This is a new topic for this series, though it touches ground our product strategy frameworks guide covers from a different angle — that guide covers prioritization discipline generally; this one applies real, specific pricing research to a decision founders often make with no research at all. Our MVP development guide covers what a first build costs and how long it takes; this guide picks up right where that one ends, covering what to actually charge for it once it exists. Pricing is one of the few product decisions a founder can change relatively cheaply after launch, which is exactly why it gets treated as an afterthought — but a real, decades-old research methodology, and a real body of peer-reviewed behavioral research, already exist to replace guesswork with something closer to evidence.
The Real Pricing Model Taxonomy
What are the actual distinct pricing models used by real SaaS companies?
Five real, distinct models recur across the industry: per-seat/per-user (charging by the number of people using the product), usage-based/consumption (charging by actual usage volume), flat-rate (a single price regardless of usage or seats), tiered (a small number of fixed packages at different price/feature levels, often including a free tier), and hybrid models that combine two or more of these — most commonly a seat-based or tiered base fee plus a usage-based component layered on top.
| Model | How It Actually Charges | Where It Fits Best |
|---|---|---|
| Per-seat / per-user | A fixed price per user with access to the product | Products where value scales cleanly with headcount using the tool, like collaboration software |
| Usage-based / consumption | Charges scale directly with actual usage volume | Infrastructure and API products where usage is the real proxy for value delivered |
| Flat-rate | A single fixed price regardless of usage or seat count | Simple products with a narrow feature set and low variance in how customers use them |
| Tiered (often with a free tier) | A small number of fixed packages at increasing price/feature levels | Products serving a wide range of customer sizes and needs from one pricing page |
| Hybrid | A base fee (seat- or tier-based) plus a usage-based component on top | Products where a fixed core value exists but usage still varies meaningfully by customer |
Usage-based pricing specifically has been the subject of extensive, real, ongoing tracking by OpenView Partners, a named venture capital firm that has published an annual SaaS Benchmarks report since well before usage-based pricing became a mainstream conversation. OpenView's own reporting has documented a real, substantial shift toward usage-based elements across the SaaS industry over the past several years — the exact percentage in any given year is worth pulling directly from OpenView's current report at openviewpartners.com's SaaS Benchmarks series rather than treated as a fixed number here, since it is measured and republished annually and this guide would otherwise go stale. Kyle Poyar, a named operator who previously worked at OpenView and now writes the widely followed Growth Unhinged newsletter, is a useful ongoing, named source for tracking how these model choices continue to evolve in practice, since much of his writing focuses specifically on real, current pricing decisions founders are making.
Enterprise and custom pricing: the sixth, unlisted model
It's worth naming a real, common pattern the five-model taxonomy above doesn't formally capture as a distinct model, but that nearly every SaaS pricing page eventually adds: a “Contact Us” or “Custom” tier sitting above the published, self-serve tiers, aimed at enterprise buyers whose needs — security review requirements, custom contract terms, volume discounts, dedicated support — don't fit neatly into a fixed, published price. This isn't really a sixth pricing model so much as an acknowledgment that the five real models above describe how self-serve pricing works, while enterprise sales frequently involves negotiated terms that deliberately sit outside the published structure. The practical decision a founder faces isn't whether to eventually add this tier — most products serving any mix of company sizes eventually do — but when: adding an unpriced enterprise tier too early, before a product has the self-serve tiers below it validated with real pricing research, risks a sales team improvising enterprise pricing without the same evidentiary grounding the rest of this guide argues for. Once real enterprise deals start closing, the specific negotiated terms across those deals become their own real, valuable data source — not for building a Van Westendorp-style survey, but for spotting recurring, specific requests (a particular integration, a specific compliance certification, a particular contract term) that keep surfacing deal after deal, which is often the clearest signal that a genuinely new, formalized tier should be carved out of what has so far been ad hoc negotiation.
Packaging: How Many Tiers, and What Goes in Each
How many pricing tiers should a new SaaS product actually have?
There is no single correct number, but the practical pattern across most SaaS pricing pages converges on two to four visible tiers plus an unpriced enterprise option — enough to serve genuinely different customer segments without forcing a prospective buyer to parse an overwhelming number of near-identical choices. The specific number matters less than whether each tier is built around a real, distinct customer need rather than an arbitrary feature split invented to create the appearance of choice.
