Pricing Page and Packaging Work Backed by Willingness Research

The pricing page is usually not the problem. Packaging is. Before a line of the page changes we establish what people will pay and for which bundle, using research with a sample that can carry the claim, 250 or more respondents per segment drawn from your customers, your churned accounts and your lost deals rather than from a panel, and only then rebuild the page around the answer. A beautifully argued page for the wrong three tiers is an expensive way to keep the same revenue. Method, and its limits: Van Westendorp to bound the acceptable range, which is a blunt instrument and treated as one. Gabor-Granger for elasticity across a short list of candidate points. Choice-based conjoint when the real question is which capability belongs in which tier, which it usually is. Each method has a characteristic failure: Van Westendorp flatters the midpoint, stated willingness overstates by a factor that varies by category, conjoint results collapse if the attribute list was written by whoever most wanted a particular answer. We name which failure applies to your result instead of handing over a chart and letting it stand unqualified. Why we do not run price A/B tests: Showing different prices to comparable consumers in the EU at the same moment is an exposure your legal counsel will not enjoy explaining, and the reputational version of that conversation happens on social media rather than in court. Beyond that: returning visitors contaminate the cells, and revenue per visitor has enough variance that the test needs several times the traffic of a conversion test on the same page. We will happily test the page, plan naming, feature table structure, toggle framing, where the contact-sales threshold sits. We will not test the number. The page itself: Whether the value metric is legible within five seconds, per seat, per workspace, per thousand calls, because a buyer who cannot predict next year's bill will not sign this year's. Tier names that describe who the tier is for rather than how good it is. Feature tables that stop at what actually differentiates, since the eighteen shared rows are cognitive load with no decision value. Annual and monthly framing, including whether the discount is stated as a percentage or as months free. And for consumer sales in the EU, displayed prices include VAT, which the Price Indication Directive requires and which we have found wrong on live sites more than once. Migration, which is most of the risk: Any packaging change lands on an existing base. You get a grandfathering plan, a segment-by-segment view of who is worse off, and the notification copy, because the churn caused by a badly announced increase costs more than the increase earns and does so within one quarter. Not included: No billing implementation in Stripe, Chargebee or anywhere else. No discounting or promotional calendar. No sales enablement, battlecards or objection handling. No ongoing competitor price monitoring, which is a subscription you can buy for less than we would charge to read it to you. Who should not buy this: Teams looking for charm pricing and a decoy tier. Ending in nine and adding a tier nobody buys is folklore, not research, and it is priced accordingly elsewhere. Teams who cannot reach 250 respondents per segment and are unwilling to have the finding stated as directional. Companies whose price has been set by the founder's instinct and where that instinct is not genuinely open to being wrong.