What is conjoint analysis?

Person completing usage and attitude study
Image of the post author Kadence International

What is conjoint analysis? It’s often lauded as an extremely effective way to gain detailed insights and conduct market research, but how does it work?

Essentially, conjoint analysis is a way of measuring the value that customers place on an existing or new product’s features. It typically works via a survey, which looks something like this:

  • Participants are shown a combination of features (called attributes) for a product. If the product is a smartphone, for example, they might be shown the price, memory size, screen resolution, and camera quality.
  • They’re then asked to compare different attributes. For example, what would they choose out of a $150 phone and a $250 phone? Do they prefer 32GB of memory or 64GB? There are several different ways to structure this, as we’ll find out.
  • After the answers have been collected, we must analyze the results to inform the right marketing decisions.

Conjoint analysis is a powerful research method used in market research to analyze and understand customer preferences. It is particularly valuable in assessing product attributes and their impact on consumer decision-making.

In this blog post, we’ll look at this process in more detail and dig deeper into the different types of conjoint analysis and the various benefits it can deliver. 

Why do conjoint analysis?

By examining various factors such as product features, price, brand, and packaging, conjoint analysis provides insights into how different attributes influence the target audience’s choices.

This market research method proves especially beneficial in different stages of the product life cycle, from initial business analysis and product design to the product launch and beyond.

Conjoint analysis aids great product managers, designers, and project managers in making informed decisions by identifying the optimal combination of features and attributes that resonate with the target audience. It leverages research techniques like mail surveys, personal interviews, focus groups, and telephone surveys conducted by skilled survey researchers to gather data on consumer preferences and opinions. This valuable information helps development teams refine and optimize the final product, ensuring it meets customer needs and expectations.

Additionally, the conjoint analysis provides insights into the competitiveness of existing products and aids in strategic planning for future product enhancements or new offerings.

There are several reasons to conduct a conjoint analysis. These include:

  • To measure and understand customer preferences for certain product features
  • To assess or predict how well a new product will do if brought to market
  • To gain an understanding of how changes in price affect demand
  • To predict future trends, for example, around the adoption of certain features

How to do conjoint analysis

Get regular insights

Keep up to date with the latest insights from our research as well as all our company news in our free monthly newsletter.

Choose the right survey type

The first stage is to decide on the correct survey type. There are several ways to do a conjoint analysis — here are the main methods.

  • Ratings-based conjoint analysis. This is where participants give each attribute a rating, for example, on a scale of 1-100.
  • Ranking-based conjoint analysis. This is where participants rank the attributes in order from best to worst. There is also a best vs worst analysis, where participants simply pick their favorite and least favorite attributes from the selection.
  • Choice-based conjoint analysis (CBC). This is the most commonly used model and the one this guide will focus on. It presents combinations of attributes to participants and asks them to choose which they prefer.

One of the most powerful advantages of choice-based conjoint analysis is that it can allow you to use modelling to predict how customers will feel about combinations they didn’t even assess.

In other words, in an extremely efficient way of predicting responses to features without having to spend a huge amount of time testing each combination.

Identify the relevant attributes (features)

Next, it’s time to decide which product attributes you want to have your respondents compare and assess. The key is not to use too many. We typically avoid using more than 5 or 6 attributes, e.g. for a car color, or engine size. We do this to reduce respondents’ cognitive load to ensure they engage with the choices presented to them. 

For each attribute, you need to add levels. For example, if your participants are assessing a smartphone, one attribute might be ‘price’, and the levels might be $200, $350, and $700.

The levels will usually reflect the different tiers of the product you’re considering selling. For the smartphone, you might be releasing a basic model, a higher-end model, and a deluxe model. The levels for attributes such as price, camera size, and memory will align with those tiers.

Levels should be chosen based on factors like:

  • How interesting and valuable they are for management — will they inform useful decisions?
  • How well do they avoid bias?
  • How realistic they are

In the CBC method, there are two commonly used models for making choices:

  • Single choice with none. This requires the participant to make one choice out of the selection. There is also the possibility to select none of the options.
  • Single choice. This is the same as above, but there is no ‘none’ option — the participant has to pick one. 

Design the questionnaire

Screener questions

Most Surveys start with some screener questions. These are general questions around demographics like the respondent’s age, job title, or purchase habits. The goal is to filter out those who won’t be a good fit for the survey based on the people you’re trying to target.

Introduce and explain

It’s important to take some time at the beginning of the survey and in your questions to clearly explain what the respondents need to do to answer the question. Surveys should be as clear and easy to follow as possible.

Create the right questions

The questions you choose and how you structure them will make or break your survey. Here are some guidelines to follow:

  • Questions should follow one another logically and be grouped together intuitively. It’s best not to confuse your participants by ordering your questions in a confusing way.
  • People often give more accurate and useful answers when you use situational questions g. For example, instead of asking, “Which phone would you buy” ask something like, “Thinking back to the last time you purchased a phone — if you had the following options instead, which would you have picked?”
  • Finish with some demographic questions so that you can further understand your customer base and analyze the results by demographic to understand any meaningful differences.

Analyze and take action

Once the survey has been written, scripted, sent out, and completed by your target group, it’s time to analyze the results and take action on them. This is perhaps the most important part of the process, as it’s where your research can really make a tangible impact.

There are several ways to analyze your results based on how you designed the survey. The most important thing is to collect and analyze your data in a way that makes it easy to draw useful conclusions and share them.

This will allow you to gain real value from the survey and present those findings to others in the company. This:

  • Helps justify your decisions and actions
  • Informs future plans and inspires new features
  • Identifies areas that need to change or improve

At Kadence, it’s our job to ensure you create and conduct the most effective surveys and market research possible, giving your brand the edge. To find out more about how we can help with conjoint analysis and more, get in touch to request a proposal.