.png)
Frameworks, core principles and top case studies for SaaS pricing, learnt and refined over 28+ years of SaaS-monetization experience.
Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form.
Conjoint analysis is a statistical research method that measures customer preferences by asking them to make trade-offs between different product features, pricing, and attributes. It reveals which features customers value most and how much they're willing to pay, making it invaluable for pricing strategy, product development, and feature prioritization—especially for SaaS companies with complex packaging decisions.
If you’ve ever wondered why your “obvious” pricing change didn’t move revenue—or worse, hurt conversions—you’re not alone. Many SaaS companies make what we call the $50M pricing mistake: changing price and packaging based on opinions, competitor pages, or internal politics instead of actual customer trade-offs.
Conjoint analysis is how you prevent that.
In this guide, you’ll learn what conjoint analysis is, how it works, when to use it for SaaS pricing, and how to decide if it’s worth the investment for your business.
What is Conjoint Analysis?
Simple definition in plain English
At its core, conjoint analysis is a way to answer:
“If I change my product’s features and price, which combination will my customers actually choose?”
Instead of asking customers directly, “How much would you pay?” or “Which feature is most important?”, conjoint analysis:
- Shows people realistic product options (e.g., different plans with different features and prices)
- Asks them to choose between these options
- Uses statistics to back out how much each feature and price point is actually worth to them
It’s like running thousands of virtual A/B tests in a single survey.
The core principle: trade-off analysis
The key idea behind conjoint is trade-offs.
In the real world, your buyer doesn’t get everything they want at the lowest possible price. They weigh:
- More features vs. lower price
- Better support vs. longer contract
- Integrations vs. implementation complexity
Conjoint analysis mimics that decision process by forcing respondents to make choices between imperfect options. By observing these choices, we can:
- Quantify which attributes really drive selection
- Estimate willingness to pay for specific features
- Simulate how changes in price/packaging will shift demand and revenue
Brief history and origin
Conjoint analysis emerged in marketing research in the 1970s, formalized in work by researchers like Green & Srinivasan (1978). Originally used by consumer goods and automotive companies, it helped them understand how customers valued things like brand, size, flavor, and price.
Over time, conjoint moved from academia into commercial research. With modern software and computational power, it’s now widely used in:
- B2C products (CPG, retail, travel)
- Telecom and subscription businesses
- B2B SaaS pricing and packaging decisions
Why it’s called “conjoint” analysis
The name comes from “considered jointly”.
Instead of asking about each product attribute in isolation—e.g., “How important is SSO?” or “Do you care about 99.99% uptime?”—conjoint measures the joint effect of all attributes together:
- How important is SSO given this price?
- How important is 24/7 support when combined with usage limits?
- How does a cheaper price offset a missing feature?
That’s exactly how real buyers think when they compare your pricing page with competitors.
How Conjoint Analysis Works
The basic methodology: showing options and recording choices
Here’s how a typical choice-based conjoint (CBC) study works for SaaS:
- Define attributes
Example SaaS attributes:
- Price per month
- Number of seats or usage limits
- Key features (SSO, advanced reporting, API access)
- Support level (email only vs. 24/7 phone)
- Contract term (monthly vs. annual)
- Define levels for each attribute
For example:
- Price: $49, $99, $199 per month
- Seats: 10, 25, 50
- Support: email only, business-hours chat, 24/7 phone
- Create product profiles (concepts)
These are combinations of attribute levels, like mini pricing cards:
Option A
- $49/month
- 10 seats
- Email support only
- Basic reporting
Option B
- $99/month
- 25 seats
- Chat support
- Advanced reporting
- Ask respondents to choose
For each task, respondents see 2–4 options and answer something like:
“Which option would you be most likely to choose?”
Sometimes there’s also a “None” option to capture those who wouldn’t buy at all.
- Repeat across multiple choice tasks
Each person completes, say, 8–12 tasks, each with different combinations.
From these repeated choices, we can estimate how much each feature and price level contributed to their decisions.
A real conjoint analysis example (SaaS)
Imagine you run a B2B analytics tool and you’re rethinking your pricing tiers. You want to know:
- How much more will people pay for SSO?
- Does “advanced reporting” deserve its own higher tier?
- What’s the revenue-optimal price for your Pro plan?
