Types of Conjoint Analysis: How They Work and When to Use Each

Choosing between product ideas rarely comes down to one factor. Price matters, but so do features, brand, and dozens of other trade-offs customers make without ever saying so out loud. The different types of conjoint analysis exist because no single survey method captures every kind of trade-off equally well.

Some types work best when a study has too many attributes to test at once. Others mirror a real checkout experience almost exactly. A few let customers build their own version of a product instead of picking from ready-made options.

In this blog, we’ll break down the six types of conjoint analysis researchers use most, walk through real examples of each, and help you match a type to your own study.

What is conjoint analysis?

Conjoint analysis is a survey-based market research method that shows customers realistic product combinations, called profiles, and measures which combination they would actually choose. Instead of asking people to rate features one at a time, it forces the same trade-offs people make in real shopping.

Researchers break every product down into attributes (the features being tested, such as brand or price) and levels (the specific options within each attribute). A skincare study, for example, might test:

• Brand: CeraVe, Neutrogena, La Roche-Posay
• Price: $12, $18, $24
• Package size: 4 oz, 8 oz, 12 oz

Statistical modeling then converts respondent choices into a part-worth utility score, a number showing how much each attribute level contributes to the overall decision. For the full mechanics behind that scoring process, see this complete guide to conjoint analysis.

How many types of conjoint analysis are there?

Search results and textbooks do not agree on this number, and that mismatch is where most of the confusion starts. Some sources list only two types. Others count a dozen or more, folding in rating-based and hybrid variants that see little real-world use today.

In practice, six types cover almost every study a market researcher runs: choice-based conjoint, adaptive choice-based conjoint, menu-based conjoint, adaptive conjoint analysis, traditional full-profile conjoint, and MaxDiff, which sits just outside the conjoint family but gets grouped with it constantly. Older or niche variants, such as rating-based conjoint and self-explicated models, still show up in academic papers but rarely in commercial research anymore.

The more useful question isn’t how many types exist. It’s which one matches your attribute count, budget, and how closely the survey needs to mirror an actual purchase.

Choice-based conjoint (CBC) analysis

Choice-based conjoint, also called discrete choice conjoint, is the most widely used type of conjoint analysis today. Respondents see two or more complete product profiles side by side and pick the one they would actually buy, the same way they would compare options in a store or online.

CBC studies run on one of a few design structures: a random design that combines attributes uniquely for each respondent, a D-optimal design that an algorithm builds to minimize bias for the given sample size, or an imported design pulled from an existing statistical package.

• Mirrors real purchasing behavior more closely than any other type
• Handles multiple attributes without overwhelming respondents
• Feeds directly into market share simulations
• Works well when price is a central research question

A coffee machine study might show three machines that vary by price, brand, and one special feature, then ask which one a respondent would buy. The resulting choices reveal exactly how much weight price carries against brand loyalty, something a rating scale could never show. Most conjoint software, including QuestionPro’s conjoint analysis tool, builds around CBC as the default method for this reason.

Adaptive choice-based conjoint (ACBC)

Adaptive choice-based conjoint blends the realism of CBC with an adaptive survey flow. It works well for products with many attributes and clear must-have features, such as laptops or software bundles, where some options are simply non-negotiable for a given buyer.

The interview runs through three phases:

1. Build-your-own step.
   Respondents configure their ideal product from the available attributes and levels.
   
2. Screening step.
   They mark which features are unacceptable, narrowing the field before any trade-offs happen.
   
3. Choice tournament.
   A series of narrowed comparisons pinpoints the final preference among a respondent’s own shortlist.

Because the survey adapts to each respondent’s own must-haves, ACBC tends to produce more engaging interviews than a standard CBC study, though it takes more design effort and a larger sample to field well.

Menu-based conjoint (MBC)

Menu-based conjoint asks respondents to build their own product or service from a menu of components, instead of choosing between a few fixed profiles. A fast-food study might have respondents select a burger, a side, and a drink, each carrying its own price and attribute levels.

MBC fits purchase situations where customers naturally mix and match, such as telecom plans, software subscription tiers, or restaurant combo meals. The output shows not just which single item people prefer, but how they bundle several items together under a budget constraint, which a fixed-profile study cannot capture.

Adaptive conjoint analysis (ACA)

Adaptive conjoint analysis was the dominant method before choice-based conjoint took over in the 1990s. Instead of full product profiles, it presents graded-pair comparisons, asking respondents to rate their relative preference between two attributes at a time, then adjusts later questions based on earlier answers.

