Gabor-Granger Revenue Curve Interpretation in Practice

A Gabor-Granger revenue curve is a pricing diagnostic. To read it well, teams need to tell demand and revenue apart, check whether the survey tested the right range, and explain the shape before trusting the peak. The highest point offers a hypothesis to test.

What the Gabor-Granger chart actually contains and why teams consistently misread it

An analyst looks at two lines on two axes and reads one against the wrong scale. That is the first mistake, before any argument about the right price begins. Demand belongs on the right axis, measured as a percentage. Revenue is on the left axis and is measured in currency. The lines answer different questions, so their slopes and heights cannot be read interchangeably.

The revenue line creates a second, less obvious trap. It shows average revenue per target customer across the whole sample, including people who would not buy. It does not show revenue per buyer. The calculation produces a hypothetical revenue index based on stated purchase intent. Teams that treat that index as booked revenue will be disappointed when actual sales come in.

Its pedigree does not change what the chart measures. The calculation can be sound while the business conclusion drawn from it is wrong.

Why the revenue curve is shaped like a hump and what each zone of that hump signals

The revenue hump expresses a trade-off between price and the share of people willing to buy. On its left side, a higher price adds more revenue than the lost buyers take away. This is the inelastic zone. On the right side, the loss of buyers outweighs the gain from charging more, so revenue falls. The peak marks the balance point within the tested range.

For a SaaS price ladder, a fall in demand need not be bad news. Demand can decline while revenue rises, provided the higher price compensates for the smaller buying share. Once that compensation stops, the revenue curve turns downward. Reading only the demand line would miss that change.

A wide, flat top leaves room to choose among nearby prices that bring in almost the same revenue, while a narrow top shows that small price changes have a bigger effect. Position also tells a story. A peak near the left edge suggests that the grid started too high, while a peak near the right edge suggests that it stopped too low. Either can show that the survey needs fixing before it can guide a price choice.

The four diagnostic signals the chart gives before the optimal price deserves any trust

The chart’s four signals form a pre-flight check. “Not Interested” measures the share who opted out before seeing any price, so it concerns audience fit. A high share suggests the sample included people who were not potential buyers. Their lack of interest cannot safely be interpreted as resistance to the prices shown.

“All Rejected” and “No Ceiling” test the boundaries of the price grid. The first records people who rejected every price, including the lowest, and warns that the starting price may be too high. The second records people who accepted every price, including the highest. Here, the true peak may lie above the tested range, so the chart’s highlighted price may be only a lower bound for the optimum.

“Optimal Price” belongs at the end of this inspection. It marks the revenue peak, but even a clear peak cannot fix a grid that stopped short of the range it needed to test. Before presenting it, the analyst should establish that audience fit and both boundary checks support an interpretation. Diagnostics come before findings, even when the software makes the winning price the biggest number on screen.

How a flat peak, a steep cliff, and a premature drop each call for a different response

A flat peak turns a forced choice into a strategic one. A lower price can bring in more volume, while a higher price can bring in more margin from fewer customers who are easier to serve. Neither choice is automatically wrong. The response should follow cost structure and growth objectives, rather than defaulting to the midpoint simply because it looks balanced.

A cliff needs a different response because its cause is unresolved. A sharp drop between adjacent tested prices could mark a psychological threshold, but it could also reflect prices spaced too far apart or the way the offer was framed. The chart cannot settle that distinction by itself. Qualitative follow-up should explain the drop before a recommendation lands on either side of it.

A steep decline from the first tested price instead points toward a grid that began too high. The response is to rerun with lower prices. Before interpreting any of these patterns, the analyst should also split the curve by segment. A cliff in the blended curve may vanish or grow sharper within a group, while a gentle slope may show that respondents share a moderate sensitivity to price. The total curve is not necessarily a shared market response.

The revenue peak and the profit peak are different prices, and confusing them is a costly default

Revenue leaves out the cost of serving the buyer. The contribution curve brings that cost back into the calculation by multiplying demand share by price minus unit cost. Its peak is usually at a higher price than the revenue peak because the margin kept counts for more than the volume kept. When unit costs are meaningful, optimizing revenue alone can leave margin behind.

That third line is often missing from the chart. Most tools show revenue by default, which makes it easy to treat the most visible maximum as the business objective. For SaaS and subscriptions with marginal costs near zero, the two curves converge. For physical goods, services with real delivery costs, or tiers with different costs, contribution is the operative measure.

ArcWave illustrates the distinction without needing a complicated argument about buyer psychology. Its pricing analysis produced a revenue peak at one price and a contribution peak at a higher price after unit cost was included. The team chose the contribution peak. That decision changed because the calculation included costs.