You might test a conjoint design with these attributes and levels:
- Price per month: $99, $149, $199
- Dashboards: 3, 10, unlimited
- Users: 5, 20, 50
- SSO: Not included, Included
- Support: Email only, Priority email + chat
A sample choice task in your survey might look like:
Which plan would you choose for your team?
| | Plan A | Plan B | Plan C |
|---------------------|-------------------------|---------------------------|------------------------|
| Price | $99 / month | $149 / month | $199 / month |
| Dashboards | 3 | 10 | Unlimited |
| Users | 5 | 20 | 50 |
| SSO | Not included | Included | Included |
| Support | Email only | Priority email + chat | Priority email + chat |
[ ] I’d choose Plan A
[ ] I’d choose Plan B
[ ] I’d choose Plan C
[ ] I wouldn’t choose any of these
Respondents complete multiple tasks like this, each with different combinations.
What the data reveals: part-worth utilities
Behind the scenes, conjoint analysis estimates “part-worth utilities”.
Think of utility as a numerical score that represents how attractive each attribute level is to buyers. For example, the analysis might tell you:
- Utility for SSO included = +30
- Utility for SSO not included = 0
- Utility for $99/month = +15
- Utility for $199/month = -15
These aren’t meaningful as absolute numbers; what matters is the difference between levels:
- SSO included vs. not included: +30 points
- $99 vs. $199: +30 points
Interpretation: In this market segment, SSO is roughly as valuable (in utility terms) as the difference between $99 and $199 per month.
From there, we can estimate willingness to pay:
“On average, respondents are willing to pay about $100/month more for SSO.”
Multiply this across accounts and contract terms, and it becomes powerful input for your pricing strategy.
From utilities to pricing insights
Once you have utilities, you can:
- Simulate pricing scenarios
- “What if we include SSO only in the Pro tier at $199?”
- “What if we lower the Standard plan from $149 to $129?”
- “What if we make support 24/7 only on Enterprise?”
Estimate market share for each scenario
Using conjoint simulators, you can estimate the percentage of respondents who would choose each offering or competitor.Find revenue-optimal price points
By combining simulated choice shares with price levels, you can plot revenue curves:
- At $149, more people buy, but each pays less
- At $199, fewer buy, but each pays more
The conjoint model helps you find the sweet spot.
- Prioritize features
- Which features truly move the needle?
- Which are “table stakes” vs. actual differentiators?
- Where should you invest roadmap effort?
Pro Tip:
Don’t stop at the first “optimal” price. Use conjoint outputs to build multiple scenarios—e.g., growth-focused, margin-focused, and adoption-focused pricing structures—and then validate your top candidates in the market.
Types of Conjoint Analysis
There are several types of conjoint analysis and related methods. For SaaS pricing, you’ll mostly see:
- Choice-based conjoint (CBC)
- Rating-based conjoint
- MaxDiff (Best-Worst Scaling)
- Adaptive conjoint
Choice-Based Conjoint (CBC)
Choice-based conjoint is the most common and realistic method used in SaaS pricing research.
Characteristics:
- Respondents choose between product configurations
- Each task shows a small set of options (usually 2–4)
- Closely mimics real-world buying behavior (choosing between vendors or plans)
Best for:
- SaaS pricing and packaging
- Comparing your plans vs. competitors
- Understanding willingness to pay for specific features
- Modeling market share and revenue impact of pricing changes
CBC is usually the recommended starting point for conjoint analysis pricing questions in B2B SaaS.
Rating-Based Conjoint
In rating-based conjoint, instead of choosing between options, respondents rate or rank each product profile.
Example:
“On a scale from 1 to 10, how likely would you be to purchase this plan?”
Pros:
- Easier for respondents when there are many combinations
- Can capture more nuanced sentiment than binary choices
Cons:
- Less realistic than actual choices
- Ratings can be inflated or inconsistent
- Harder to simulate forced trade-offs
Rating-based conjoint is used less often for pricing strategy today, but can be useful when:
- You’re exploring early product concepts
- You want to understand preference structure without making people choose between options
MaxDiff Analysis (Best-Worst Scaling)
MaxDiff isn’t conjoint in the classic sense, but it’s tightly related and often used alongside conjoint in SaaS pricing research.
