It also used to be expensive. Corus is changing that.
In the toolbox of the market researcher, few things are as powerful as a well done conjoint analysis. In this article, we will dive into the why and the how of this particular market research tool.
What is Conjoint?
Conjoint analysis assesses how much each variable in a study contributes to consumer choice. In the sense of the output, it’s that simple. Conjoint analysis dates back to the early 1970s (so almost 50 years ago!) as a method of testing econometric theory and choice logic. Paul Green is often credited with the first published papers on conjoint.1 It is fitting that someone with a color for a last name would invent a study type that very often includes testing color preference… or maybe that’s just us nerds that found that a funny happenstance. Anyway, a conjoint analysis is then a multiple-regression analysis of the attributes that make up the choice that a person makes.2
Let’s say for example that Apple is getting ready to launch a brand new iPhone. As they get ready to launch the product, they want to figure out what to focus on in their advertising – is it the new screen resolution? The color options? New app features? Something else? By performing a thorough conjoint analysis, they discover that storage capacity is the best opportunity to price discriminate and drive a higher ASP (average sale price).
Sounds simple enough right? But simulating reality and ensuring that respondents have digested the choice that they are actually making with ‘fake money’ can be difficult. What is even more difficult is deducing stated vs. derived needs. A customer may claim that they are price sensitive, but if presented correctly, conjoint analysis can uncover that they are, in fact, more feature sensitive.
So how is a conjoint study conducted through the Corus platform, and why is it better? Our platform enables you to build a conjoint ‘spinner’ within a broader complex survey, adding branding and skip logic to multiple conjoint questions effortlessly based on the respondent. You can also make price within a conjoint a function of a prior question in case you’re offering insurance, or a service plan, for example. It gets better if you’re looking to perform a study internationally – human translation is integrated and you can conduct a conjoint in dozens of countries within hours of starting the process.
Our collaborative tools and automation also make the most challenging part of conjoint, developing “the spinner”, easier. This is essentially developing the choice architecture of the study – what factors do you want to include and what options for each factor do you want to test? Because all other pieces have been automated, and our platform is collaborative, you can spend what precious little time you have confirming that you and the team are testing the correct factors.
Finally, The transparent and easy check-out process through Corus means that your survey is up and running immediately, and doesn’t need to be distributed through some third party means. Once you have your responses, you have the ability to either run conjoint yourself (if you so choose) or immediately engage an analyst to perform the work.
Then you have an instant connection with 80M respondents around the globe matching all manner of demographic sets, so you no longer have to “recruit panel” – even more time saved! (Side note: we’ll let you in on a little secret here, but all market research firms are using the same tools to recruit panel, so don’t let them convince you that you have to pay a ton for it).
Corus has taken great steps to make the research process as easy, fast, and cost effective as possible. Certainly others have tried to tackle these three things with conjoint analysis, with varying degrees of success, but we saw opportunity for a much for powerful and streamlined process. So we decided to roll conjoint into the platform, and provide clear pricing as we have done for other research results. In the end, you get conjoint data in a perfectly clean dataset alongside the responses to other questions and respondent metadata. Regression can be run easily for the full set or for subsets of respondents to guide strategy. Don’t know how to do that? We have vetted professionals ready to help you generate the insights that will persuade your business to make the right decisions to respond to customer needs.
If you’re interested in this type of work, feel free to download this document for reference: Survey Products
Learn about other data analysis products here: Data Analysis