Pricing Strategy for New Ventures: Price Before You Build

Ask a venture team about their pricing and you’ll almost always get the same answer, delivered in the same slightly embarrassed tone: “we’ll figure that out closer to launch.” I’ve heard it from two-person startups and from corporate ventures with eight-figure budgets, and it’s worth noticing what the sentence actually claims — that the single variable determining whether the business exists can be bolted on after every other decision has been made.

The best evidence we have says this is exactly backwards. Madhavan Ramanujam and Georg Tacke of Simon-Kucher & Partners — the consultancy that does more pricing work than anyone on earth — published Monetizing Innovation in 2016, built on the firm’s research across thousands of client engagements, and the headline number is brutal: around 72% of new products miss their revenue or profit targets, or fail outright. Their diagnosis isn’t bad products or bad marketing. It’s sequence. Companies design the product, build the product, and then ask what to charge — at which point the willingness-to-pay conversation isn’t an input to design, it’s a hostage negotiation with sunk cost. Their prescription is one sentence: have the willingness-to-pay talk with customers early, before you build, and design the product around what you learn. Price isn’t a number you put on the thing. It’s evidence about whether the thing should exist.

This is part nine of this series, and it’s here on purpose. Everything earlier — the structures, the labs, the metered funding — is about keeping new ventures alive long enough to find a business. Pricing is where you find out whether there’s a business to find.

The four ways monetization fails

Ramanujam and Tacke’s most useful contribution isn’t the statistic; it’s the taxonomy. When new products fail to monetize, the failures cluster into four types, and each one is a different disease with a different cure. Knowing which one you’re incubating is half the diagnosis.

Feature shock: too much for too many. The team crams in every feature anyone proposed, because each addition feels like added value, and produces an over-engineered product that confuses buyers and costs too much to build. Their canonical reading is Amazon’s Fire Phone (2014): Dynamic Perspective’s four face-tracking cameras, the Firefly recognition button, a premium $199-on-contract price matching the iPhone — a pile of genuinely clever engineering that customers hadn’t asked for and wouldn’t pay iPhone prices to get. Within weeks the price fell to 99 cents on contract; Amazon took a $170 million write-down. Nobody had established what any of those features were worth to a buyer before building all of them. Feature shock is what “we’ll figure out pricing later” looks like when it ships.

Minivation: the underpriced success. The product works, customers love it, revenue arrives — and it’s a fraction of what it should have been, because the team priced timidly and can never claw back up. This is the failure nobody files as a failure, which is what makes it insidious: the post-mortem never happens because there’s champagne instead of a corpse. Most of Simon-Kucher’s minivation examples are anonymized clients for an obvious reason — no company issues a press release announcing it left half the market’s willingness to pay on the table. But every pricing practitioner has seen it: the modest win that a single early WTP conversation would have revealed as a potential blockbuster.

Hidden gem: the value you don’t recognize. Something genuinely valuable sits inside the company — a by-product, a dataset, an internal tool — and never gets monetized because it doesn’t fit how the company thinks about what it sells. The classic examples orbit Kodak, which is why this failure type should feel familiar if you read the first post in this series: a company whose processes and values are tuned to one business model will systematically fail to see value that monetizes differently, even when it invented that value. The hidden gem is the innovator’s dilemma expressed as a pricing failure — the asset isn’t starved by the resource-allocation process, it’s never entered into it at all.

Undead: the product that should have died. The answer to a question no customer was asking, kept alive by internal conviction long past the point the evidence said stop. Their example is the Segway — pre-launch predictions of a transportation revolution, actual sales in the low tens of thousands over its first years against forecasts of thousands per week. The tell is that nobody ever tested whether people would pay for it at anything like the price it needed. An early willingness-to-pay conversation is the cheapest kill signal you can buy, and the undead product is what happens when a venture skips it and the funding gates that should have caught it rubber-stamp instead.

Notice the shared root. All four failures are information failures: in each case, a conversation about money, held before the build, would have changed the design, the price, or the go decision. That conversation is the discipline. Everything else in this post is technique for having it well.

How do you actually test willingness to pay?