Instead of full product profiles, MaxDiff shows lists of features or benefits and asks:
“Which of these is most important and which is least important?”
Example for a SaaS platform:
- Single sign-on (SSO)
- Custom dashboards
- Priority support
- Unlimited API calls
Respondent selects one as Best and one as Worst. Across many tasks and respondents, you get a ranked list of feature importance with ratio-scale scores.
Best for:
- Feature prioritization
- Designing good-better-best tiers
- Understanding what to include/exclude before running a full conjoint
However, MaxDiff alone doesn’t give you willingness to pay; you use it to narrow the feature set and then layer conjoint on top.
Adaptive Conjoint Analysis
Adaptive conjoint uses respondent answers in real time to “adapt” which questions they see next.
Pros:
- Can handle more attributes than traditional conjoint
- More efficient for complex products
- Reduces survey fatigue
Cons:
- More complex to design and analyze
- Less transparent to stakeholders
- Not always necessary for typical SaaS pricing questions
In SaaS, adaptive conjoint can be helpful when:
- Your product has many modules or add-ons
- You’re pricing a platform with dozens of features
- You want deep insight at the individual account level
Which type of conjoint to use when
Here’s a simple decision framework:
| Situation | Recommended Method |
|----------|--------------------|
| You want to redesign pricing tiers or compare competitors | Choice-Based Conjoint (CBC) |
| You want to understand which features matter most before pricing | MaxDiff, then CBC |
| You have very complex products with many attributes | Adaptive conjoint or carefully designed CBC |
| You’re doing early-stage concept testing | Rating-based conjoint or MaxDiff |
Common Mistake:
Jumping straight into a complex adaptive conjoint because it “sounds advanced.” In most SaaS pricing projects, a well-designed choice-based conjoint combined with MaxDiff for feature screening gives 95% of the value with far less complexity.
When to Use Conjoint Analysis for Pricing
Ideal scenarios for SaaS
Conjoint analysis shines in situations where price, features, and packaging all interact. For B2B SaaS, that typically means:
- New product launches
- You’re entering a new market and need to set a price
- You’re unsure how to bundle features into good-better-best tiers
- You want to position against incumbents with confidence
- Packaging and tier redesign
- Your current plans are confusing or misaligned with value
- You suspect you’re underpricing key features (e.g., SSO, API, advanced analytics)
- Sales complains about endless custom discounts and one-off deals
- Feature monetization and add-ons
- You’re releasing major new functionality and wondering:
- Should it be a paid add-on or included in Pro/Enterprise?
- How much extra will customers pay for it?
- Segment-specific pricing
- You serve multiple segments (SMB, mid-market, enterprise)
- You want to see how preferences and willingness to pay differ by segment, industry, or role
Company stage and revenue suitability
Conjoint analysis is a high-leverage, higher-investment method. It’s typically best for SaaS companies with:
- $1M+ ARR (minimum)
- Clear signs of product-market fit
- Multiple pricing/packaging decisions at stake (not just “Should we be $19 or $24?”)
It becomes especially valuable when:
- You’re at $5M–$300M ARR
- Pricing mistakes now have seven- or eight-figure impact
- You want to move from “gut-feel pricing” to a disciplined, data-driven pricing strategy
When NOT to use conjoint analysis
Conjoint analysis is not always the right tool. It’s often overkill when:
- You’re pre–product-market fit
- You’re testing only a simple price change on an existing plan
- Your ACV is very low (e.g., <$100/year) and price changes don’t move the margin needle
- You have no budget for robust sampling and analysis
In those cases, faster and cheaper methods may be more appropriate.
Alternatives for smaller or earlier-stage companies
If conjoint isn’t a fit yet, consider:
Van Westendorp Price Sensitivity Meter
For a quick read on price ranges and acceptable bandsSimple willingness-to-pay questions
Carefully written surveys using open-ended or bounded WTP questionsGabor-Granger technique
Ask directly about purchase intent at different fixed pricesA/B testing on your pricing page
For high-traffic, low-touch businesses, this can validate simple price moves
These methods are less powerful than conjoint but can still provide directional guidance until you’re ready for a full conjoint study.
