Pricing Quantitative Research for B2B SaaS: A Practical Guide

Pricing research is a broad field. There are surveys, experiments, interviews, and statistical models, and within each category, multiple methods to choose from.

For B2B SaaS companies working on pricing or packaging, four quantitative methods come up most often: MaxDiff, Build-Your-Own, Van Westendorp, and Conjoint Analysis. Each serves a different purpose, and knowing which one to reach for and when can save significant time and sharpen the quality of your pricing decisions.

This article covers all four. For each method, I explain what it is, when it is most useful, and what it looks like in practice.

The goal is not to run all four. The goal is to pick the right tool for the question you are trying to answer.

Start with Customer Interviews

Before running any survey, spend time talking to customers.

Quantitative research is only as good as the hypotheses behind it. Customer interviews help you understand which features customers actually value, how they describe their own problems, and what tradeoffs matter to them. Without this foundation, you risk designing surveys around the wrong variables.

Through interviews, you will typically surface a shortlist of features or capabilities that customers consistently mention. These become the inputs for your quantitative work.

Method 1: MaxDiff — Prioritize What Customers Value

What it is

MaxDiff (Maximum Difference Scaling) measures the relative importance of a set of items. Respondents are shown small sets of typically four to five options and asked two questions: which item is most important, and which is least important. This repeats across multiple sets, with items rotating in and out so that every item appears roughly the same number of times.

The forced best-and-worst choice is what makes MaxDiff analytically powerful. Unlike a standard rating scale where respondents might rate everything as "very important," MaxDiff forces real prioritization.

When to use it

MaxDiff is best used early in the research process, when you need to prioritize a list of features or positioning themes. It answers the question: of everything we offer, what do customers care about most?

It is particularly useful when you have a long list of potential features and need to identify which ones genuinely drive perceived value versus which ones are nice-to-have.

What it does not tell you

MaxDiff does not attach pricing to items. It tells you what customers prefer, not what they will pay for.

Example

Imagine you are a project management SaaS company considering which features to highlight in your next pricing tier. You have ten candidate features and want to know which ones customers rank highest.

A MaxDiff question might look like this:

From the following features, select the one that is MOST important to you and the one that is LEAST important to you.

• Real-time collaboration across teams
• Custom workflow automation
• Advanced reporting and analytics
• Single sign-on (SSO)
• API access for integrations

After running this across multiple sets, you might find that custom workflow automation and real-time collaboration consistently score as the top two, while SSO and API access score lower among your mid-market segment. That insight directly informs both your packaging and your messaging.

Method 2: Build-Your-Own (BYO) — Explore Packaging from Scratch

What it is

Build-Your-Own research presents respondents with a menu of features and asks them to select what they would include in a product at a given price. Typically, the exercise is run at multiple price points, and researchers observe which features are consistently selected regardless of price versus which features only appear at higher price points.

When to use it

BYO is most valuable when you are starting from scratch, when you do not yet have a packaging hypothesis and need to understand how customers naturally group features and what they expect within a given price range.

It is also useful for identifying which features are considered table stakes (always selected) versus premium differentiators (only selected at higher price points).

What it does not tell you

BYO gives you directional packaging themes and rough price ranges. It does not give you precise willingness to pay or statistically rigorous preference weights. Think of it as discovery research, not validation research.

Example

You are a CRM platform company launching a new pricing structure. You present respondents with a list of twelve features and ask them to build their ideal plan at a $50 per seat per month price point.

You are building a CRM plan priced at $50/seat/month. Select the features you would expect to be included:

• Contact and deal management
• Email integration
• Pipeline reporting
• AI-assisted follow-up suggestions
• Custom fields and objects
• Territory management
• Advanced forecasting
• Dedicated account manager
• API access
• Data enrichment
• Workflow automation
• Revenue intelligence dashboards

You then repeat the exercise at $100 and $150 per seat per month.

You might find that contact management, email integration, and pipeline reporting are selected at every price point. These are your Core tier features. AI-assisted follow-up and workflow automation start appearing consistently at $100. Revenue intelligence and dedicated account management only appear at $150.

This gives you a natural three-tier packaging hypothesis before you run any more sophisticated analysis.

Method 3: Van Westendorp — Identify the Acceptable Price Range

What it is

Van Westendorp is a straightforward price survey built around four questions:

• At what price would this product seem so cheap that you would question its quality? (Too cheap)
• At what price would this product seem like a great deal? (Acceptable low end)
• At what price would this product start to feel expensive, but you would still consider buying it? (Acceptable high end)
• At what price would this product be so expensive that you would not consider buying it? (Too expensive)

By plotting the distribution of responses across all four questions, you can identify the acceptable price range, the zone where most respondents find the price credible and reasonable.

In practice, I often simplify this to just the two boundary questions: too cheap and too expensive. This gives you a clean picture of the price corridor without over-complicating the survey.

When to use it

Van Westendorp works best when your packaging is already defined and you simply need to understand the price range customers will accept. It is particularly useful for straightforward, clearly understood capabilities, for example, pricing an AI writing add-on or a standalone analytics module.