The practical failure mode worth naming directly is a tier structure that reflects how the engineering team happened to build feature flags, rather than how customers actually think about their own needs. A pricing page with five tiers whose differences are hard for a prospective buyer to parse at a glance — three tiers all clustered within a narrow price band, distinguished by features most buyers don't understand well enough to know if they need — adds decision friction without adding real clarity. The more durable approach, consistent with the persona-based methodology Patrick Campbell documented at Price Intelligently, is to build tiers around genuinely distinct customer segments identified through the same kind of research covered earlier in this guide — a self-serve individual or small team, a mid-sized team with a specific set of collaboration or integration needs, and a larger organization needing security and administrative controls — rather than starting from an arbitrary feature list and working backward into tiers.
A closely related, practical question worth naming directly: which specific feature to use as the “wall” that separates a free or lower tier from a paid one. The honest answer depends entirely on what genuinely constitutes core value versus what constitutes expansion value for a specific product, which is exactly the kind of product-specific judgment the pricing research covered throughout this guide is meant to inform, not replace. A wall placed too early — gating a feature a user needs just to experience the product's core value at all — suppresses adoption before a prospective customer has any reason to believe the product is worth paying for. A wall placed too late, gating only features a small minority of power users ever reach, leaves revenue on the table from customers who would have happily paid earlier. Getting this right is less about a universal rule and more about genuinely understanding, from real usage data once the product has any, where a typical customer's actual value realization happens relative to where the paywall currently sits.
The Van Westendorp Price Sensitivity Meter
What is the Van Westendorp Price Sensitivity Meter, and does it actually work?
It is a real, named pricing research methodology developed by Dutch economist Peter van Westendorp in 1976, which asks potential customers four specific questions about a product to map out an acceptable price range — not a single number, but a band bounded by where a price feels too cheap to trust and where it becomes too expensive to consider at all.
The four real, specific questions the methodology asks are: at what price would this product be so cheap you'd question its quality; at what price would it be a bargain, a good deal for the money; at what price would it start to feel expensive, though you'd still consider it; and at what price would it become too expensive to consider at all. Plotting the responses to these four questions across a sample of potential customers produces intersection points that define a realistic, defensible price range — rather than a founder's single guess at a number, or a price copied directly from whatever a competitor happens to charge. The methodology is old enough, and well-documented enough across market-research literature, that it doesn't require special tooling to run — a founder can administer these four questions in a simple survey to a sample of real prospective customers before finalizing a price.
What the methodology can't tell a founder
It's worth being direct about the real limitations of a nearly 50-year-old survey methodology, precisely because overstating what it can do would undercut the case for using it at all. Van Westendorp asks people to imagine paying for a product, which is a meaningfully different mental exercise than actually reaching for a credit card — stated willingness to pay in a hypothetical survey question reliably differs from real, observed purchasing behavior, a gap well known throughout survey-based market research generally, not a flaw specific to this one method. The technique also asks about a single product in isolation, which means it doesn't naturally surface how a price will actually be judged against specific, named competitors a buyer is comparing against in the moment of decision — a real purchase decision, unlike a survey response, usually happens with several competing options visible at once. None of this means the exercise isn't worth running; it means its output is a well-reasoned starting range to test against real purchasing behavior after launch, not a guaranteed, final answer that replaces watching what actually happens once real money is on the table.
Other Real Pricing Research Methods
Is Van Westendorp the only real, credible way to research pricing, or are there other established methods?
No — conjoint analysis is a second, well-established market-research technique, applied to marketing since the 1970s, that asks respondents to choose between different bundles of features and prices to infer which specific attributes actually drive willingness to pay. Patrick Campbell, founder of Price Intelligently (later ProfitWell, now part of Paddle), built a widely cited, named methodology combining buyer-persona segmentation with willingness-to-pay surveys specifically for SaaS pricing.
Conjoint analysis works differently from Van Westendorp in a way worth understanding directly: rather than asking about price in isolation, it presents respondents with several different hypothetical product configurations — each bundling a different combination of features at a different price — and asks them to choose between them repeatedly. Because the specific combinations are varied systematically across the survey, the technique can statistically isolate which individual features are actually driving a customer's willingness to pay more, rather than only establishing an acceptable price range for a single, fixed product as Van Westendorp does. It requires more survey infrastructure and a larger sample to run well, which is part of why Van Westendorp remains the more commonly reached-for method for an early-stage product with a smaller base of prospective customers to survey.