Conjoint Analysis vs Other Pricing Research Methods
How does conjoint analysis compare to other pricing research methods?
Van Westendorp Price Sensitivity Meter (PSM)
Respondents answer four questions:
- At what price is it too cheap (quality concerns)?
- At what price is it cheap (good value)?
- At what price is it expensive (but still worth considering)?
- At what price is it too expensive?
Pros:
- Quick, inexpensive
- Good for defining a price band
Cons:
- Doesn’t handle features or packaging
- Based on self-reported rather than trade-off behavior
Best for:
- Early-stage price exploration
- Commoditized products with simple features
Direct willingness-to-pay (WTP) surveys
You ask:
“What is the maximum you’d be willing to pay for this product per month?”
Pros:
- Simple, easy to implement
- Can be run with smaller samples
Cons:
- People underestimate or overestimate their true WTP
- Doesn’t model trade-offs between price and specific features
Best for:
- Early directional research
- Supplements to more robust methods
Gabor-Granger
You show a product description and ask:
“Would you purchase at $X per month?”
If yes, ask at a higher price. If no, ask at a lower price.
Pros:
- More structured than open-ended WTP
- Outputs a demand curve by price point
Cons:
- Does not account for feature differences or competitive context
- Can still be biased by hypothetical responses
Best for:
- Simple single-plan products
- When you’re only testing price, not packaging
A/B testing
You show different prices (and possibly packages) to different visitors and compare performance.
Pros:
- Based on real buyer behavior
- Strong evidence if you have enough traffic and conversions
Cons:
- Requires significant traffic and control
- Risky: You can lose real revenue while testing
- Hard to test complex multi-attribute changes or multiple combinations
Best for:
- Incremental optimization of an already decent pricing page
- Confirming specific changes after research
Conjoint vs. other methods: comparison table
| Method | Handles Features + Price | Simulates Trade-offs | Works with Low Traffic | Good for SaaS Packaging? |
|----------------------------|--------------------------|----------------------|------------------------|---------------------------|
| Conjoint Analysis | Yes | Yes | Yes (survey-based) | Excellent |
| Van Westendorp (PSM) | No | No | Yes | Limited |
| Direct WTP Surveys | No (unless you add them) | Weakly | Yes | Limited |
| Gabor-Granger | Partially | Limited | Yes | Moderate |
| A/B Testing | Yes | Yes | Needs traffic | Good for tweaks |
Conjoint is often the most powerful when you’re making big pricing and packaging decisions. Other methods are useful for early exploration or incremental refinement.
Real-World Example: SaaS Pricing Study
To make this concrete, here’s an anonymized example from a B2B SaaS company.
The pricing challenge
A mid-market SaaS company (~$25M ARR) offering workflow automation had:
- 3 pricing tiers that had grown organically over time
- Sales pushing heavy discounting and custom deals
- Confusion over which features belonged in which tier
- A looming price increase they were nervous about
They wanted to:
- Re-architect tiers to better reflect value
- Understand willingness to pay by segment
- Reduce discounting without killing win rates
Study design and methodology
We designed a choice-based conjoint study with:
Attributes:
Price per month (as ACV ranges)
Number of workflows
User seats
Integrations (standard vs. premium)
SSO
Support level
Implementation services
Segments:
SMB (<100 employees)
Mid-market (100–1000)
Enterprise (>1000)
Sample size:
~400 qualified decision-makers and influencers
Split across segments and industries
We also ran a MaxDiff exercise up front to prioritize features and avoid an overloaded conjoint design.
Key findings and insights
Some of the surprising outcomes:
- SSO was more valuable to mid-market than to the very largest enterprises, who already assumed it was table stakes and focused more on advanced administration.
- Customers were willing to pay significantly more for implementation and onboarding services than the company had been charging.
- A subset of SMB customers valued premium integrations enough to choose a higher-priced tier that included them.
From the conjoint simulator, we identified:
- A new three-tier structure (with a clearer good-better-best ladder)
- Optimal price bands for each tier
- Features that were safe to move up-market without hurting conversion
- Features that needed to remain broadly available to avoid churn
Business impact and results
After implementing the new packaging and pricing:
- Average contract value (ACV) increased by ~18% within 6 months
- The discount rate in new deals dropped by 30%+
- Win rates held steady or improved in the mid-market segment
- Sales had a clearer narrative for each tier’s value proposition
The cost of the conjoint study was a tiny fraction of the incremental annual revenue unlocked by better pricing decisions.
