Generic Freezer on bottom White None None $999 13
- 14. Market Simulation Example www.vernonresearch.com FreezeMaster with No Warranty 14
- 15. Market Simulation Example www.vernonresearch.com FreezeMaster with 1-Year Warranty 15
- 16. Market Simulation Example www.vernonresearch.com FreezeMaster with 3-Year Warranty 16
- 17. Market Simulation Example www.vernonresearch.com FreezeMaster with 5-Year Warranty 17
- 19. Conjoint Analysis Summary Conjoint analysis is… •a great way to determine which bundles of attributes will make the best products or services. •the most reliable means of determining the best price of a product or a service. •extremely powerful and accurate in predicting how markets will react to modified or new products/services. •useful in helping understand the relative importance and usefulness of various attributes during product development. www.vernonresearch.com 19