01 / Institute of Directors · supplier case study
Appeal first, then a price range
Current and prospective members rated three packages and answered price-range questions. The result was IoD Advance.
example_iod →Price and proposition testing
Price and proposition testing before members vote with their feet.
Price and proposition testing checks a fee change, a new tier or a new package before you launch it. Pick the method that answers your question, sample the right people, then confirm it on what new joiners actually do.
Test before you change the fee, a tier, the package, the benefits
From question to pricethe order we work in
- 01Decidewhat would change, and for whomSCOPE
- 02Askrange, price, package or benefitsMETHOD
- 03Samplecurrent, lapsed and prospectsWHO
- 04Test livenew joiners first, then renewalsBEHAVIOUR
- 05Readby segment, against cohort renewalDECIDE
Decided onwhat people do
300
Respondents: a common starting sample for a conjoint study
Sawtooth Software, practitioner guidance ↗32%
Of 69 published choice experiments used fewer than 100 people
de Bekker-Grob et al. 2015, peer-reviewed ↗01/ when to test
Test before the fee or the package changes.
A price test earns its cost when a change is hard to undo. An index-linked rise needs good notice, not research. A new tier, a new package or a new membership model needs evidence first.
Test the proposition before the price. The Royal College of Anaesthetists is asking members to choose between three category models first, and will explore fees once a model is preferred. Price a package people do not want and you learn nothing useful.
Fee mechanics, tiers and the fee-increase letter are on membership pricing. The wider review sits on the membership review hub.
A rise your terms already describe
No study. Clear notice and a way out.
Notice onlyA rise above your usual rule
One survey question, then a joiner test.
Light testA tier or concession added
Price range, then a live test on joiners.
Range + liveBenefits rebundled and repriced
Benefit ranking and package choice, then live.
Full studyCategories or the whole offer changed
Proposition first, price second, then phased.
Staged study02/ methods compared
Five methods, five different questions.
There are two ways to learn what people will pay: ask them, or watch them. A peer-reviewed review of willingness-to-pay methods sorts the asking into direct questions and indirect choice tasks, and the watching into market data and experiments.
Each method answers a different question. Van Westendorp gives a range. Gabor and Granger’s price ladder gives demand at set prices. Conjoint shows trade-offs inside a package. Best–worst scaling, sold as MaxDiff, ranks benefits. A live test shows behaviour.
Van Westendorp PSM
AnswersWhere the price feels too cheap or too expensive
AsksFour open price questions about one offer
Good forA first range for a new tier
Weak spotPrice awareness, not intent to buy
Gabor-Granger
AnswersShare who would join at each set price
AsksWould you join at £x? Then higher or lower
Good forChoosing between a few fees
Weak spotOne offer, no competitors, stated not real
Choice-based conjoint
AnswersWhat each benefit and price is worth in a choice
AsksPick one of several packages, repeated
Good forDesigning tiers and bundles
Weak spotCost, design skill, hypothetical choices
→ The usual method for a new package.
MaxDiff (best–worst)
AnswersWhich benefits matter most, in rank order
AsksMost and least valued in small sets
Good forCutting a long benefit list
Weak spotRanks value; does not price it
Live test on new joiners
AnswersWhat people actually do at the new price
AsksNothing: change the offer, watch conversion
Good forConfirming before existing members see it
Weak spotSlow, and needs enough joiners
→ The evidence the others predict.
03/ method chooser
Start from the question, not the method.
Pick what you need to know. The cards that answer it stay.
Most membership studies combine two: a benefit ranking or package choice to shape the offer, and a price question to set the range. Then a live test confirms it. Benefit ideas themselves are on member benefits.
- Direct · 4 questions
Van Westendorp
Where members think the price turns from too cheap to too expensive.
- Direct · price ladder
Gabor-Granger
Would you join at £x? Then higher or lower. Gives a demand curve for one offer.
- Indirect · choice sets
Choice-based conjoint
Packages with varied benefits and prices. What people pick reveals trade-offs.
- Indirect · best and worst
MaxDiff
Most and least valued benefit from small sets. Ranks a long list cleanly.
- Revealed · live
Test on new joiners
Change the join price or offer for a period. Compare conversion, then first renewal.
- Revealed · records
Your own past changes
Renewal by cohort before and after your last fee change. Free, if the data is clean.
Method classes follow Breidert, Hahsler and Reutterer’s peer-reviewed review: direct surveys, indirect surveys, and observed or experimental data.
04/ van westendorp price sensitivity
Van Westendorp: draw your own price curves.
Four questions, four curves, one range. Enter your own survey medians, or paste your respondents’ answers.