Start with what not to do, because it’s the corporate default. Cost-plus pricing — total your costs, add a target margin, call it a price — feels rigorous and is almost information-free: it encodes everything about you and nothing about the customer. For a mature commodity business it’s at least stable. For a new venture it’s lethal in both directions — it underprices genuine innovation (your costs say nothing about the value created) and overprices weak products (your costs don’t care that nobody wants it). Competitor-based pricing is one notch better and shares the flaw: it imports someone else’s assumptions about a product that isn’t yours, and for genuinely new categories there’s often no honest comparable anyway. Value-based pricing — anchoring price to the customer’s next-best alternative plus the differentiated value you deliver against it — is the only method that forces you to learn something. Which is precisely why teams avoid it: it requires conversations that might return an answer you don’t like.

The oldest structured tool for those conversations is the Van Westendorp Price Sensitivity Meter (1976). You ask each respondent four questions about the product: at what price would it be so cheap you’d doubt its quality? At what price is it a bargain? At what price does it start feeling expensive, but you’d still consider it? At what price is it too expensive to consider at all? Plot the cumulative curves and the intersections give you an acceptable price range and an indicative optimal point. Be honest about what it does and doesn’t tell you: it maps the acceptable range — where you’re not obviously mispriced — from stated attitudes, with no competitive context and no actual purchase commitment. People are systematically better at bracketing than at predicting their own behavior. Van Westendorp is a rangefinder, not a cash-register forecast, and treating its “optimal price point” as gospel is how surveys replace evidence.

Ramanujam’s field version is even simpler, which is why it gets used: describe the concept, then ask three questions. What would you consider an acceptable price? An expensive price? A prohibitively expensive one? Then the follow-up that carries most of the information: why? The absolute numbers matter less than the spread and the reasoning — a tight cluster with articulate reasons means a real reference point exists in the customer’s head; answers scattered across an order of magnitude mean the value proposition itself is unclear, which is a product problem wearing a pricing costume. In my experience the “why” conversations are the single highest-yield discovery interviews a venture runs, because money makes people honest. A customer will praise anything in a feedback session. Ask them to bracket a price and their actual mental model of the value falls out.

Two structural moves follow from the interviews. First, segment by willingness to pay, not demographics. The segmentation that matters for a new venture isn’t “enterprises vs. SMBs” or “millennials vs. boomers” — it’s the clusters that emerge when you sort by what people value and what they’ll pay for it. Two identically sized companies in the same industry can sit at opposite ends of the WTP distribution because one has the burning problem and one doesn’t. Second, package around the segments — the good-better-best structure exists because it converts a WTP distribution into revenue instead of forcing one price to lose at both ends. Three tiers, each anchored to a real segment’s needs, with the fence features (the ones your high-WTP segment can’t live without) held out of the cheap tier. Build the tiers from interview data and the packaging designs itself; build them from a features spreadsheet and you get feature shock in triplicate.

The pricing metric matters more than the price

Here’s the decision teams spend the least time on and should spend the most: not how much to charge, but what unit to charge for. Per seat, per transaction, per gigabyte, per resolved ticket, per successful outcome — the metric determines how your revenue scales with your customer’s value, and getting it right routinely matters more than getting the number right, because a good metric self-corrects while a bad one caps you forever.

The qualitative pattern from the modern usage-based companies is instructive. Snowflake charges for consumption — compute and storage actually used — so a customer’s bill grows precisely as their use of the product grows, and Snowflake’s celebrated net revenue retention is substantially a pricing metric achievement: expansion is built into the unit of charge rather than dependent on a sales rep renegotiating seats. Twilio charges per message and per minute, which meant a two-person startup could start at pocket-change prices and grow into a seven-figure account without ever hitting a repricing wall. The older, stranger, best example is Rolls-Royce’s “power by the hour”: airlines don’t buy jet engines, they buy hours of thrust, with maintenance bundled — the metric aligned so tightly with customer value that the customer’s success and the vendor’s revenue became the same number. The test for a metric is exactly that alignment: does the customer’s bill grow when, and only when, the value they receive grows? Per-seat pricing fails this test more often than people admit — plenty of products deliver value that has nothing to do with headcount — and it persists mostly because it’s easy to administer. Choose the metric during discovery, not after: it’s a design decision, and some products need rebuilding (metering, entitlements, billing plumbing) to support the right one.

