Founders are the worst judges of what their software is worth

Ask a SaaS founder how they arrived at their price and the answer tends to come in one of two forms. Some added a margin to what it costs them to deliver the product. Others studied the competition's pricing page and settled somewhere in the neighbourhood. Both answers sound prudent, the kind of thing that survives a board meeting without raised eyebrows. Both are wrong, and wrong in precisely the same way.

Neither method consults the person paying. Yet the customer is the only party whose judgment determines whether a deal closes, renews or grows. What they weigh is not the vendor's cost base or a rival's list price. It is the value they expect to receive. Value-based pricing begins there, with a question that sounds obvious and is rarely asked in earnest: what is this worth to the buyer, and what will they pay for it?

The comfortable methods are the expensive ones

Cost-plus pricing has the appeal of arithmetic. It is simple to calculate and simple to defend. But the number it produces bears no relation to anything the customer cares about. Nobody buying software knows or cares what the vendor pays its engineers. They care about what the product does for them. The model is also brittle. When costs rise, margins compress. When a competitor cuts prices, cost-plus has no reply. The profit line drifts flat, then slides.

Competitor-based pricing is a modest improvement, because it at least reflects what the market will tolerate. Its flaw is subtler. It borrows someone else's strategy for a product that exists precisely because it is different. Price a distinctive product like its nearest alternative and you manage to be wrong twice: you undercharge the customers who prize the difference and overcharge the ones who barely notice it.

The cost of this indifference can be measured. A study of 512 SaaS companies found that a 1% improvement in monetization lifted the bottom line by 12.7%. The same gain in retention delivered 6.71%, and in acquisition just 3.32%. Pricing is the most powerful lever a software company has. It is also the one most reliably left alone.

A price is the output of a system, not a decision

The common misreading of value-based pricing is that it simply means charging more. It does not. It means building the machinery that tells you what to charge, and three parts of that machinery must be in place before any number can hold.

The first is a set of buyer personas grounded in evidence rather than imagination. The familiar version, with its adjectives and stock photography, is useless here. What matters is data on who buys, which outcomes they pursue, which features they value and what they will pay. Willingness-to-pay research, usage analysis and structured customer interviews supply it. Without that foundation, every pricing decision is a guess wearing the costume of a strategy.

The second is a value metric, the unit a company charges for. Twilio charges per API call. DocuSign charges per signed document. Stripe takes a percentage of the revenue it processes. In each case the metric rises as the customer succeeds, so revenue expands without anyone picking up the phone. A good metric grows with the customer's value, remains predictable enough for a buyer to budget, and tracks the vendor's cost of delivery. The HOPE framework offers a useful filter, scoring candidates on operational ease, meaning whether the metric can be tracked and billed, and on customer clarity, meaning whether the buyer understands why they are paying for it. Strong on both, and it becomes the primary lever. Clear but hard to operate, and it belongs on the roadmap. Weak on both, and it should be discarded.

The third is a choice about positioning. Knowing what customers will pay is not the same as deciding how much of it to take. Penetration pricing captures less than perceived value in order to accelerate adoption. Neutral pricing takes a fair share. Premium pricing takes more, which demands genuine differentiation and a tolerance for lower volume. None is correct in the abstract. The answer depends on a company's stage, its competitive position and the kind of growth it wants.

Founders consistently underprice their own products

Willingness to pay sounds like something only a customer's subconscious could reveal. In practice, two well-established methods, run in sequence, measure it with surprising reliability.

The Van Westendorp method comes first. It asks respondents four questions: at what price would the product seem so cheap that its quality becomes suspect, at what price does it begin to feel expensive, at what price is it too expensive to consider, and at what price is it a bargain? The answers mark out an acceptable range with a floor and a ceiling. That range tells a company where to play. It does not tell it where to land.

The Gabor-Granger method narrows the field. Respondents see specific price points inside the range and say whether they would buy at each one. The result is a demand curve, which shows the price that maximizes revenue and exactly what a company forfeits by stepping one rung lower.

Together the two methods replace instinct with evidence. I have run both through Qualtrics in more than 20 SaaS engagements, and the pattern barely varies. The data points to prices 20 to 40% higher than the founding team believed customers would accept. Founders, it turns out, are among the least reliable judges of what their own products are worth.

Packaging is where pricing strategy becomes visible

The pricing model settles what a company charges for. Packaging settles how that charge is presented to different kinds of buyer. Good-Better-Best remains the dominant structure, and for sound reasons: it mirrors how most markets segment, and buyers reliably anchor to the middle option. Prices across the three tiers should run at roughly one, three and eight to ten times the entry level. Tiers should be named for a customer outcome or stage of growth, not for a count of features.

Two failures recur with depressing regularity. The first is designing tiers on a whiteboard instead of from evidence. Invented tiers tend to produce a middle option nobody wants and an enterprise tier whose only function is to make the middle look reasonable. Tiers built from willingness-to-pay research and usage data reflect where customers actually cluster and what they will pay at each level.

The second failure is placing the core job behind an add-on. If a customer cannot do the main thing the product promises without paying extra, then every tier is incomplete. That is not monetization. It is a competitive weakness with a price tag attached.

The companies that win never stop repricing

The final misconception is that pricing is a project with an end date. The evidence says otherwise, and says it bluntly. A study of 96 SaaS companies with more than $5M in annual recurring revenue compared their ratio of customer lifetime value to acquisition cost. Companies with no pricing function managed 1.68. Those that reviewed prices once a year reached 3.23. Those that optimized continuously reached 11.09.

The lesson is uncomfortable for teams that treat pricing as a launch task. Every feature that adds value to the product is also a question about price. If the price does not move with the value, the difference simply goes uncollected. A quarterly rhythm keeps the two in step: diagnose performance through net revenue retention, average revenue per user and churn; gather qualitative signal from sales, customer success and support; run targeted experiments on specific segments; then make permanent whatever works.

AI makes the case for value pricing impossible to ignore

If value-based pricing is advisable for conventional software, it is unavoidable for AI. Seat-based pricing collapses the moment an agent does the work. The customer's value climbs while the number of seats holds steady or falls, which means the vendor is penalized for its own success. No pricing model survives that contradiction for long.

For most AI products the answer is a hybrid. A platform floor covers the cost of delivery and gives the customer's finance team a predictable line item. A variable component, tied to usage or to outcomes, captures the upside as the customer grows.

Outcome-based pricing goes furthest. The customer pays per resolved ticket, per qualified lead or per share of the savings achieved. It prices the work rather than the access, which makes it the most defensible model for agents. But it rests on four conditions: a measurable result, an agreed baseline, an attribution rule settled before the engagement begins, and a floor that protects margin when volume dips. Remove any one of them and the model tends to dissolve into argument.

The question worth asking is the one nobody asks

The companies that price well are not cleverer than their rivals. They have simply learned to ask a different question. Not what the product costs to build, and not what the competition charges, but what it is worth and to whom. The answer is rarely comfortable, because it exposes how much has been left on the table. It is almost always higher than anyone in the room expected.

Cost-plus pricing protects the margin. Competitor pricing protects the ego. Only value-based pricing protects the business.