Patrick Campbell's methodology at Price Intelligently, documented across multiple interviews he has given (including SaaS Club's and Acquired.fm's podcasts), follows a real, repeatable sequence: define distinct buyer personas, survey each persona specifically on feature preferences and willingness-to-pay using Van-Westendorp-style economic questions, align pricing tiers to what the data actually shows about each persona's preferences, and revisit the whole exercise on a recurring basis — Campbell has spoken specifically about quarterly repricing reviews rather than treating an initial price as permanent. The company he founded was acquired by Paddle, a real, named billing and payments company, folding this research practice into a broader subscription-commerce platform.
The Psychology of Price Endings
Does ending a price in .99 or 9 actually change how much people buy, according to real research?
Yes — this is real, peer-reviewed, and somewhat counterintuitive. A field experiment by Eric T. Anderson and Duncan I. Simester, published in Quantitative Marketing and Economics in 2003, found that an identical mail-order clothing item sold better priced at $39 than at the lower price of $34 — a real, controlled demonstration that a 9-ending price can outsell a genuinely lower, round-numbered price for the same item.
Anderson and Simester's study is worth citing precisely because it is a real field experiment with an actual randomized price manipulation, not a survey asking people how they think they'd react to different prices — the same physical item, sold to real customers through a real mail-order catalog, at genuinely different randomly assigned prices (Anderson & Simester, “Effects of $9 Price Endings on Retail Sales,” Quantitative Marketing and Economics, 2003). A separate, earlier study by marketing professor Robert M. Schindler and co-author Thomas M. Kibarian, published in 1996, found a real, measurable sales increase from prices ending in “.99” compared to otherwise identical prices ending in “.00.” Both studies point toward the same underlying, real phenomenon: price endings function as a genuine behavioral signal to buyers, not merely an aesthetic choice, even though the effect runs counter to a naive assumption that lower is always simply better. It is worth being direct that this guide has deliberately excluded several specific aggregate percentage figures (claims of a fixed “24% average lift” from 9-ending prices) that circulate widely online without a traceable primary source — the two named, peer-reviewed studies above are the specific, citable findings this guide relies on instead.
Anchoring Effects
What is anchoring, and is there real research showing it affects what people are willing to pay?
Anchoring is the well-documented cognitive bias where an initial number a person is exposed to — even one they know is arbitrary — measurably shifts their subsequent judgments, including how much they're willing to pay. Dan Ariely, George Loewenstein, and Drazen Prelec's 2003 paper in the Quarterly Journal of Economics demonstrated this directly with a real, now-famous experiment: MIT students who wrote down the last two digits of their own Social Security number before bidding on products bid systematically higher when those digits happened to be higher, despite the number having no logical connection to the products' value.
The paper's real title, “Coherent Arbitrariness: Duration-Sensitive Pricing of Hedonic Stimuli Around an Arbitrary Anchor,” captures its central finding precisely: once an arbitrary anchor is introduced, subsequent judgments about relative value remain internally consistent (a person who bids higher on one item, having anchored high, will bid consistently higher on similar items too) even though the anchor itself was random and disconnected from the product's actual value (Ariely, Loewenstein & Prelec, “Coherent Arbitrariness,” Quarterly Journal of Economics, 2003). The direct, practical implication for pricing a new product: whatever number a prospective customer encounters first — a competitor's price, an early, informal quote, a number mentioned in an exploratory sales conversation — functions as a real anchor for everything that follows, which means the order in which pricing information is presented, and the very first number a prospect hears, is not a neutral or inconsequential detail. A founder who leads a pricing conversation with a high-end tier before mentioning a lower one, for instance, is deliberately using this same, real, well-documented effect — not a manipulative trick invented by SaaS companies, but the same cognitive mechanism Ariely, Loewenstein, and Prelec demonstrated with a completely unrelated random number.
Price as a Quality Signal
Does a higher price actually make a product seem better to buyers, according to real research?