Launch: penetration, skimming, and the freemium trap

The classic launch question is penetration versus skimming, and the real tradeoff is sharper than the textbook version. Penetration — price low to grab share fast — is justified when there are genuine learning-curve economics or network effects, where early volume compounds into a structural advantage that later margin can harvest. Skimming — start high, serve the high-WTP segment first, walk down the curve — fits products with strong differentiation and no winner-take-all dynamics. The asymmetry everyone underweights: prices move down easily and up almost never. Your launch price is an anchor that sets the reference point for every future negotiation, and a venture that anchors low without a compounding reason has usually just committed minivation on day one. Raising prices on an installed base is one of the most painful motions in business; cutting them is a press release. When the strategic case is ambiguous, that asymmetry says start higher.

Which brings us to freemium, the most common WTP-destroying decision in software — not because free tiers never work, but because of why teams choose them. Freemium chosen as a deliberate weapon (a metered on-ramp into a usage-priced product, a network-effects land grab with a modeled conversion path) can be sound. Freemium chosen out of fear — “nobody will pay for this yet, let’s go free and figure out monetization later” — is the willingness-to-pay conversation deferred indefinitely, at scale. You accumulate a user base selected precisely for unwillingness to pay, learn nothing about value, and eventually face converting people whose entire relationship with you was founded on not paying. If the honest reason for your free tier is that you’re afraid of the pricing conversation, the free tier is the feature shock of business models: something for everyone, value evidence for no one.

Corporate ventures deserve a specific warning here, because they underprice new products with remarkable consistency, and the mechanism is structural rather than stupid. A venture subsidized by a parent P&L faces no survival pressure from its price: payroll clears whether the price is right or half of right, so pricing courage has no forcing function. Add the parent’s instinct to price the new thing “accessibly” so adoption charts look good at the quarterly review, and you get ventures that hit usage milestones while quietly proving nothing about value. This is one more argument for the metered-funding discipline from earlier in this series: a venture funded against validated-learning milestones, where revenue quality is a gate, can’t hide behind subsidized pricing. If the parent’s money means customers never have to vote with theirs, you’ve bought growth statistics and sold the only evidence that mattered.

Pricing is an experiment stream, not a launch decision

The deepest reframe is temporal. Teams treat pricing as a decision — made once, near launch, under deadline pressure, then defended. Treat it instead as an experiment stream that starts before the build and never fully stops. Before building: WTP interviews shape what gets built at all and what metric it’s built to meter. During the build: Van Westendorp-style ranging and packaging tests refine tiers. After launch: revisit deliberately and often — quarterly is not too frequent in the first couple of years, because your early price was set at your point of maximum ignorance, and everything you learn afterward (who actually buys, what they compare you to, which features fence the tiers) is information your launch price didn’t have.

And keep the epistemology straight: a paying customer is the strongest evidence a new venture can produce. Interest is cheap, pilots are polite, letters of intent are theater — a signed invoice is a costly signal, which is exactly why it’s the validated-learning milestone that outranks the others. A venture that can’t yet show willingness to pay hasn’t validated a business; it’s validated attention. Every method in this post is really a way of pulling that strongest-evidence moment as early in the venture’s life as possible, when changing course is still cheap.

Which sets up the last post in this series. The money you charge isn’t just validation — it buys the only resource that actually kills startups when it runs out. Not competition, not technology: time. How to count it, and how to know whether you’re default alive or default dead, is where we go next.

Further reading

• Madhavan Ramanujam & Georg Tacke, Monetizing Innovation (2016) — the spine of this post; the four failure types and the case for the early WTP talk, from the firm that’s seen more pricing data than anyone.
• Peter Van Westendorp, “NSS Price Sensitivity Meter” (ESOMAR, 1976) — the original four-question method; read it alongside a modern critique so you use it as a rangefinder, not an oracle.
• Rafi Mohammed, “The Good-Better-Best Approach to Pricing” (Harvard Business Review, 2018) — the practical case for tiered packaging.
• Kyle Poyar’s OpenView work on usage-based pricing — the best running qualitative account of the Snowflake/Twilio-style shift from seats to consumption.