Running a Conjoint Study: Step-by-Step Process
If you’re considering a conjoint study for your B2B SaaS pricing, here’s how it typically works.
Step 1: Define objectives and key questions
Before touching survey tools, get crystal clear on:
- What decisions will this study inform?
- Price points?
- Feature-tier mapping?
- Packaging structure?
- What are you willing to change based on the results?
- Which segments do you need to understand separately?
Common objectives:
- “Redesign our pricing tiers for self-serve and sales-led customers.”
- “Determine how much more we can charge for SSO and advanced analytics.”
- “Decide whether to make a new feature an add-on or include it in Pro.”
Step 2: Select attributes and levels
Attributes are the building blocks of your conjoint:
- Price
- Seats/usage
- Key features or modules
- Support/SLAs
- Contract terms
For each attribute, define levels that are:
- Realistic (you’d actually offer them)
- Distinct (meaningful differences)
- Limited (too many levels = cognitive overload)
Example:
- Price: $99, $149, $199, $249
- Users: 10, 25, 50, 100
- Support: Email only, Business-hours chat, 24/7 support
This step often requires cross-functional alignment (product, finance, sales).
Step 3: Choose conjoint type and experimental design
For most B2B SaaS pricing research, you’ll choose:
- Choice-Based Conjoint (CBC) as the core method
- Possibly add MaxDiff for feature prioritization
Then you’ll work on the experimental design:
- How many attributes per concept?
- How many levels per attribute?
- How many choice tasks per respondent?
- How many alternatives per task?
Tools like Sawtooth Software, Conjointly, or Qualtrics help generate efficient designs that ensure:
- You can estimate utilities accurately
- Respondents aren’t overloaded
Step 4: Determine sample size requirements
Sample size depends on:
- Number of attributes and levels
- Number of segments you want to analyze separately
- Budget and timeline
As a rule of thumb for B2B SaaS:
- N = 200–300 total is often the minimum for robust insights
- N = 400–600 is common if you’ll slice by 2–3 segments (e.g., SMB vs. mid-market vs. enterprise)
For very high-ACV, narrow markets, you may work with smaller but highly qualified samples.
Step 5: Program the survey
This includes:
- Intro and screener questions (to ensure the right audience)
- Contextual description of your product/category
- MaxDiff (if used) for feature importance
- Conjoint choice tasks (8–12 tasks per respondent is typical)
- Profiling questions (company size, role, industry, tech stack)
You’ll want to:
- Use clean, jargon-free language
- Make the tasks look like real pricing cards or plan comparisons
- Test on desktop and mobile
Step 6: Recruit respondents
For B2B SaaS, the best respondents are:
- Actual buyers or strong influencers
- Who match your ICP (industry, company size, region)
- And have real experience with tools like yours
You can source them via:
- Professional panel providers with B2B targeting
- Your own customer and prospect lists (carefully sampled)
- Partnering with specialized B2B research agencies
Incentives should reflect the seniority and time required (e.g., gift cards, donations, or credits).
Step 7: Analyze results
This is where the statistics come in. Common techniques include:
- Hierarchical Bayesian (HB) estimation for CBC
- Latent class segmentation to find distinct preference clusters
Outputs typically include:
- Part-worth utilities for each attribute level
- Relative importance of attributes
- Market share simulations across pricing scenarios
- Willingness-to-pay ranges for key features
At Monetizely, we pair pricing strategists with specialist statistical analysts to ensure the model is both technically sound and business-relevant.
Step 8: Apply insights to your pricing strategy
Analysis is only valuable if it changes real decisions. This is where we:
- Build and compare pricing and packaging scenarios
- Visualize market share and revenue impacts of each scenario
- Tailor recommendations for:
- Self-serve vs. sales-assisted motions
- SMB vs. enterprise offers
- Existing customers vs. new customers
We then help translate insights into:
- Revised pricing pages
- Sales enablement materials
- Upgrade paths and migration strategies for existing customers
Timeline and resource requirements
Typical conjoint engagements for B2B SaaS run:
- 2–4 weeks for design and programming
- 1–3 weeks for fieldwork (data collection)
- 2–4 weeks for analysis, simulation, and recommendations
Total: 5–10 weeks, depending on complexity and internal decision cycles.