Yes — this is a well-established, peer-reviewed finding, not just an intuitive marketing claim. Akshay R. Rao and Kent B. Monroe's 1989 integrative review in the Journal of Marketing Research, synthesizing a substantial body of prior experimental research, established a statistically significant positive relationship between price and buyers' perceived quality of a product.
Rao and Monroe's paper, “The Effect of Price, Brand Name, and Store Name on Buyers' Perceptions of Product Quality: An Integrative Review,” is a real, peer-reviewed meta-analysis published in the Journal of Marketing Research in 1989, meaning its finding isn't drawn from a single study but from a synthesis of many prior experiments examining the same question (Rao & Monroe, Journal of Marketing Research, 1989). The practical relevance for a founder pricing a first product is direct and worth stating plainly: a price set too low, out of a fear of scaring away early customers, doesn't just leave revenue on the table — per this research, it can genuinely signal lower quality to the very customers a founder is trying to win over, working against adoption rather than helping it. This doesn't argue for pricing arbitrarily high; it argues for treating a very low, defensive starting price as carrying a real, measurable cost of its own, not a safely neutral choice.
A Real, Documented Pricing Model Change
Is there a real, documented example of a company changing its pricing model with reported results?
New Relic, a real, publicly known infrastructure-monitoring company, shifted from a bundled, infrastructure/SKU-based pricing structure to a hybrid model combining a per-user component with usage-based telemetry pricing, a change documented from the outside by Zuora's own “Subscribed” blog as part of its broader coverage of usage-based pricing adoption.
It's worth being precise about the sourcing here, since this guide holds itself to a high bar on exactly this kind of claim: the account of New Relic's pricing shift, including a reported roughly 15% increase in committed annual recurring revenue following the change, comes from Zuora's own published account of New Relic's transition, not from a primary New Relic press release or investor filing independently verified in this research. Zuora is a real, named subscription-billing company with a direct commercial interest in usage-based pricing adoption, which doesn't make its account false, but does mean it should be read as a vendor's account of a customer's outcome rather than an independently audited result. Anyone citing the specific 15% figure in a high-stakes context should confirm it directly against New Relic's own investor materials or press communications before relying on it as an audited number.
What is safe to take from this case regardless of the exact percentage is the shape of the underlying decision: New Relic moved from a pricing structure that didn't track how customers actually used the product (bundled infrastructure SKUs) toward one that did (usage-based telemetry combined with a per-user base), a real, well-documented pattern across the broader usage-based pricing shift OpenView has tracked industry-wide. The lesson worth taking isn't “usage-based pricing produces a specific 15% lift” — it's that a pricing model misaligned with how value is actually delivered and consumed is a real, fixable problem, and companies large enough to have the internal data to notice this misalignment have, in real, documented cases, chosen to fix it rather than treat their initial pricing choice as permanent.
The practical mechanics of changing a pricing model without breaking existing customers
A real pricing model change, like New Relic's, is a genuinely harder execution problem than choosing the new model in the first place, because it has to account for a real, existing customer base already paying under the old structure. This connects directly to the same discipline our guide on sunsetting or migrating a legacy product covers for a different kind of transition: a pricing change that forces every existing customer onto new terms immediately, with no transition period, risks exactly the kind of customer backlash that guide documents in the Google Reader and G Suite Legacy cases. The more common, lower-risk approach real companies use is grandfathering — letting existing customers keep their current price and terms for some period, or indefinitely, while the new pricing structure applies only to new customers going forward. This doesn't mean a pricing change never reaches existing customers at all; it means the transition gets sequenced deliberately, with clear communication about what changes, when, and why, rather than silently changing what a customer is billed without adequate notice.
Annual vs. Monthly Billing and Discounting
Does offering an annual billing discount actually matter, and how large should it typically be?
Annual billing discounted against the monthly-equivalent price is standard, near-universal practice across SaaS pricing pages, for a straightforward, real underlying reason: it trades a discount for prepaid, more predictable revenue and meaningfully lower payment-churn risk, since a customer who has already paid for a year cannot cancel mid-cycle the way a monthly subscriber can. The specific discount percentage varies by company and isn't governed by a single, authoritative recommended figure this guide could verify — it is a real, company-specific trade-off between the value of prepaid cash and reduced churn risk against the discount required to make prepayment attractive enough for a customer to choose it.