When ACA still makes sense:

• Studies testing more attributes than a respondent could reasonably compare in full profiles
• Early-stage product design research, before pricing becomes the central question

When to avoid it:

• Pricing-sensitive studies, since ACA does not simulate a direct purchase decision the way CBC does
• Any study that needs market share simulation as an output

Traditional (full-profile) conjoint analysis

Traditional conjoint, also called full-profile conjoint, is the original method developed in the 1970s. Respondents rank or rate complete product profiles that show every attribute at once, rather than choosing between separate concepts.

It works best for smaller studies with only a handful of attributes. Ranking a long list of full profiles gets exhausting fast, so traditional conjoint rarely scales past three or four attributes before respondent fatigue sets in. Researchers still reach for it, though, when they need a simple conjoint analysis example to explain the concept to stakeholders who have never seen one before.

Is MaxDiff a type of conjoint analysis?

Not technically, though the two get confused constantly. MaxDiff, short for maximum difference scaling or best-worst scaling, shows respondents a list of individual items, such as features or benefits, and asks them to pick the most and least important in each set. It never combines items into full product concepts the way conjoint does.

• Conjoint measures how attributes combine to drive a single purchase choice
• MaxDiff ranks a long list of standalone items against each other, without building them into concepts

The two methods pair well together. Teams often run MaxDiff first to narrow a long attribute list down to the handful that matter most, then build a conjoint study around those survivors. See this comparison of conjoint analysis vs. MaxDiff for a full breakdown, or explore QuestionPro’s MaxDiff software directly.

Which type of conjoint analysis fits your study?

The table below lines up all six types side by side, so you can match one to your attribute count, budget, and research goal at a glance.

Conjoint analysis types in action: 3 Quick examples

Seeing each type applied to a real research question makes the differences easier to remember than any definition alone.

SaaS pricing, using CBC

A US project management software company tests three pricing tiers against feature bundles like storage limits and integrations, using choice-based conjoint to see which combination customers would actually pay for, not just say they want.

Employee benefits, using ACBC

An HR team building a new benefits package asks employees to first configure their ideal mix of health coverage, remote work days, and retirement match, then screens out combinations the budget cannot support before running trade-off comparisons on what remains.

Restaurant menu design, using MBC

A US fast-casual chain uses menu-based conjoint to see how customers build a meal from entrees, sides, and drinks, revealing which combinations drive the highest average order value.

How to choose the right type of conjoint analysis

Matching a type to your study comes down to four questions, answered in order.

1. How many attributes are you testing?
   Six or fewer points toward CBC. More than that usually calls for ACBC or ACA.
   
2. Is price a central question?
   If yes, rule out ACA and lean toward CBC or ACBC, since both simulate a real purchase decision.
   
3. Do customers naturally bundle products?
   Menu-based conjoint fits purchase situations built from separate, combinable parts.
   
4. What’s your sample size and budget?
   Smaller studies with tight budgets often do better with traditional full-profile conjoint, or a MaxDiff screening step first to trim the attribute list.

How QuestionPro supports different types of conjoint analysis

QuestionPro’s conjoint analysis tool is built around choice-based conjoint, the type most researchers reach for first. A guided task creation wizard lets you enter attributes and levels, choose a design type such as random or D-optimal, and set constraints like fixed tasks or prohibited combinations, without writing a line of code.

Once responses come in, the platform calculates part-worth utility scores and relative importance automatically, then supports market share simulations so you can test how a proposed product might perform before it launches. Reports export to Excel, CSV, or HTML for deeper analysis alongside QuestionPro’s broader market research software.

Common mistakes when choosing a conjoint analysis type

Most conjoint studies do not fail because of bad execution. They fail because the wrong type was matched to the question being asked.

• Defaulting to CBC out of habit.
  It’s the most popular type for good reason, but a bundled-purchase study almost always performs better as menu-based conjoint.
  
• Testing too many attributes in a full-profile design.
  Traditional conjoint breaks down fast past four attributes, since ranking long profile lists exhausts respondents before they finish.
  
• Skipping a MaxDiff screening step.
  Teams that jump straight into a 12-attribute CBC study often see noisier data than a shorter study built from a pre-screened attribute list.
  
• Ignoring sample size minimums for the chosen type.
  Under-sampling a segmented ACBC study makes the resulting part-worths unreliable, even when the type itself was the right pick.

Picking a type is a decision, not a formality

The type of conjoint analysis you choose shapes every number the study produces afterward. A well-matched method turns a vague sense of what customers want into a specific, defensible answer about price, features, and trade-offs.

Start with your attribute count and your research question, not the type that happened to work on your last project, and the right method usually becomes clear on its own.

Frequently Asked Questions (FAQs)