Internal stakeholders usually include:
- Product or pricing lead (driver)
- Finance/RevOps
- Product marketing
- Sales leadership for sign-off
Common Mistakes and How to Avoid Them
Too many attributes (cognitive overload)
Trying to cram every single feature into the conjoint is a recipe for bad data.
Avoid by:
- Using MaxDiff to narrow to key drivers first
- Grouping related items into broader attributes (e.g., “Advanced analytics” bundle)
Unrealistic combinations
If your design allows impossible combos (e.g., “Enterprise support at the lowest price”), you’ll get distorted results.
Avoid by:
- Adding constraints to the experimental design
- Carefully reviewing sample concepts before launch
Wrong audience/sample
If your respondents aren’t similar to your real buyers, your insights won’t generalize.
Avoid by:
- Tight screening criteria (decision-maker vs. casual user)
- Validating with your sales and CS teams who really buys
Ignoring competitive context
Conjoint in a vacuum can overestimate your pricing power.
Avoid by:
- Including competitor-like profiles where appropriate
- Framing tasks as, “Imagine you’re choosing between these vendors…”
Not testing the questionnaire
A single confusing question can tank response quality.
Avoid by:
- Running a pilot test with 10–20 respondents
- Watching real people complete the survey (if possible)
Misinterpreting statistical significance
Not every difference is meaningful in practice.
Avoid by:
- Looking at confidence intervals and effect sizes
- Focusing on material business impact, not just statistical noise
Pro Tip:
Treat conjoint outputs as decision support, not gospel. Combine them with qualitative feedback, sales input, and financial modeling to build a robust pricing strategy.
Tools and Software for Conjoint Analysis
Several tools can support a conjoint study for SaaS pricing.
Sawtooth Software
- Considered the industry standard for conjoint and choice modeling
- Very flexible and powerful for complex CBC and adaptive conjoint designs
- Often used by professional pricing and market research firms
Best for: Teams working with experts who can fully leverage its capabilities.
Qualtrics
- Enterprise survey platform with conjoint modules
- Good for organizations that already use Qualtrics for broader research
- Integrates well with other customer insight programs
Best for: Larger organizations with internal research or analytics teams.
Conjointly
- Online platform focused specifically on conjoint and pricing research
- More guided interface with templates and recommendations
- Suitable for mid-market teams that want more self-service options
Best for: SaaS companies that want to run occasional conjoint studies without building heavy in-house expertise.
R packages (for technical teams)
If you have strong technical and statistical skills, open-source options exist:
support.CEs,conjoint,bayesm, and others in R- Highly flexible but require serious quant skills and time
Best for: Data science teams with bandwidth to own the full modeling process.
When to DIY vs hire experts
DIY can work when:
- You’re running a simple conjoint with a few attributes
- You have internal analytic capabilities
- The decision risk is moderate
Hiring experts is usually better when:
- You’re making high-stakes pricing decisions (e.g., enterprise plans, global rollout)
- You need to align multiple stakeholders behind one pricing direction
- You want a partner to go from research → strategy → implementation
For most B2B SaaS companies in the $5M–$300M ARR range, partnering with specialists yields a much better ROI on the study.
Conclusion
Conjoint analysis is one of the most powerful pricing research methods available to B2B SaaS companies. By forcing respondents to make realistic trade-offs between features, packages, and prices, it helps you:
- Quantify willingness to pay
- Design value-aligned pricing tiers
- Prioritize features and investments
- Simulate how changes will affect conversion, market share, and revenue
It’s especially worth the investment when:
- You’re past $1M ARR and pricing decisions have real revenue impact
- You’re rethinking pricing and packaging or launching a major new product
- You want to move from opinion-based to evidence-based pricing
If you’re unsure whether conjoint analysis is the right fit—or which pricing research approach would deliver the best ROI—getting expert guidance upfront can save substantial time and cost.
Not sure if conjoint analysis is right for your pricing strategy? Book a free 30-minute pricing consultation with our team to discuss your specific situation and explore the best research approach for your SaaS business.