What stated purchase intent actually predicts — and how far to discount it

Even a well-designed grid measures what people say they would buy. It does not measure what they actually purchase. Hypothetical bias arises because spending imaginary money is easier than handing over real money. Stated intent therefore tends to overstate real purchase behavior.

There is a counterargument, but not a clean escape. Some respondents understate willingness to pay in an effort to keep prices down, which can offset some overstatement. Conjointly argues that these biases mostly cancel out, but that cancellation is imperfect. It does not make the survey equivalent to a transaction.

The practical response is an explicit discount to stated intent. How large that adjustment should be is a judgment call, and the curve itself cannot supply the answer. That is a major source of uncertainty in the recommendation. Even a clear cliff or tidy peak comes from hypothetical choices, so neither is as precise as a result from actual sales. The curve can still guide the choice if the recommendation makes that limit clear.

Where the curve goes silent: competition, features, and the reasoning behind the bend

Gabor-Granger isolates price, which leaves several business questions unanswered. It cannot separate the value of individual features or decide which features belong in a bundle. When the decision depends on those trade-offs, conjoint analysis is the appropriate method.

Competition creates another gap in the interpretation. A bend could reflect a competing product at that price, a category norm, a psychological threshold, or a perceived quality floor. Those explanations are not interchangeable. The curve shows where willingness to buy changes, but it does not identify which comparison or belief produced the change.

The practical fix is to pair the price ladder with qualitative follow-up. The quantitative work locates the peak and the cliff, while the conversation investigates the reasoning behind them. AI-moderated interviews can bring that follow-up into the same exchange. To make a recommendation defensible, ask why respondents accepted or rejected the offer at the point where their answer changed.

How the ArcWave launch shows interpretation dependence in a real pricing decision

ArcWave’s launch shows how one curve can support opposing readings. Read as genuine consumer interest, the results supported a premium launch. Read differently, they could have supported anchoring to a lower bundle price. The peak alone could not decide between them.

Segment behavior supplied the missing context for that choice. The brand-aware and traveler segments were willing to pay more, with a revenue peak above the blended market peak. Those differences, together with brand signals, made the premium interpretation more defensible.

The team also used the analysis to build a price architecture comprising a DTC price, an MSRP, and a holiday promotional price, with a floor intended to avoid value erosion and an upper bound. This was a broader decision than choosing the top of a line. The case demonstrates interpretation dependence, because costs, segment behavior, and qualitative context determined how the numerical result could be used. Without that context, the chart would have left the central launch question unresolved.

Running Gabor-Granger faster and with richer qualitative output using modern research infrastructure

The classic workflow separates a static survey from the conversation needed to explain it. Manual branching handles the price ladder, and qualitative work comes later. That separation becomes a problem when a cliff appears and a recommendation is already due. Modern research infrastructure can shorten the cycle from weeks to hours while bringing the follow-up into the ladder itself.

Conveo lets teams ask why at each price point in the same conversation. A yes-or-no response supplies the curve, while the explanation helps distinguish a threshold from a framing effect or another interpretation. Speed is useful because it makes that explanation available during the pricing work.

Synthetic AI consumer panels give teams a way to try out the grid first. Teams can test different price ranges, segments, and product framings with a modeled audience before they turn to live respondents. The recommended hybrid workflow uses those rapid runs to explore boundaries and segment hypotheses, then validates the promising range with a smaller, focused live panel study.

A reading protocol: the sequence in which to interpret a Gabor-Granger curve before recommending a price

For the next chart, use the following sequence before recommending a price. First, check audience fit and confirm that “All Rejected” and “No Ceiling” are within acceptable bounds. If either boundary check fails, repair the grid before interpreting its maximum. Second, match each line to its axis, with demand on the percentage scale and revenue on the currency scale. Do not read slope until that distinction is clear.

Third, look at the hump’s width and position as well as its highest point. Record whether it offers room to choose, signals fragility, or runs into the edge of the tested range. Fourth, flag every cliff and use qualitative follow-up to distinguish a threshold from a grid artifact or framing effect. Explain the drop before choosing either side. Fifth, plot contribution when unit costs are meaningful and use that curve for the cost-sensitive decision.

Sixth, state the discount applied to purchase intent and make clear that it is a judgment. Seventh, look at segments before drawing conclusions, so the blended average does not hide responses that conflict. Revisit any pattern whose meaning changes within a group. Eighth, treat the resulting peak as a hypothesis and validate it with an A/B test or a geo test before full commitment.