TL;DR
- Three proven methodologies: Van Westendorp identifies acceptable price ranges through four perception questions. Gabor-Granger builds purchase-intent demand curves across discrete price points. Conjoint analysis reveals how customers trade off price against features.
- Sample size matters: Gabor-Granger and Van Westendorp tests need 150–200 respondents per segment minimum. Live A/B price tests require 1,000+ observations per variant for statistical significance.
- B2B SaaS is relatively inelastic: Enterprise customers typically show elasticity coefficients between -0.5 and -0.9. SMB customers run -1.2 to -1.8. Segment your analysis — aggregate elasticity obscures actionable signals.
- Apply a revealed preference discount: Survey-based tests consistently overstate purchase intent by 10–20% compared to actual buying behavior. Adjust your revenue projections accordingly.
- Build the full revenue model: A price increase that looks good on conversion metrics can produce net-negative ARR within 12 months if it raises churn. Run the complete unit economics before acting.
Pricing is one of the highest-leverage decisions an operator makes, and one of the least systematically tested. Most SaaS companies set prices based on competitive benchmarking, internal cost calculations, or founder intuition — then leave them unchanged for years. Companies that use rigorous value-based pricing methodologies report up to 25% higher revenue growth compared to those relying on cost-plus or competitor-based approaches alone.
The gap between companies that test pricing and those that do not is not primarily a resource gap. It is a methodology gap. Most teams do not have a structured template for designing a test, collecting data, analyzing results, and translating findings into a defensible pricing decision. This post provides exactly that.
The template below covers three established methodologies — Van Westendorp, Gabor-Granger, and conjoint analysis — with a test design worksheet, data collection fields, analysis framework, and decision criteria for each. Use whichever method fits your stage and objective; the decision framework at the end applies to all three.
The Three Methodologies: When to Use Each
Before selecting a testing approach, it helps to understand what each method actually measures and where each one breaks down.
Van Westendorp Price Sensitivity Meter
Developed by Dutch economist Peter van Westendorp in 1976, the Price Sensitivity Meter (PSM) is a survey-based technique that maps consumer price perception across four distinct thresholds. It is the right starting point when you are launching a new product, entering a new tier, or have no behavioral baseline to work from.
The method asks four questions for each respondent:
- Too cheap: "At what price would this product be so inexpensive that you would question its quality?"
- Bargain: "At what price would this product represent a great value — a real bargain?"
- Expensive but acceptable: "At what price would this product start to feel expensive, but you would still consider buying it?"
- Too expensive: "At what price would this product be so expensive that you would not consider buying it, regardless of quality?"
The four responses are plotted as cumulative distribution curves. Their intersections define four key thresholds: the Point of Marginal Cheapness (PMC), the Point of Marginal Expensiveness (PME), the Acceptable Price Range (APR) between them, and the Optimal Price Point (OPP) where the maximum number of respondents find the price neither too cheap nor too expensive.
The PSM's primary limitation for SaaS is that it is static — it captures a single moment of perception without accounting for feature changes, competitive context, or the effect of pricing on retention rates and LTV. It is best used as a first-pass instrument, not as the final word on pricing.
Gabor-Granger Method
Developed by economists Clive Granger and André Gabor in the 1960s, the Gabor-Granger method measures purchase intent across a range of discrete price points to construct a demand curve. It is the most direct route to identifying the revenue-maximizing price point within a defined range.
The mechanics are straightforward: respondents are shown a series of price points (typically five) and asked at each one whether they would purchase. The response to each price generates a percentage of respondents who would buy at that level. These percentages, when plotted against price, produce a demand curve — and when multiplied by price, a revenue curve whose peak identifies the optimal price.
The Gabor-Granger approach has a known limitation: because respondents see all price points sequentially, earlier exposures can anchor later responses (a "price anchoring" bias). Research design typically mitigates this through randomized price order presentation. A minimum of 150 respondents is recommended for stable curves; below 100, the confidence intervals become too wide to support a pricing decision.
Conjoint Analysis
Conjoint analysis is the most rigorous and resource-intensive of the three methods. Rather than asking respondents to evaluate a product at a given price in isolation, conjoint analysis presents a series of product configurations — each with different combinations of features and price — and asks respondents to choose between them or rank their preferences.
The statistical output, typically derived through Choice-Based Conjoint (CBC) analysis, reveals part-worth utilities: how much value respondents assign to each feature attribute and to each price level. This allows you to calculate willingness to pay for specific features, identify which feature combinations justify premium pricing, and model how demand shifts when you add or remove features at different price points.