It is also a good fit when speed matters. The survey is short, easy to field, and fast to analyze.

What it does not tell you

Van Westendorp does not measure price sensitivity relative to specific features or force tradeoffs. It tells you the acceptable range, not the optimal price point within that range, and not how demand changes as price moves through that range.

Example

You are preparing to launch an AI summarization add-on for your document management platform. The feature is well defined. You need to know what price range the market will accept.

Your Van Westendorp question set might look like this:

We are considering an AI summarization add-on that automatically generates summaries for uploaded documents and meeting notes.

• At what monthly price per seat would this add-on seem so cheap that you would question its quality?
• At what monthly price per seat would this add-on seem like a great deal?
• At what monthly price per seat would this add-on start to feel expensive, but you would still consider it?
• At what monthly price per seat would this add-on be too expensive to consider?

If respondents cluster the "too cheap" responses around $5, the "great deal" around $15, the "getting expensive" around $30, and the "too expensive" around $50, your acceptable price range is roughly $15 to $30 per seat per month.

You now have a validated corridor to work within and a clear ceiling to avoid breaching.

👉 See related post: Mastering the AI Pricing Survey

Method 4: Conjoint Analysis — Test Tradeoffs and Stress-Test Packaging

What it is

Conjoint analysis is the most analytically rigorous of the four methods. It presents respondents with a series of product configurations, each varying across multiple attributes including price, and asks them to choose their preferred option.

By analyzing which configurations respondents choose across many comparisons, conjoint reveals the relative importance of each attribute and the willingness to pay for individual features. It also enables simulation modeling: you can test how demand shifts as you adjust features or prices across different tiers.

When to use it

Conjoint is best used when you already have a packaging hypothesis and want to stress-test it. It requires more design effort upfront, typically four to six attributes including price, and is more demanding for respondents than the other methods.

Used well, conjoint analysis is the most powerful tool in the pricing research toolkit. It lets you run "what if" simulations: what happens to conversion if we move from $75 to $90? What if we remove a feature from the mid tier?

What it does not tell you

Conjoint requires a well-formed hypothesis to be useful. If you go into a conjoint study without a clear sense of your packaging options and price range, you will struggle to design the attributes correctly. This is why BYO and Van Westendorp typically come first.

Example

Building on the CRM example above, you now have a packaging hypothesis from your BYO research and a price corridor from Van Westendorp. You want to validate the packaging and understand demand sensitivity across tiers.

A conjoint question might present respondents with three product profiles side by side:

Which of the following CRM plans would you choose?

Respondents choose across dozens of such configurations, with attributes and prices varying systematically. The resulting data tells you the utility value of each feature and how demand shifts as prices change.

You might discover, for example, that workflow automation drives significantly more conversion to the mid tier than AI follow-up suggestions, which informs both your packaging and your feature marketing.

You Do Not Need to Run All Four

One of the most common mistakes in pricing research is over-engineering the process. Here is a practical guide to selecting the right method:

• If you need to prioritize features or positioning themes: MaxDiff alone may be sufficient.
• If you are starting from scratch on packaging: BYO is your best first step. MaxDiff can complement it.
• If your packaging is defined and you need a price range quickly: Van Westendorp is the right tool. It is fast, easy to field, and directionally reliable.
• If you need to validate packaging and model demand sensitivity: Conjoint analysis is the right investment.
• If you are pricing a simple add-on or new feature: Van Westendorp is usually all you need. A straightforward AI add-on does not require a conjoint study.
• If you are repricing a complex multi-tier product: Start with BYO and Van Westendorp, then move to conjoint for validation.

Most companies also have a solid informal understanding of which features customers value through support interactions, sales conversations, and product usage data. If that foundation is already strong, you can often skip MaxDiff and BYO entirely and go straight to conjoint.

Survey Distribution: Who to Ask and How

Existing customers

Existing customers are the easiest to access and the most valuable for most pricing research. They understand your product deeply and can give informed feedback on features, packaging, and price.

They are especially useful when you are researching new features or add-ons, because they are not worried about changes to their existing contract. They evaluate the new capability on its own merits.

Common distribution channels for existing customers include email outreach and in-app surveys. For shorter surveys like Van Westendorp or MaxDiff, tools like Typeform or Youform work well.

Non-customers and market panels

If you need to understand the broader market, especially for pricing a new product, re-pricing existing products, or entering a new segment, you can purchase survey samples from research panel providers.

These providers let you define respondent criteria by industry, company size, job title, or geography, and include screening questions to qualify respondents. Reputable providers include Conjointly and Respondent.io.

Panel research is particularly useful for conjoint studies, where you need a larger, statistically sufficient sample to generate reliable preference weights.

Closing Thought

The best pricing decisions are grounded in data: internal signals like deal analysis and feature usage, and external research like the methods covered in this article.

Build a lightweight but consistent customer research practice. Run surveys more frequently than you think you need to.

The companies that make pricing a continuous discipline rather than a one-time project consistently make faster and more confident pricing decisions.

⚡Want a Fast AI Pricing Diagnostic?

I share pricing diagnostics in a 15 to 20 minute intro call.

📩 helenchou@helenc.cc