No box in the template asks what you abandon. The method that does measure a trade-off breaks down above six attributes, where most positioning starts.
A value proposition template cannot be wrong, because none of its boxes asks what you are giving up.
Look at the standard shape. Who you serve. What you do. What benefit they get. Why you are different. Every field is additive. You can complete all of them, present the result, get nods around the table, and have decided nothing at all, because nothing in the exercise required anyone to abandon a customer, a use case, a channel or a claim.
The definitive statement on this is thirty years old and unusually blunt:
“Trade-offs occur when activities are incompatible. Simply put, a trade-off means that more of one thing necessitates less of another.”
And, from the same article:
“The essence of strategy is choosing what not to do. Without trade-offs, there would be no need for choice and thus no need for strategy. Any good idea could and would be quickly imitated.”
Followed by the sentence most templates quietly deny:
“the absence of trade-offs is a dangerous half-truth that managers must unlearn. Quality is not always free.”
The article’s own example is the useful one. A soap brand built its position on a residue-free formula, pharmacy distribution, detailing to dermatologists, advertising in medical journals, and a slower, costlier manufacturing process. In choosing that, it “said no to the deodorants and skin softeners that many customers desire in their soap” and “gave up the large-volume potential of selling through supermarkets.”
Two abandonments, stated. That is what a value proposition looks like when it has been decided rather than filled in.
The method that measures a trade-off, and its ceiling
There is a real technique for measuring what people will give up, and it dates from 1971. Instead of asking how important each benefit is, which invites everyone to say all of them, conjoint analysis presents complete profiles and asks the respondent to rank or choose among them. Importance is then inferred from choices that forced a sacrifice.
That reversal is the whole point, and it explains why the question “how important is fast delivery to you” produces nothing usable. Asked alone, every attribute is important. Asked against a price and a lead time, one of them yields.
Now the constraint that matters most for a value proposition. The definitive review of the method, published in 1990 by two of its architects, recommends:
“we recommend the use of full-profile conjoint analysis if the number of attributes can be kept down to (say) six or fewer factors.”
Their reason, on the same page: beyond that there is “severe information overload on the respondents. When faced with such tasks, respondents resort to simplifying tactics and the resulting part-worth estimates may distort their true preference structures.”
Read that against the average value proposition slide, which lists eight to twelve benefits. The reference method for measuring trade-offs breaks down at the exact point where most positioning statements begin. A proposition with ten pillars is not merely unfocused. It is, in a measurable sense, unevaluable.
Two further facts keep this honest. Reliability across more than thirty studies from 1973 to 1984 has a median around 0.75, which is respectable and not perfect. And the measured importance of an attribute depends partly on the analyst: holding the range of variation constant, adding intermediate levels to an attribute increases its estimated importance. A measured trade-off is not a raw fact about your customers. It is the product of an experimental design, and the design has to be declared.
What people say they would pay, against what they pay
The other half of a value proposition is price, and here the measurement problem is quantified across hundreds of studies.
The tightest estimate comes from a meta-analysis restricted to studies that reported a monetary willingness to pay and used the same elicitation mechanism for the hypothetical and the real value. Twenty-eight studies, 83 observations. The calibration factor, meaning stated value divided by real value, has a mean of 2.60 and a median of 1.35.
A broader meta-analysis of 29 studies, including both willingness to pay and willingness to accept, concluded that “on average subjects overstate their preferences by a factor of about 3 in hypothetical settings”.
And for consumer goods specifically, a 2020 meta-analysis covering 77 studies, 115 effect sizes, 24,347 hypothetical observations and 20,656 real ones found a mean hypothetical bias of 21 percent, with one counterintuitive result: indirect methods overstate real willingness to pay more than direct methods do. Conjoint analysis does not mechanically protect you.
The spread between 1.35 and 3 is not a disagreement. It is the lesson. The tighter the protocol, the smaller the gap, and the gap never reverses direction. People overstate, always in the same direction, by an amount that depends on how carefully you asked.
One more finding, less quoted and useful: the ratio of standard deviations between real and hypothetical responses is about 0.6. Hypothetical answers are not only higher. They are more scattered.
The benchmark against actual purchase data is the last piece. A 2011 study compared four approaches against real purchases and found that only two passed both the statistical and decision tests: an incentive-compatible mechanism, and an incentive-aligned conjoint. Both make the respondent face a real consequence. Neither is what a positioning workshop does.
Two popular methods that measure no trade-off at all
The four-question price method. It appears in every product marketing toolkit: at what price would this be too expensive, too cheap, expensive but acceptable, so cheap you would doubt the quality. Its origin is a 1976 paper delivered at a trade congress. There is no founding article in a peer-reviewed journal, and no published derivation of why the intersections of its cumulative curves should mean anything. The four questions and their crossing points are conventions set by their author.
The one peer-reviewed evaluation of it is balanced and worth quoting in full:
“Results indicate that the Van Westendorp Price Sensitivity Meter yields biased results because of its hypothetical nature and its focus on minimum customer resistance. However, the authors find it to be a method of high predictive quality for eliciting willingness-to-pay since the measurement results are comparable to those of the incentive-aligned Becker-DeGroot-Marschak mechanism.”
