SaaS Price Elasticity Measurement: How to Test Pricing Before You Commit

Pricing changes are one of the highest-leverage decisions in SaaS — and one of the least rigorously analyzed before execution. The typical process: compare prices to competitors, poll the team, maybe ask a few customers, then change prices and watch what happens to conversion rates and churn.

This is an expensive way to learn. A price increase that triggers 15% churn instead of the expected 5% can take 12–18 months of customer acquisition to recover from. Measuring price elasticity before committing to a change doesn't eliminate uncertainty, but it narrows the range of outcomes from "unknown" to "probabilistic."

This post covers the methods — survey-based, behavioral-data-based, and experimental — for measuring price sensitivity before you change prices.

Understanding Price Elasticity in B2B SaaS

Price elasticity of demand (PED) measures the relationship between price changes and demand changes:

PED = % change in quantity demanded ÷ % change in price

A PED of -0.5 means: if you raise prices 10%, demand drops by 5%. A PED of -1.5 means: if you raise prices 10%, demand drops by 15%.

Inelastic demand (PED between 0 and -1): Customers are relatively insensitive to price. High switching costs, strong product differentiation, mission-critical use cases, and limited alternatives all drive inelasticity.

Elastic demand (PED below -1): Customers respond strongly to price changes. Commodity markets, low switching costs, multiple substitutes, and budget-sensitive buyers drive elasticity.

B2B SaaS typically sits in the inelastic range, but the variance by segment is significant:

• Enterprise with high switching costs and deep integration: PED typically -0.2 to -0.4
• Mid-market with moderate integration: PED typically -0.4 to -0.7
• SMB with low switching costs and direct competitors: PED typically -0.6 to -1.2

Understanding which segment your PED applies to is essential — a price change that is safe for your enterprise segment may be highly risky for your SMB base.

Method 1: Van Westendorp Price Sensitivity Meter (PSM)

The Van Westendorp PSM is the most widely used survey method for initial pricing research. It requires no controlled experiment and can be run in weeks.

The four survey questions:

1. At what price would you consider the product so cheap that you would question its quality?
2. At what price would you consider the product inexpensive but still worth buying?
3. At what price would you consider the product starting to get expensive, but you'd still consider buying it?
4. At what price would you consider the product so expensive you would not consider buying it?

Interpreting the results:

Plot the cumulative frequency distributions for all four questions. The intersections identify:

• Point of Marginal Cheapness (PMC): Intersection of "not cheap enough" and "too cheap" — price below this is perceived as suspiciously cheap
• Point of Marginal Expensiveness (PME): Intersection of "not too expensive" and "too expensive" — price above this loses majority interest
• Optimal Price Point (OPP): Intersection of "too cheap" and "too expensive" — the narrowest resistance point
• Acceptable Price Range: Between PMC and PME

PSM limitations:

• Stated preferences ≠ actual purchase behavior. Customers consistently say they'd pay more in surveys than they do in practice (social desirability bias).
• Discount actual price from PSM results: reduce the OPP by 10–20% to approximate real willingness to pay.
• Minimum sample: 150 responses per segment for statistical validity. Run separate PSM surveys for each ICP segment — the acceptable range can vary 2x between SMB and enterprise buyers.
• PSM works best for the full product price, not for pricing components (add-ons, tiers, etc.) where conjoint analysis is more appropriate.

Method 2: Gabor-Granger Price Testing

Gabor-Granger is a survey method that directly measures purchase probability at different price points:

Ask respondents: "At a price of $X per month, how likely are you to purchase this product?" using a scale from 0 (definitely not) to 100 (definitely would). Repeat for each price point being tested.

Plot the purchase probability against price to generate a demand curve. The area under the curve weighted by price represents expected revenue — the revenue-maximizing price is where price × probability is maximized.

Gabor-Granger advantages over PSM:

• Direct revenue optimization (you're finding max-revenue price, not just acceptable range)
• Easier to analyze and communicate
• Works for component pricing (can test specific add-on or tier prices)

Limitations:

• More affected by hypothetical bias than PSM
• Requires careful price list construction — if your test prices are anchored by the first question, subsequent responses are biased

Method 3: Cohort-Based Behavioral Analysis

The most reliable price elasticity data is behavioral — how did real customers respond to real price differences? If your product has had multiple price points over time, or if you currently sell at different prices to different segments or geographies, cohort analysis can approximate elasticity.