Peter van Westendorp presented the price sensitivity meter at the ESOMAR Congress in 1976. Each person names the price at which the offer feels too cheap, a bargain, getting expensive and too expensive. Where the curves cross marks a range of acceptable prices.
He warned that this kind of price consciousness “should never be equated with propensity to buy” (van Westendorp, reprinted by Research World). A later peer-reviewed analysis adds that the range can be wide and shift with small changes in the sample. Treat it as a first range, not a price.
Your inputs
0 lines read · need 5 or more
Price sensitivity meter
£51–£141Range of acceptable prices (PMC to PME)
£65Optimal price point (too cheap = too expensive)
£88Indifference price (cheap = expensive)
Your fee sits inside the range
£110 is between £51 and £141. Test it on behaviour before you rely on it.
Our arithmetic on your inputs; no respondent data is built in. Curves and crossing points follow van Westendorp’s 1976 ESOMAR paper. “Your medians” models each answer as normally spread around your median — a sketch, not a finding. Price consciousness is not propensity to buy. Nothing you type leaves your browser.
The curves and crossing points follow van Westendorp’s paper. No respondent data is built in: every curve comes from your inputs.
05/ stated vs revealed
What people say versus what they pay.
Asking someone what they would pay is the weakest evidence. The Breidert review lists why: it puts price under a spotlight, people have no reason to answer truthfully, and a stated value need not become a purchase. Across the studies it reviews, willingness to pay is higher when nobody has to buy.
Choice tasks do better because they mimic a decision. A peer-reviewed comparison against real purchases found hypothetical conjoint biased, yet it may still point to the right price. Methods with real stakes did best.
Members who would lose something by leaving — a designation, a licence — are less sensitive to dues. Bowman calls these entry and exit costs. Read results by segment for that reason.
How well each method predicts what people pay
- Asking “what would you pay?”No effect
Direct asking focuses on price, gives no incentive to be truthful, and stated value need not become a purchase.
Evidence: direct questionBreidert, Hahsler & Reutterer, 2006 · peer-reviewed review - Interviews against a field testNo effect
In Stout’s 1969 comparison, demand moved with price in the field test; interviews showed no significant change, even reversals.
Evidence: direct questionStout 1969, as reviewed by Breidert et al. · peer-reviewed review - Van Westendorp’s four questionsNo effect
A descriptive “workhorse”; its acceptable range can be wide and move with small changes in the sample.
Evidence: direct questionLipovetsky, 2006 · peer-reviewed - Choice-based conjointHelps
Showed hypothetical bias against real purchases, yet may still lead to the right demand curve and price decision.
Evidence: indirect surveyMiller et al., Journal of Marketing Research, 2011 · peer-reviewed - Conjoint with real stakesHelps
Incentive-aligned methods passed both statistical and decision tests against real purchase data.
Evidence: incentive-alignedMiller et al., 2011 · peer-reviewed - Live tests in the marketHelps
Real behaviour without an artificial setup; the costs are money and the time it takes to see results.
Evidence: real behaviourBreidert et al., 2006 · peer-reviewed review
The grade names the kind of evidence a method produces, not the quality of any one study. Miller et al. tested consumer products, not memberships: read it for the mechanism.
06/ who to sample
Who to ask, and how many.
Current members tell you about renewal. Lapsed members tell you where the price broke. Prospects tell you about joining, which is where a new price is felt first. A study of only loyal members tests the people least likely to leave.
For a conjoint study, Sawtooth Software’s practitioner guidance starts at 300 respondents, with at least 200 in each subgroup you want to compare. Small samples are common: in a peer-reviewed review of 69 choice experiments, 32% used fewer than 100 people, and most did not say how they chose their sample size.
For question design, see membership survey questions and the survey hub.
Current members
- Renewal at the new price
- Split by category and tenure
- Include first-year members
Lapsed members
- Where price broke the habit
- What would bring them back
- Recent lapsers remember best
Prospects
- Joining at the new price
- Event buyers, enquirers, students
- A panel if your list is thin
Employers who pay
- Corporate and employer-paid fees
- Budget cycles and thresholds
- Ask the payer, not only the member
300Respondents to start a conjoint study
200Per subgroup you want to compare
500+Times each level should appear; 1,000 safer
Sawtooth Software, sample size rules of thumb · practitioner guidance from a conjoint software vendor, not a study.
Sample sizes in 69 published choice experiments
- Did not clearly report a sample-size method71%
- Fewer than 100 respondents32%
- More than 600 respondents23%
- More than 1,000 respondents9%
07/ live testing
Confirm it live, on new joiners first.
A survey predicts; a live test shows. Change the join price for new members, keep existing members untouched, and read conversion and first renewal.