For most SaaS teams below $20M ARR, CBC conjoint is overkill as an initial pricing instrument. It requires specialized survey design software, statistical expertise to analyze correctly, and significantly larger samples. Its natural home is evaluating major pricing model changes — moving from per-seat to usage-based, or restructuring tier boundaries — where the interaction between features and price is what you need to understand.
Price Elasticity Test Template
The following template is structured in four sections: test design, data collection, analysis framework, and decision criteria. Fill in the design section before fielding any survey. Complete the data collection and analysis sections as results come in. Use the decision criteria section to translate findings into a specific pricing recommendation.
Section 1: Test Design Worksheet
Test Design Fields
Test objective
[ ] Identify acceptable price range for new product/tier
[ ] Find revenue-maximizing price point within existing range
[ ] Evaluate how customers trade price against specific features
[ ] Validate a proposed price increase before rollout
Method selected
[ ] Van Westendorp [ ] Gabor-Granger [ ] Conjoint (CBC)
Target segment
Define by: company size / revenue band / role / industry / current plan tier:
_____________________________________________
Price range to test (Gabor-Granger / Conjoint)
Floor: $________ Ceiling: $________ Step intervals: $________
Number of price points: ________ (recommend 5–7)
Required sample size
Target n per segment: ________ (minimum 150; recommend 200+)
Confidence level required: [ ] 90% [ ] 95% [ ] 99%
Respondent recruitment method
[ ] Email survey to existing customer base (segment: _____________)
[ ] Survey panel (e.g., Respondent.io, Wynter, Qualtrics panel)
[ ] In-app prompt at defined trigger event
[ ] Live A/B test (new visitors only)
Test window
Start date: ________________ End date: ________________
Minimum duration for live tests: 2 full billing cycles (or 4 weeks, whichever is longer)
Section 2: Data Collection Fields
Van Westendorp Data Collection
For each respondent, record:
- Respondent ID | Segment tag | Date collected
- Q1 — Too cheap price: $________
- Q2 — Bargain price: $________
- Q3 — Expensive but acceptable price: $________
- Q4 — Too expensive price: $________
- Respondent role: ________ | Company ARR/revenue band: ________
- Current pricing tool used (if any): ________
Aggregate outputs to calculate:
- Point of Marginal Cheapness (PMC): intersection of "too cheap" and "bargain" cumulative distributions
- Point of Marginal Expensiveness (PME): intersection of "expensive but acceptable" and "too expensive" cumulative distributions
- Optimal Price Point (OPP): intersection of "too cheap" and "too expensive" distributions
- Acceptable Price Range (APR): PMC to PME
Gabor-Granger Data Collection
For each price point tested, record:
- Price point | n respondents shown | n who said "would purchase" | Purchase intent % | Revenue index (price × intent %)
- Price Point 1: $________ | n= ________ | Would buy: ________ | Intent %: ________ | Revenue index: ________
- Price Point 2: $________ | n= ________ | Would buy: ________ | Intent %: ________ | Revenue index: ________
- Price Point 3: $________ | n= ________ | Would buy: ________ | Intent %: ________ | Revenue index: ________
- Price Point 4: $________ | n= ________ | Would buy: ________ | Intent %: ________ | Revenue index: ________
- Price Point 5: $________ | n= ________ | Would buy: ________ | Intent %: ________ | Revenue index: ________
Derived outputs:
- Revenue-maximizing price point: $________ (highest revenue index)
- Demand elasticity between adjacent price points: ΔQ%/ΔP% for each interval
- Price point of sharpest demand drop (highest elasticity): $________ to $________
Live A/B Test Data Collection
For each variant, record weekly:
- Variant label | Price shown | n visitors | n trials started | n converted to paid
- Visitor-to-trial rate | Trial-to-paid rate | Revenue per visitor
- Statistical significance at current sample size (use chi-square or Z-test for proportions)
- Segment breakdown: traffic source / geography / device type (to check for randomization balance)
Stop conditions:
Pause test if conversion delta exceeds 40% in either direction before reaching minimum sample size. Document stop reason.
Section 3: Analysis Framework
Once data is collected, run the following analysis sequence before drawing conclusions.
Analysis Steps
Step 1 — Validate sample quality
Check response distribution for logical consistency. In Van Westendorp, flag and remove any respondent where "too cheap" > "bargain," or "bargain" > "too expensive" — these represent invalid responses that will distort intersection calculations.
Step 2 — Segment before aggregating
Calculate elasticity and price thresholds separately for each defined segment before producing an aggregate result. Aggregate elasticity frequently obscures segment-specific signals. Research from Price Intelligently found that elasticity can vary by a factor of 2–4x between enterprise and SMB cohorts in the same product.