Steps:

1. Identify cohorts of customers acquired at meaningfully different price points (at least 10–15% difference)
2. Compare conversion rates, time-to-close, 90-day churn, and 12-month NRR across cohorts
3. The difference in conversion rates between price cohorts provides a behavioral elasticity estimate

Limitations:

• Historical price changes are often confounded with product changes, market changes, and team composition changes. Cohorts acquired 2 years apart at different prices aren't a controlled experiment.
• This method works best when there is a recent, clean price test to analyze.

Method 4: Staged Price Testing (New Customers Only)

The safest controlled test of pricing is to change prices for new customers while keeping existing customers at their current rate. This isolates the conversion-rate elasticity effect without exposing existing ARR to churn risk.

Process:

1. Define the new price point you want to test (typically 15–25% above current for an increase test)
2. Launch new pricing for new customer sign-ups (new page, new pricing visible to new visitors)
3. Measure conversion rate (free-to-paid or trial-to-paid) for 6–8 weeks at the new price
4. Compare to the conversion rate from the previous 6–8 weeks at the old price (with seasonality adjustment)

Sample size calculation:

For a 5% absolute conversion rate and a test designed to detect a 20% relative change (from 5% to 4% or 6%), you need approximately 3,800 visitors per variant. Use a statistical power calculator with 80% power and 5% significance level to size your test based on your actual traffic.

Limitations:

• Conversion rate elasticity ≠ retention elasticity. New customers may convert at a lower rate under higher pricing, but existing customers may churn more (or less) in response to price increases. Staged testing measures conversion impact but not retention impact.
• Seasonal effects can confound short tests — compare same-period data from the prior year to adjust.

Method 5: Retention Analysis After Historical Price Changes

If you've ever raised prices for existing customers, the churn rate in the 90 days following that change provides direct evidence of price elasticity:

Analysis framework:

1. Calculate the churn rate in the 90 days immediately before the price change (baseline)
2. Calculate the churn rate in the 90 days immediately after the price change
3. Attribute any statistically significant increase above baseline to the price change
4. PED estimate: divide the % change in churn rate by the % change in price

Example:

• Baseline churn: 2.5%/month
• Post-price-increase churn (15% price increase): 3.5%/month
• Churn increase: +1 percentage point = +40% relative change
• PED estimate: -40% demand change / +15% price change = PED of approximately -2.7 (elastic)

This PED estimate suggests the price increase was too aggressive — a 15% increase that generates a 40% relative churn increase destroys ARR. Recalibrate the target price increase downward.

Applying Elasticity Data to Pricing Decisions

Once you have elasticity estimates, the decision framework for a price change:

If PED is -0.4 to -0.6 (moderately inelastic): A 20% price increase generates approximately 8–12% demand reduction. Revenue impact: $100 existing ARR × 0.80 remaining customers × 1.20 new price = $96 — a 4% net ARR reduction from price-increase churn, offset by expansion revenue from the 80% who stay. Long-term win if acquisition rate holds or improves.

If PED is -0.7 to -1.0 (approaching elastic): A 20% price increase generates 14–20% demand reduction. Revenue impact is approximately neutral or slightly negative. Recommended approach: 10–12% increase with strong grandfathering of existing customers, staged rollout to new customers first.

If PED is below -1.0 (elastic): Price increases are revenue-destructive at this elasticity. The correct response is not to accept permanent underpricing, but to first improve the product differentiation or switching cost to reduce elasticity before raising prices.

The SaaS price increase playbook covers the execution process after you've determined the right price point through elasticity measurement.

The Measurement Stack

For a complete pricing intelligence function, run all three methods in sequence:

1. Van Westendorp PSM (weeks 1–4): Identify the acceptable price range and optimal price point for new buyers. Run by ICP segment. Discount by 15% from survey results.
2. Cohort behavioral analysis (weeks 1–4, parallel): Extract real conversion and churn data by historical price point from your CRM and billing system.
3. Staged new-customer test (weeks 4–12): Test the price point identified by PSM with real new customers. Measure conversion rate at statistical significance.
4. Post-increase retention analysis (ongoing): If you proceed with a price increase, track churn for 90 days post-implementation and compare to the pre-increase baseline.

The combination of stated-preference survey data and behavioral data from cohort and staged testing produces the most reliable estimate of true price elasticity in your market. Neither method alone is sufficient — stated preferences overestimate willingness to pay, and behavioral data without survey context lacks the causal explanation for observed patterns.