A fee change compounds. Model what a tested change does to your own subscription income over five years, at your renewal rate and joiner flow. The renewal slider starts at 82%, the median in Marketing General’s 2026 report (a US survey). Use your own.
Read renewal by cohort, first-year and mature separately; the formula is on membership retention rate.
01 · Baseline
Record the before
Join conversion and first-year renewal for the same months last year.
02 · Joiners
Change the join offer only
New price or package for a set period. Existing members see nothing new.
03 · Compare
Read conversion against baseline
Same season, same channels. One change at a time.
04 · Renewals
Roll out with notice
Announce before the reminder, with the reason and the way out.
05 · Cohort
Watch first renewal
The test is not over until the tested cohort renews.
Your three levers
Subscription income over 5 years
- Now £150k
- Yr 1 £150k
- Yr 2 £150k
- Yr 3 £152k
- Yr 4 £153k
- Yr 5 £156k
+4%income after 5 years
895members in year 5
180joiners a year just to stand still
Income grows
Keep the renewal rate here and the fee rise does the rest.
Our arithmetic on your inputs; 1,000 members at £150 are example starting values. The 82% default is the median renewal rate in Marketing General’s 2026 report, a US survey. Joiners and fee change are yours to set.
The sequence is our method. The model is arithmetic on your inputs; no outcome is implied.
08/ reading results
How far can you trust the result?
Read a result by segment, never as one average. A price most members accept can still lose the early-career members you most need.
Separate what the data shows from what you think it means. The MRS Code of Conduct asks researchers to distinguish facts, opinion and interpretation (rule 60), and to publish only findings the data supports (rule 61). Hold your own board paper to the same standard.
Score your study here before you take it to the board.
Is your price result ready for the board? · 0/5 answered
01What kind of evidence is it?
02Who answered?
03Can you read it by segment?
04Was the package tested before the price?
05Will a live test confirm it?
Answer all 5 to see your result.
Our scoring, built on the method classes in Breidert et al. (2006). Not a statistical test.
09/ ethics and UK GDPR
Research, not a sales pitch in disguise.
Price research with members is still research. The MRS Code of Conduct bars selling “under the guise of research” (rule 4), requires participants’ permission (rule 30) and bars client goods or vouchers as research incentives (rule 25). It binds MRS members and accredited companies; it is a sound standard for anyone.
Under UK GDPR, the ICO’s guidance says there is no specific lawful basis for research; a membership body will usually rely on legitimate interests. Processing counts as statistical only if the results are not used to make decisions about particular people (ICO). Pricing for everyone qualifies; re-pricing one member from their answers does not. The ICO notes this guidance is under review after the Data (Use and Access) Act.
Before the survey goes out · 0/8 in place
Tick what your study already does.
RuleWhat it means for a price test
MRS rule 4No selling or upgrading under cover of research
MRS rule 25No free membership or vouchers as the incentive
MRS rule 43Re-contact only with permission given at the time
MRS rule 49Anonymity kept unless people agree otherwise
ICO, research provisionsLegitimate interests is the likely lawful basis
ICO, statistical purposesResults not used for decisions about one person
MRS Code of Conduct 2023 (sector body) and ICO guidance (regulator). Not legal advice.
10/ real examples
Who has tested, and how.
Three published examples. Two are supplier case studies, so read them as the supplier’s account.
None of these is our work. Insight that shapes the whole offer is covered on insight-informed membership; category design on membership categories.
02 / V&A · supplier case study
MaxDiff, TURF and contingent valuation
Current, recently lapsed and a national panel ranked features and valued them, to repackage and reprice.
example_va →03 / Royal College of Anaesthetists · the body’s own page
Categories first, fees second
Members review three category models; fees follow once a model is preferred.
example_rcoa →Sources: Clusters, Institute of Directors case study — secondary, the supplier’s account; DJS Research, V&A case study — secondary; RCoA, reviewing membership — over 65% of its 2025 survey respondents supported developing and simplifying membership.
11/ questions
Price and proposition testing FAQs.
Price and proposition testing is research that checks a membership offer and its price before launch. Proposition testing asks which package or benefits people want; price testing asks what they would pay for it. It usually combines a survey method, such as conjoint analysis or Van Westendorp, with a live test on new joiners.
A survey method from van Westendorp’s 1976 ESOMAR paper. Each respondent names the price at which an offer feels too cheap, a bargain, getting expensive and too expensive. The crossing points of the cumulative curves mark a range of acceptable prices. Van Westendorp himself warned that price consciousness should not be equated with propensity to buy.
A price ladder: respondents say whether they would buy at a set price, then at a higher or lower one, which gives the share willing to pay at each price. It comes from André Gabor and Clive Granger’s work on consumer price willingness (ESOMAR, 1966). It tests one offer in isolation and relies on stated intent.
