4 Pricing Research Methods for SaaS Founders

What Is Pricing Research, and Why Do Most SaaS Founders Skip It?

Pricing research is the structured alternative to guessing — and most SaaS founders skip it entirely. Only 6% of SaaS companies have done sophisticated research on buyer needs and willingness to pay, per OpenView’s 2021 survey of 2,200 companies. Another 48% have done none at all.

Founders searching for help hit a wall immediately. A search for “pricing research methods for SaaS founders” returns pricing model content in 6 of the top 7 results — tiered pricing, usage-based pricing, per-seat pricing. None of it answers the actual question: how do you find the right number once you’ve picked a model?

Six of seven top search results cover pricing models, not pricing research methods.

This gap compounds as pricing models keep shifting. The move away from per-seat pricing toward usage-based and outcome-based models makes it harder to guess a number that holds — you need real customer input, not last year’s competitor benchmark. That evidence gap is also a trust gap: buyers increasingly question SaaS pricing they can’t see the logic behind, and research is how you build a price you can defend.

The good news: running this research no longer requires a consulting budget. It requires knowing which of four methods to use, and when.

How Does the Van Westendorp Price Sensitivity Meter Find Your Acceptable Price Range?

The Van Westendorp method asks four questions and gives you a range, not a number. Developed by Dutch economist Peter van Westendorp in 1976, it remains one of the most widely used pricing survey techniques in market research (Wikipedia, 2026). You ask each customer:

• At what price would this be so cheap you’d question the quality?
• At what price does this start to feel like a bargain?
• At what price does this start to feel expensive, but you’d still buy it?
• At what price would this be so expensive you wouldn’t buy it?

Plot the four response curves and you get three critical intersections: the Point of Marginal Cheapness, the Point of Marginal Expensiveness, and the Indifference Price Point — the acceptable range between them, plus the point where “cheap” and “expensive” perceptions cancel out (Sawtooth Software, 2026).

The four Van Westendorp curves intersect to mark the acceptable price range.

Run this before you’ve committed to a number — it’s built for products without an obvious reference price. You want it when you’re launching something new, not when you’re auditing an existing subscription tier against three established competitors.

Van Westendorp gives you a range. It won’t tell you if customers will actually pull out their card at a specific number inside that range — that’s a different question, and a different method.

How Does Gabor-Granger Test a Specific Price Point?

Gabor-Granger doesn’t ask customers to imagine a range — it asks a yes-or-no question at a real number. Developed in the 1960s by economists André Gabor and Clive Granger, the method shows one respondent one price, then adjusts: say yes, the next price is higher; say no, it drops. This repeats until each respondent’s ceiling is found (IntelliSurvey, 2025).

• Every respondent lands somewhere on a demand curve — the price where they switched from “yes” to “no.”
• Aggregate those switch points and you get the percentage of customers willing to buy at each price level.
• Multiply demand by price at every point and you find the revenue-maximizing price — not just the highest one customers will tolerate (QuestionPro, 2025).

Run this on a product you already sell, not one you’re about to launch. Gabor-Granger assumes customers already have some reference point for value — new, unfamiliar products don’t give clean answers.

Gabor-Granger adjusts price up or down per respondent to trace the demand curve.

Gabor-Granger tells you what price maximizes revenue on your existing product. It says nothing about which features are actually driving that willingness to pay — that’s a feature question, and it needs a different method.

How Does MaxDiff Reveal Which Features Customers Actually Value?

Two methods down — Van Westendorp and Gabor-Granger both test a single price at a time. The next two work differently: they test features and bundles instead.

MaxDiff exists because rating scales lie. Ask customers to rate 15 features from 1–5, and you get a wall of 4s and 5s — everyone claims everything matters (SurveyMonkey, 2026). MaxDiff, short for Maximum Difference Scaling, fixes this by showing small sets of 4–5 items at a time and forcing a best/worst choice for each set (B2B International, 2026).

• Show respondents 8–30 features total — fewer and you might as well rank them directly; more and the survey gets exhausting (GLG, 2025).
• Every choice is relative, not absolute — there’s no scale to inflate, so priorities separate cleanly (Bentley University, 2026).
• Repeat across 150–400 B2B respondents and you get a statistically stable hierarchy, not a guess dressed up as data (B2B International, 2026).

Run this before you commit to a roadmap or a tier structure, not after. MaxDiff is an earlier-stage tool — it narrows a long list down to what’s worth testing further, rather than finalizing a price or package (Werk Insight, 2025).

MaxDiff forces trade-offs where rating scales let every feature score a 5.

MaxDiff hands you a ranked list of what matters. Turning that ranking into an actual tier structure and a price is a packaging question — and that’s where Conjoint Analysis takes over.

How Does Conjoint Analysis Price Bundles and Tiers?

Conjoint analysis is the only method on this list that tests price and features together. Choice-Based Conjoint (CBC), the version most SaaS teams use, shows respondents complete packages — not isolated prices or isolated features — and asks them to pick the one they’d buy (Monetizely, 2025). That choice forces a real trade-off, the same one your customer makes at checkout.

• Respondents choose between full profiles: a feature set bundled with a specific price, repeated across several rounds.
• The output is a set of utility scores — what each feature is actually worth to the customer, in price terms, not opinion.
• Those utilities let you simulate demand for any tier structure before you build it (Umbrex, 2026).

Sawtooth Software recommends a minimum of 200 respondents for a basic CBC study — fewer, and the trade-off patterns don’t stabilize enough to trust. Run this when you’re redesigning tiers or deciding what goes into “good, better, best” — not for a single new price point.

Conjoint analysis shows which features drive customers toward a higher-priced bundle.

Running conjoint well means understanding what your packaging decisions actually cost you to maintain — the same discipline that applies to tracking your own AI infrastructure spend as it scales. Get the tier structure wrong and you’re not just leaving revenue on the table; you’re building support and engineering overhead around packages nobody actually wanted.

Four methods, four different questions. The next section shows you which one to run first, based on where you actually are.

The Pricing Proof Ladder — Which Method Should You Run First?

Four methods, four different questions — and most founders don’t need all four at once. The Pricing Proof Ladder sequences them by what you actually know about your product right now, from “I have no idea what to charge” to “I need to redesign my tiers.” Climb it in order, and each method builds on what the last one told you.

The Pricing Proof Ladder — four methods, sequenced by what you already know.

Conclusion

The Pricing Proof Ladder gives you a starting point, not a mandate: Van Westendorp for a new product, Gabor-Granger for an existing one, MaxDiff before a roadmap call, Conjoint when tiers need rebuilding. The real shift isn’t which method you pick — it’s running any of them at all.

Only 6% of SaaS companies have done this kind of research, and self-serve tools have removed the excuse that it’s too expensive. Start with the rung that matches where you are today.

If you’re still deciding how to structure the price itself, the shift away from per-seat pricing is the next thing worth understanding.