The underlying logic is worth making explicit, since it connects directly back to the anchoring research covered earlier in this guide: presenting a monthly price first and an annual-equivalent discounted price second (or vice versa) shapes how large the discount feels, exactly the same anchoring mechanism Ariely, Loewenstein, and Prelec's research describes in an entirely different context. A founder deciding how to structure and present this choice is making a real pricing-psychology decision, not simply picking a discount percentage in isolation. It is also worth being honest about a real trade-off on the business side that discounting for annual prepayment doesn't eliminate: prepaid annual revenue looks attractive on a cash-flow basis in the short term, but if the underlying product doesn't retain that customer past the prepaid period, the discount has simply moved a churn event further into the future rather than preventing it — a distinction worth tracking via renewal rate specifically, not just initial annual-plan signup rate.
Freemium Conversion: What the Data Shows
What percentage of freemium users actually convert to paid, according to real, credible data?
A real, named, dated report — the 2026 Free-to-Paid Conversion Report, co-published by Kyle Poyar's Growth Unhinged, ChartMogul, and ProductLed, based on a survey of 200 B2B software products conducted in January 2026 — found a median free-to-paid conversion rate of 8%, with roughly a 10x spread between the top and bottom quartile of performers.
This is a genuinely useful, real, dated number worth anchoring to specifically because of how carefully its methodology is described: a named survey of 200 real B2B software products, conducted in January 2026, with a stated median and quartile spread rather than a single, unsourced average. It is worth being direct about a widely circulated, competing claim this guide deliberately excludes: a generic “freemium converts at 2–5%” figure appears across a large volume of marketing and SEO content with no traceable original study behind it, and this guide could not verify a credible, named, dated primary source for that specific range. Rather than repeat an unsourced number just because it circulates widely, this guide relies on the Growth Unhinged/ChartMogul/ProductLed report's 8% median, while being transparent about its own methodology limits: a self-selected sample of 200 products, skewed toward companies in the roughly $1–10 million ARR range, which may not represent every stage or category of SaaS product evenly.
The 10x spread between top and bottom quartile performers in the same report is arguably the more actionable finding than the median itself: it means conversion rate is not primarily a fixed property of the freemium model in general, but something individual products vary on enormously based on execution — which free-tier limits are chosen, how clearly the paid tier's additional value is communicated, and how well the product's own onboarding actually demonstrates value before asking for payment. A founder disappointed by an early conversion rate well below 8% shouldn't necessarily read that as evidence freemium itself doesn't work for their category — the same report's own data shows enormous, real variance driven by execution, not just by model choice.
Freemium is a distribution strategy, not just a pricing choice
It's worth naming a distinction that gets lost when freemium is discussed purely as a pricing model: for many products, its real function is customer acquisition, not monetization directly. A free tier exists to let a genuinely useful product spread with minimal friction — no purchase decision, no approval process, nothing standing between a prospective user and trying the product — with the expectation that a meaningful share of that broader base converts to paid later, either directly or by influencing a paid decision elsewhere in their organization. Judging a freemium strategy purely on its direct conversion rate, without accounting for this broader distribution effect, risks underrating a free tier that is working exactly as intended even when its own direct paid-conversion number looks modest relative to the report cited above. This is precisely why the report's own quartile spread matters as much as its median: a product converting well below the median on direct paid conversion might still be using its free tier successfully as a distribution and word-of-mouth engine, a benefit that a single conversion-rate number doesn't fully capture on its own.
Setting Initial Pricing With No Existing Customers
Bringing the research above together into concrete guidance for a founder with no existing customer base to survey yet:
- 1
Run a lightweight Van Westendorp survey against prospective customers, not existing ones
The methodology only requires access to people who could plausibly buy the product — early waitlist signups, people who've expressed interest, or a small paid survey panel matching the target customer profile, not paying customers who don't yet exist.
- 2
Choose a pricing model based on how value actually scales, not on what looks simplest to build
Per the real taxonomy above, a per-seat model only makes sense if value genuinely scales with headcount using the product — copying a competitor's model without checking this fit is a common, avoidable mistake.
- 3
Treat the first price as a real hypothesis to test, not a permanent commitment
Per Patrick Campbell's own documented practice at Price Intelligently, revisiting pricing on a recurring basis — not just once at launch — is standard practice among companies that take pricing research seriously.