Note the second criticism, because it is exactly the failure mode of an additive value proposition. The method optimizes for minimum customer resistance: it finds the price that annoys the fewest people, not the price that captures the most value from the people you actually want. And all four of its questions concern the price of one object. None of them asks you to choose between two configurations, so it measures no trade-off whatsoever.
Agreement scales. The other way a proposition gets falsely validated is by asking twenty prospects to rate each benefit on a scale from strongly disagree to strongly agree. Everything comes back important, and the exercise is declared confirmed.
The methodological literature explains why, with numbers:
Across 41 studies, the correlation between answers to a statement and answers to its reverse averages negative 0.22. It should approach negative 1.
Across 10 studies, 52 percent agree with a statement while only 42 percent disagree with its opposite.
Across 8 studies, 14 percent more people agree with a statement than express the same opinion in a forced-choice version of the question.
Across 7 studies, 22 percent of respondents agree with both a statement and its contrary, while only 10 percent disagree with both.
The summary from the same source: “taken together, these methods suggest an acquiescence effect averaging about 10%.”
And acquiescence is a stable trait, not noise: its cross-sectional reliability averages 0.65 across 29 studies. Which means a benefit-rating exercise measures the respondent’s disposition to agree at least as much as the value of the benefit. The recommended alternative is stated plainly: “a better practice is to offer respondents a choice between alternative statements.”
That is the whole argument in one sentence. Force a choice, or you are measuring agreeableness.
Before anyone reaches for a framework, it is worth knowing the state of the field.
A 2017 review in the Journal of the Academy of Marketing Science, the most authoritative treatment of the customer value proposition available, concludes that the concept remains poorly understood and poorly implemented; that relatively little research has been published on it compared with the work on value generally; and that the literature is fragmented and lacks a solid theoretical foundation. The authors are reduced to proposing a conceptual model with research propositions and illustrating it with four case studies.
Propositions to be tested. Not results.
So there is no published evidence that a better-formulated value proposition causes a better commercial outcome. Anyone claiming otherwise is extrapolating from a framework, not citing a finding.
Which does not make the exercise worthless. It makes the justification different. The reason to state what you give up is not that a study proved it lifts revenue. It is that a position nobody can be excluded from is one nobody can be chosen for, and that the trade-offs you refuse to make in a document get made anyway, later, by whoever answers the next request for proposal.
On that last point the evidence is genuinely split, and both sides deserve naming. A 2000 meta-analysis of the generic strategy literature found that cost and differentiation work well as high-level descriptors but that the theoretical claim about their performance “is not yet supported” by the empirical record. A 2007 study of 2,351 business units using national survey data found a significant positive relationship between strategic purity and performance, with pure strategies never doing worse and often better than hybrid ones. Write that both ways or do not write it.
One sentence per abandonment. Not “we focus on mid-size manufacturers”, which excludes nobody in practice, but “we do not serve companies below fifty employees, we do not sell through resellers, and we do not compete on lead time.”
Six claims at most. Above that, the reference method for measuring what customers trade off stops working. Treat it as a hard ceiling on how many things a position can be about.
Never validate with an agreement scale. Present two configurations that differ, and ask which one they would buy. Everything else measures agreeableness, at about ten points of it.
Discount every stated price. The direction is always the same and the magnitude depends on how carefully you asked. A number obtained without consequence is an upper bound, not an estimate.
Write down what would prove you wrong. If the abandonment costs you nothing, you did not abandon anything. If losing the deals you said you did not want does not hurt, they were not deals you were winning.
Because it is additive. Every box asks what you offer and none asks what you give up, so it can be completed without any decision having been made. The founding statement on strategy says the opposite: a trade-off means more of one thing necessitates less of another.
How do I actually measure what customers will trade off?
Conjoint analysis, where respondents rank complete profiles and the analyst infers importance from choices that force a sacrifice. But it has a ceiling: its founding review recommends six attributes or fewer, warning that beyond that respondents resort to simplifying tactics that distort their preferences.
Can I trust what people say they would pay?
Not at face value. The tightest meta-analysis, restricted to studies using the same elicitation on both sides, found a median calibration factor of 1.35 and a mean of 2.6. A broader one found people overstate by about a factor of three. A consumer-goods meta-analysis found a mean bias of 21 percent.
Is the four-question price sensitivity method reliable?
It has no published peer-reviewed theoretical derivation. The one peer-reviewed evaluation found it yields biased results because it is hypothetical and because it targets minimum customer resistance, meaning the price that displeases the fewest people rather than the price that captures the most value.
Is there evidence that a better value proposition improves results?
No. A 2017 review in the Journal of the Academy of Marketing Science concludes the concept remains poorly understood and poorly implemented, that relatively little research exists, and that the literature is fragmented and lacks a solid theoretical foundation. What exists is a research agenda, not results.