Step 3 — Apply revealed preference discount (survey tests only)
Survey-based purchase intent consistently overstates actual buying behavior. Apply a 15–20% downward adjustment to intent percentages before calculating revenue projections. This is a standard correction in commercial pricing research.
Step 4 — Calculate price elasticity coefficients
For each price interval in a Gabor-Granger test: E = (ΔQ% / ΔP%). An elasticity of -1.0 means a 10% price increase drives a 10% reduction in purchase rate. Values between 0 and -1.0 are inelastic (price increase grows revenue). Values below -1.0 are elastic (price increase reduces revenue). B2B SaaS enterprise segments typically fall between -0.5 and -0.9. SMB segments typically fall between -1.2 and -1.8.
Section 4: Decision Criteria
Price test results only become a pricing decision when they are evaluated against four criteria. Use the following checklist before finalizing any recommendation.
Decision Criteria Checklist
1. Statistical validity
[ ] Sample size meets minimum threshold per segment
[ ] Confidence interval at target significance level does not span the decision threshold
[ ] Test ran for the full planned duration (no early stops based on preliminary results)
2. Revenue impact model
[ ] Revenue projection built at segment level, not aggregate
[ ] Revealed preference discount applied to survey-based intent figures
[ ] Churn sensitivity modeled: at what churn uplift does the revenue gain become net-negative?
3. Competitive context
[ ] Proposed new price benchmarked against 3–5 direct competitors
[ ] No test result drives price above the top of your competitive cluster unless supported by demonstrated differentiation
[ ] Confirm whether competitors have changed pricing in the last 90 days
4. Implementation plan
[ ] Grandfathering policy defined for existing customers
[ ] Communication timeline and messaging drafted
[ ] Churn monitoring window established (minimum 60 days post-rollout)
[ ] Rollback threshold defined: if churn rises above ____% in 30 days, revert
Price Elasticity Benchmarks by Segment
Having a benchmark to compare your results against helps distinguish a normal outcome from an anomaly that warrants further investigation. The following ranges are drawn from published pricing research and industry analysis.
| Segment | Typical Elasticity Range | Interpretation |
|---|---|---|
| B2B SaaS — Enterprise | −0.5 to −0.9 | Inelastic; moderate price increases feasible with low conversion risk |
| B2B SaaS — Mid-Market | −0.8 to −1.3 | Near unitary; price increases require value narrative to protect conversion |
| B2B SaaS — SMB | −1.2 to −1.8 | Elastic; price increases will reduce conversion rate meaningfully |
| Consumer SaaS / PLG | −1.5 to −2.5 | Highly elastic; freemium and trial mechanics reduce price pressure at entry |
| Tenured customers (2+ years, deep adoption) | −0.3 to −0.6 | Strongly inelastic; switching cost and workflow dependency dampen sensitivity |
| New prospects (first 30 days of evaluation) | −1.8 to −2.5+ | Highly elastic; no switching cost, full competitive optionality |
The practical implication of these ranges is that the population you test on determines the result you get. If your Van Westendorp or Gabor-Granger sample skews toward tenured enterprise customers, you will produce an inelastic curve that looks like you have broad pricing power. If it skews toward early-stage prospects, you will see the opposite. Be deliberate about who you recruit for each test.
Connecting Test Results to Operating Decisions
A price elasticity test produces numbers. Converting those numbers into a durable pricing decision requires connecting them to the rest of your operating model. A 20% price increase that generates a 15% uplift in revenue per account looks attractive in isolation. But if it also drives a 4% increase in monthly churn, the math inverts within 12 months — the incremental revenue per account is offset by the accelerated account loss.
This is where most pricing exercises break down. The team runs a survey or an A/B test, identifies a revenue-maximizing price point, and acts on it — without modeling the retention implications. A complete pricing decision model needs four inputs: the conversion impact (from the test), the churn sensitivity (estimated from cohort data or stress-tested against a range), the LTV change at the new price, and the impact on new logo acquisition economics.
Operators using Fairview can pull these inputs from a single view — conversion data, cohort churn rates, and margin per account — without assembling them manually across disconnected tools. The decision framework above maps directly to the data Fairview surfaces, so the move from test results to a pricing recommendation stays grounded in your actual operating numbers, not just the survey output.
ProfitWell research indicates that properly managed price increases — informed by elasticity testing, communicated clearly, and paired with grandfathering for existing customers — result in less than 3% additional churn when implemented correctly. The difference between a clean execution and a disruptive one is almost entirely in the preparation: knowing your elasticity, building the revenue model before you act, and having a defined rollback threshold in place.