When you are designing a package: which benefits to bundle into which tier, and at what price. Respondents choose between packages that vary benefits and prices, and the choices reveal trade-offs. A peer-reviewed comparison found hypothetical conjoint biased against real purchases, yet still capable of pointing to the right price.
MaxDiff, or best–worst scaling, shows people small sets of benefits and asks which matters most and least. Repeated across sets, it ranks a long list on a common scale. It was introduced by Finn and Louviere (1992); Louviere and colleagues set out its advantages over rating scales. It ranks benefits but does not price them.
It depends on the method and how many segments you compare. Sawtooth Software’s practitioner guidance starts a conjoint study at 300 respondents, with at least 200 per subgroup. In a peer-reviewed review of 69 choice experiments, 32% used fewer than 100 people. A live test needs enough joiners in the test period to see a difference.
All three, where you can. Current members tell you about renewal, lapsed members show where the price broke, and prospects show whether people will join at the new price. Employers matter too where they pay the fee. Surveying only loyal members tests the people least likely to leave.
Test on new joiners first. Changing the join price for a set period leaves existing members untouched and shows real behaviour. Roll a rise to renewals only with notice and a way out; the fee-increase letter is on our membership pricing page. Then read first-year and mature renewal by cohort.
Yes. The ICO’s guidance on the research provisions says there is no specific lawful basis for research; legitimate interests is the likely one for a membership body. Results used for statistics must not drive decisions about particular people (ICO). The MRS Code of Conduct also bars selling under the guise of research. Not legal advice.
15 minutes · video or phone
Book 15 minutes on your next price test.
Book 15 minutes to talk through the change you are planning, the method that fits it and who to sample.
- 0115 minutes, video or phone
- 02The method for your question
- 03Who to sample, and how many
- 04A plain next step
Pick a day that suits · live availability
Book 15 minutes · no obligation
Test the offer, then the price, then launch.
The right method for your question, the right people in the sample, and a live test before members see it.
13/ sources
Every claim, and where it came from
Every source was read on 8 October 2026. US, vendor, practitioner and supplier sources are labelled.
- Breidert, Hahsler and Reutterer, Innovative Marketing — A review of methods for measuring willingness-to-pay (2006)Peer-reviewed review
- Peter H. van Westendorp, ESOMAR Congress — NSS Price Sensitivity Meter (PSM) — a new approach to study consumer perception of price (1976)Conference paper, sector body library
- Research World (ESOMAR) — Reprint of van Westendorp’s 1976 paper — the four questions and the caution on propensity to buyOriginal paper, reprinted
- Lipovetsky, International Journal of Operations and Quantitative Management — Van Westendorp price sensitivity in statistical modeling (2006)Peer-reviewed
- Gabor and Granger, Journal of the Royal Statistical Society Series C — On the price consciousness of consumers (1961)Peer-reviewed, UK
- Gabor and Granger, European Marketing Research Review — The pricing of new products (1966)Conference paper, sector body library
- Green and Srinivasan, Journal of Marketing — Conjoint analysis in marketing: new developments with implications for research and practice (1990)Peer-reviewed, US
- Finn and Louviere, Journal of Public Policy & Marketing — Determining the appropriate response to evidence of public concern: the case of food safety (1992) — best–worst scaling introducedPeer-reviewed
- Louviere et al., International Journal of Research in Marketing — An introduction to the application of (case 1) best–worst scaling in marketing research (2013)Peer-reviewed
- Miller, Hofstetter, Krohmer and Zhang, Journal of Marketing Research — How should consumers’ willingness to pay be measured? (2011)Peer-reviewed
- Bowman, Nonprofit and Voluntary Sector Quarterly — Toward a theory of membership association finance (2017) — entry and exit costsPeer-reviewed, US
- de Bekker-Grob et al., The Patient — Sample size requirements for discrete-choice experiments in healthcare (2015)Peer-reviewed
- Sawtooth Software — Sample size rules of thumb for a choice-based conjoint study (2020)Practitioner guidance, vendor, US
- Marketing General Inc. (Tony Rossell) — Just released: the 2026 Membership Marketing Benchmarking Report — 82% median renewalIndustry survey, US
- Market Research Society — MRS Code of Conduct 2023Sector body code, UK
- Information Commissioner’s Office — The research provisions: principles and grounds for processingRegulator guidance (under review)
- Information Commissioner’s Office — The research provisions: what is research-related processing?Regulator guidance (under review)
- Clusters Insights & Analytics — Institute of Directors: price and proposition testingSupplier case study — secondary
- DJS Research — Membership proposition research for the V&ASupplier case study — secondary
- Royal College of Anaesthetists — Reviewing RCoA membershipMembership body’s own page