CBC vs ACBC Conjoint: When to Use Each

What's the Difference Between CBC and ACBC?

Choice-based conjoint (CBC) and adaptive choice-based conjoint (ACBC) both measure how people value product features by presenting trade-off scenarios. The difference is in how those scenarios are constructed. CBC shows every respondent the same pre-generated set of choice tasks. ACBC builds the choice tasks dynamically during the interview, adapting to each person's earlier responses.

For most studies, CBC is the right call. It's simpler to design, cheaper to field, and the analysis is well-understood. ACBC earns its place when you have too many attributes for CBC to handle cleanly, when your sample is small, or when you need deeper individual-level precision.

How CBC Works

In a standard CBC study, respondents see 10-15 choice tasks. Each task shows 3-5 product profiles (combinations of attribute levels) plus a "none of these" option. The respondent picks the one they'd choose.

Every respondent sees the same set of tasks, generated before the survey launches using an experimental design algorithm. The design ensures balanced coverage: every attribute level appears roughly the same number of times, and the effects of different attributes can be separated statistically.

After data collection, the analysis (typically hierarchical Bayesian estimation) produces part-worth utilities for each attribute level, relative importance scores, and inputs for market simulation.

CBC's strength is simplicity. The design is straightforward, the respondent experience is predictable, and the statistical properties are well-documented. It works reliably for studies with 4-7 attributes and sample sizes of 300+.

How ACBC Works

ACBC structures the interview in three stages, and each stage informs the next:

Stage 1: Build Your Own (BYO)

The respondent constructs their ideal product by selecting their preferred level for each attribute. This creates a starting configuration and gives the algorithm an initial read on their preferences.

Stage 2: Screening

The system generates product profiles near the respondent's ideal configuration and asks: "Would you consider this?" Some profiles are slight variations on the ideal. Others introduce trade-offs by swapping levels.

During screening, the respondent can also flag "must-have" levels (e.g., "I won't consider anything without unlimited users") and "unacceptable" levels (e.g., "I'd never pay more than $99/month"). These constraints narrow the consideration set for the final stage.

Stage 3: Choice Tasks

The system presents choice tasks, but unlike CBC, these tasks are constructed on the fly based on the BYO preferences, screening responses, and stated constraints. Each respondent gets a personalized set of trade-offs focused on the decisions that actually matter to them.

ACBC's strength is efficiency. Because each respondent's tasks are tailored to their preferences, ACBC collects more useful information per person than CBC. This means smaller samples can produce comparable or better individual-level estimates.

Head-to-Head Comparison

When to Choose CBC

CBC is the default for a reason. Choose it when:

Your study has 4-7 attributes. CBC handles this range comfortably. The pre-generated design ensures clean estimation of all main effects, and with 300+ respondents, you'll get reliable individual-level utilities via HB.

You have adequate sample. If you can reach 300-500 respondents without budget strain, CBC gives you everything you need. The larger sample compensates for the fact that each respondent sees a generic (not personalized) set of tasks.

Pricing is a primary objective. CBC is the gold standard for pricing research. The static design ensures that every price level gets equal exposure across respondents, which produces clean willingness-to-pay estimates. ACBC's adaptive screening can sometimes under-expose extreme price levels if they get filtered out early.

You want broad platform compatibility. Most survey platforms with conjoint support offer CBC. ACBC is primarily available through Sawtooth Software's Lighthouse Studio and a handful of enterprise research platforms. If you're committed to a specific platform, check ACBC availability before planning around it.

You need simplicity. CBC designs are easier to QA, easier to explain to stakeholders, and the analysis pipeline is more standardized. If this is your team's first conjoint study, start with CBC.

When to Choose ACBC

ACBC justifies its added complexity in specific situations:

You have 8+ attributes and can't cut any. When stakeholders insist on testing 10 features, CBC becomes unwieldy. Respondents can't cognitively process 10-attribute choice tasks. ACBC's staged approach (BYO narrows the field, screening applies constraints, choice tasks focus on remaining trade-offs) makes 8-12 attributes feasible without overwhelming respondents.

Your sample is constrained. Niche B2B audiences, rare disease patients, or specialized professional segments often cap out at 100-200 respondents. ACBC's adaptive design extracts more information per person, producing viable estimates where CBC would require twice the sample.

You need "must-have" and "deal-breaker" data. The screening stage explicitly captures which feature levels are non-negotiable. This goes beyond what utility scores reveal. Knowing that 35% of respondents flag "HIPAA compliance" as a must-have is directly actionable for product roadmap decisions.

Respondent engagement matters. ACBC interviews run longer than CBC (15-25 minutes vs. 10-15), but respondents consistently report finding them more engaging and realistic. The BYO section feels like configuring a product, not filling out a survey. If you're worried about drop-off in a long survey, ACBC's engagement advantage can offset the extra time.

Practical Considerations

Cost Difference

ACBC costs more in three ways: software licensing (Sawtooth's ACBC module carries a premium), design time (the three-stage interview requires more setup and testing), and analysis time (the adaptive data structure adds complexity). For a typical study, expect ACBC to add 20-40% to total project cost compared to CBC.

Analysis Differences

Both methods produce part-worth utilities and support market simulation. ACBC adds the BYO configuration data and must-have/unacceptable flags as supplementary outputs. The HB estimation process is similar, though ACBC's individual-level data tends to be richer because each respondent's tasks were tailored to their preference space.

One nuance: because ACBC doesn't expose every respondent to the full attribute-level space, some level combinations get less coverage than in CBC. For main effects, this rarely matters. For interaction effects between specific attributes, CBC's balanced design can be more reliable.

Hybrid Approach

Some researchers use ACBC for the initial study (to identify the most important attributes and levels with a small sample) and then follow up with a CBC study on the shortlisted attributes with a larger sample. This gets you ACBC's exploratory power and CBC's estimation precision.

Frequently Asked Questions

Can I switch from CBC to ACBC mid-project?

Not without redesigning the study from scratch. The interview structure, experimental design, and data format are fundamentally different. Decide before you begin fielding.

Does ACBC work for pricing studies?

It can, but CBC is generally preferred for pricing-focused research. ACBC's screening stage may filter out extreme price levels before respondents see them in choice tasks, which can reduce the precision of price sensitivity estimates at the tails.

Is ACBC always better with small samples?

For individual-level estimates, yes. ACBC's adaptive design produces more stable individual utilities at n=150 than CBC would at the same sample size. But "small sample" is relative. If you can reach 300+, CBC's simpler design and broader software support make it the easier choice.

What sample size does ACBC need?

Minimum 100-150 for aggregate results, 200-300 for reliable individual-level estimates. That's roughly 40-50% less than CBC for comparable precision, because each respondent's data is more informative.

Related Guides

• Conjoint Analysis: Complete Guide -- Full methodology overview including CBC basics
• How to Design a Conjoint Study -- Design decisions for CBC studies
• Conjoint Analysis Sample Size Requirements -- How sample needs differ by method
• Conjoint Analysis Software -- Platform comparison including ACBC support
• MaxDiff vs Conjoint -- When a simpler method is the better fit
• Conjoint Analysis Survey Template -- CBC survey template to get started

Run CBC conjoint in minutes -- try Quali-Fi free for 14 days.