- 4
Resist defaulting to the lowest defensible price out of fear
Per Rao and Monroe's peer-reviewed research on price as a quality signal, an unusually low initial price carries a real, measurable cost of its own — it is not a safely neutral, risk-free starting choice.
None of this requires expensive market-research infrastructure to execute at a small scale. A founder can run a genuine, if informal, Van Westendorp survey using nothing more than a simple online form sent to a few dozen real prospective customers, and can apply the anchoring and price-ending research covered above without needing to run their own peer-reviewed experiment — the research already exists, and simply being aware of it while setting an initial price is enough to avoid several of the most common, avoidable mistakes a first-time founder makes when pricing a product purely by instinct.
Common Pricing Mistakes This Research Helps Avoid
Bringing the whole guide together into the specific, recurring mistakes a founder makes when pricing purely by instinct, and the specific piece of research covered above that corrects each one:
| The Instinct-Driven Mistake | What the Research Actually Shows |
|---|---|
| Copying a competitor's pricing model without checking fit | The real pricing model taxonomy: a model only works when it matches how value genuinely scales for the specific product, not because a competitor happens to use it |
| Guessing a single price rather than researching a range | Van Westendorp's 1976 methodology produces a realistic, defensible price range from real prospective customers, not a founder's single unresearched guess |
| Rounding prices to clean numbers out of a sense of simplicity | Anderson & Simester (2003) and Schindler & Kibarian (1996): price endings function as a real, measurable behavioral signal, not a purely aesthetic choice |
| Pricing defensively low to avoid scaring off early customers | Rao & Monroe (1989): price carries a real, statistically established relationship to perceived quality — an unusually low price has its own real cost |
| Treating the launch price as permanent | Patrick Campbell's documented practice at Price Intelligently: pricing research is a recurring, quarterly exercise for companies that take it seriously, not a one-time decision |
| Building tiers around engineering feature flags instead of customer segments | Real tier packaging works best when built around genuinely distinct customer needs, identified through the same persona-based research covered in this guide |
A Practical Framework
Bringing the research above together into an actual sequence for a founder pricing a product for the first time:
Run a real Van Westendorp survey before finalizing a number
Even a small, informal survey of real prospective customers, using the four real questions from Peter van Westendorp's 1976 methodology, beats a guess or a number copied from a competitor.
Match the pricing model to how value actually scales
Choose from the five real models — per-seat, usage-based, flat-rate, tiered, hybrid — based on what genuinely tracks the value the product delivers, not what's easiest to implement in billing code.
Be deliberate about anchoring and price-ending effects
The order prices are presented in, and whether they end in a round number or a 9, are not neutral details — both are the subject of real, peer-reviewed research, not marketing folklore.
Treat the first price as a hypothesis, and repeat the research
Per Price Intelligently's own documented practice, pricing research is a recurring exercise, not a one-time decision made before launch and never revisited.
Frequently Asked Questions
What are the five real SaaS pricing models?
Per-seat/per-user (charging by headcount), usage-based/consumption (charging by actual usage), flat-rate (a single fixed price), tiered (fixed packages at different levels, often with a free tier), and hybrid (combining a base fee with a usage-based component). Each fits different products depending on how value actually scales for the customer.
What is the Van Westendorp Price Sensitivity Meter?
A real pricing research methodology developed by Dutch economist Peter van Westendorp in 1976, which asks four specific questions (too cheap, a bargain, expensive, too expensive) to map out a realistic acceptable price range from real prospective customers, rather than relying on a founder's single guess.
Is there a real, credible alternative to Van Westendorp for pricing research?
Yes — conjoint analysis, a market-research technique applied to marketing since the 1970s, asks respondents to choose between hypothetical product-and-price bundles to isolate which specific features actually drive willingness to pay. It requires more survey infrastructure than Van Westendorp, which is why Van Westendorp is more commonly used by early-stage products with a smaller customer base to survey.
Does ending a price in .99 or 9 actually change buying behavior, according to real research?
Yes — a 2003 field experiment by Eric T. Anderson and Duncan I. Simester (Quantitative Marketing and Economics) found an identical item sold better at $39 than at the lower price of $34. A separate 1996 study by Robert M. Schindler and Thomas M. Kibarian found a real sales increase from .99-ending prices versus .00-ending prices.
What is anchoring, and does it really affect what people are willing to pay?
Yes — Dan Ariely, George Loewenstein, and Drazen Prelec's 2003 Quarterly Journal of Economics paper demonstrated this directly: MIT students who wrote down the last two digits of their Social Security number bid systematically higher on products when those digits happened to be higher, despite the number being logically unrelated to the products' value.
Does a higher price actually make a product seem better to buyers?
Yes, per peer-reviewed research — Akshay R. Rao and Kent B. Monroe's 1989 integrative review in the Journal of Marketing Research, synthesizing a substantial body of prior experiments, established a statistically significant positive relationship between price and buyers' perceived quality.
Is there a real, documented example of a company changing its pricing model with reported results?
Yes — New Relic shifted from a bundled infrastructure/SKU pricing structure to a hybrid per-user-plus-usage model, documented by billing company Zuora's own account, which reported roughly a 15% increase in committed ARR following the change. That specific figure comes from Zuora's account of the transition, not an independently audited New Relic filing, and should be verified directly before being cited as a hard fact.
What is the real freemium-to-paid conversion rate, according to credible data?
The 2026 Free-to-Paid Conversion Report (Growth Unhinged, ChartMogul, and ProductLed, surveying 200 B2B software products in January 2026) found an 8% median conversion rate, with roughly a 10x spread between top and bottom quartile performers — a more credible, dated, named source than the commonly repeated but unsourced "2-5%" figure.
How should a founder set initial pricing with no existing customers to survey?
Run a lightweight Van Westendorp survey against prospective customers (waitlist signups, expressed interest, or a small panel matching the target profile), choose a pricing model based on how value actually scales, and treat the first price as a hypothesis to revisit — not a permanent decision made once before launch.
Is it safer to price low when launching a new product?
Not necessarily — per Rao and Monroe's peer-reviewed research, price functions as a real quality signal to buyers, meaning an unusually low price can genuinely undercut perceived quality rather than simply being a safe, risk-free choice. A very low starting price carries its own real cost.
How is this guide different from your Product Strategy Frameworks guide?
Our Product Strategy Frameworks guide covers general prioritization discipline for any product decision. This guide applies real, specific pricing research — Van Westendorp, peer-reviewed pricing psychology, real conversion data — to the pricing decision specifically.
How is this guide different from your MVP Development guide?
Our MVP Development guide covers what a first build costs and how long it takes to ship. This guide picks up after that — what to actually charge for the product once it exists, using real research methods rather than guesswork.
What are the real limitations of the Van Westendorp methodology?
It asks people to imagine paying for a product, which reliably differs from actual purchasing behavior — a known gap across survey-based market research generally. It also asks about a single product in isolation, without directly surfacing how a price compares against named competitors. It's a well-reasoned starting range to test against real behavior after launch, not a final, guaranteed answer.
How large should an annual billing discount be compared to monthly pricing?
There's no single authoritative recommended figure — it's a real, company-specific trade-off between the value of prepaid, more predictable revenue with lower payment-churn risk against the discount needed to make prepayment attractive enough for a customer to actually choose it.
What is the safest way to change a pricing model without alienating existing customers?
Grandfathering — letting existing customers keep their current price and terms for a defined period or indefinitely, while the new pricing applies only to new customers. This avoids the kind of abrupt, forced change that produces real customer backlash, and connects directly to the same sequencing discipline covered in our guide on sunsetting or migrating a legacy product.
Is freemium primarily a pricing model or a distribution strategy?
Both, but its distribution function is easy to underrate. A free tier's real value for many products is letting the product spread with minimal friction, with a meaningful share of that broader base converting later — which means judging a freemium strategy purely on its direct conversion rate can miss a real distribution benefit the number doesn't fully capture.
None of the research in this guide replaces a founder's own judgment about their specific product and market — Van Westendorp, conjoint analysis, and the peer-reviewed psychology research above all inform a pricing decision, they don't make it automatically. What they do replace is the far more common alternative: a number chosen by instinct, copied from a competitor without checking whether the underlying model actually fits, or set defensively low out of a fear that has its own real, measurable cost. A founder who runs even one of these real, decades-old research methods before finalizing a price is already doing more diligence on this decision than most first-time founders do.