How to Price a SaaS Product When You Have No Benchmark

A founder building a genuinely new category of workflow software once told me he’d priced his product at $49/month because that’s what a semi-adjacent competitor charged, even though his product solved a different problem for a different buyer with a completely different value ceiling. Eighteen months later, after finally running real willingness-to-pay research, he moved to a $400/month starting price and lost almost no customers in the transition — the market had been willing to pay 8x what he’d been charging, and the only reason he hadn’t found that out sooner was that “just match what looks similar” felt safer than actually researching the number.

This is the trap almost every founder in a genuinely novel category falls into: when there’s no direct competitor to anchor against, the instinct is to grab whatever adjacent pricing feels roughly comparable and hope it’s close enough. It usually isn’t, and the gap tends to run in the direction of underpricing, because founders systematically underestimate the value their product creates relative to the pain it removes.

Price the Value Created, Not the Feature List

The starting move when there’s no benchmark is abandoning feature-based pricing logic entirely and building from the value metric instead — the specific, quantifiable outcome your product creates for the customer, expressed in terms the customer already tracks internally (hours saved, revenue protected, cost avoided, risk reduced).

This requires actually calculating the dollar value of that outcome for a representative customer, not estimating it loosely. If your product saves a specific role 6 hours a week, multiply that by a reasonable fully-loaded hourly cost for that role (not just base salary — include benefits and overhead, which usually adds 25-40%) and you have an annualized value number. If your product prevents a specific category of costly mistake — a compliance violation, a missed renewal, a data error that causes downstream rework — estimate the frequency and average cost of that mistake without your product, and that becomes your value anchor.

A standard rule of thumb in value-based pricing is capturing somewhere between 10-30% of the value created as the price, with the exact percentage depending on how directly attributable the value is to your product versus how much it depends on the customer’s own execution. A product that directly and unambiguously creates the value (automated calculation, direct time replacement) can capture toward the higher end of that range; a product that enables value but still requires significant customer effort to realize it should price toward the lower end, since a meaningful part of the outcome isn’t solely attributable to the tool.

Run Actual Willingness-to-Pay Research Before Setting a Number

Guessing at value capture percentages is still guessing. The way to replace guesswork with real data, even pre-launch or in early stages, is structured willingness-to-pay research with actual prospective buyers — not friends, not existing investors, people who genuinely represent your target buyer and have real budget authority or influence over a purchase like this.

The Van Westendorp Price Sensitivity method remains the most practical structured approach for a genuinely new category, because it doesn’t require the respondent to have a benchmark either — it asks four questions about the same product: at what price would this be so cheap you’d question the quality, at what price would this be a bargain, at what price would this start to feel expensive but you’d still consider it, and at what price would this be too expensive to consider at all. Plotting the responses from a sample of 30-50 qualified prospects across these four questions produces a range, not a single number, but that range is dramatically more grounded than an adjacent-competitor guess, because it’s coming directly from people who actually represent the buying population.

The critical detail most people get wrong when running this research: it has to happen with people who understand the specific problem being solved, described concretely, not an abstract product pitch. Show them the actual value proposition, ideally with a demo or mockup, before asking the pricing questions, because pricing questions answered against a vague description produce vague, unreliable answers. The specificity of the problem framing directly determines the reliability of the pricing data you get back.

Anchor Against Adjacent Alternatives, Even Imperfect Ones

Even without a direct competitor, there’s almost always an adjacent alternative your buyer would consider — a manual process, a spreadsheet-based workaround, a more generic tool being used as a substitute, or simply the cost of not solving the problem at all (status quo cost). Understanding what that alternative currently costs, in time or money or risk, establishes a real-world anchor point that’s more relevant than an unrelated competitor’s sticker price.

If your buyer currently pays a contractor $3,000/month to do manually what your software automates, that $3,000 figure is a far more relevant anchor than the pricing of some other SaaS tool that happens to occupy a similar mental category. Interview a handful of prospective customers specifically about what they currently do to solve this problem and what that costs them — not just in direct spend but in time, error rate, and opportunity cost — before finalizing a number. This “cost of the alternative” framing also becomes core sales messaging once pricing is set, because it gives the sales team a concrete way to justify the number that has nothing to do with what any competitor charges.

Structure Before Number: Decide the Pricing Model First

A mistake that compounds an already-uncertain price point is deciding the actual dollar figure before deciding the pricing model — flat fee, per-seat, usage-based, tiered by feature access, or some hybrid. The model shapes how the number needs to be structured and how it scales with customer size, and getting this decision wrong creates problems no amount of number-tweaking fixes later.

Usage-based or outcome-based models tend to fit products where value scales directly and visibly with a measurable unit of activity (transactions processed, records handled, actions completed) and where the buyer will find this framing intuitive rather than anxiety-inducing about unpredictable bills. Per-seat models fit products where value is tied to how many people are actively using the tool day-to-day, and they have the advantage of being easy for a buyer to forecast, but they can undercharge for products where a small number of seats create enormous value (in which case a hybrid — a base fee plus usage — often captures value better than seats alone). Flat-fee tiered models work best when the product’s value is relatively consistent regardless of scale, or when usage-based billing would introduce more sales friction than it’s worth for the segment you’re selling to.

For a genuinely novel product, running the willingness-to-pay research above against 2-3 different model structures, not just different dollar amounts within one model, often reveals that the model itself is a bigger lever on realized revenue than the specific number chosen within any single model.

Build in a Deliberate Price Increase Path From Day One

Because early pricing on a novel product is inherently an estimate, not a certainty, the smartest structural decision is designing the go-to-market with an explicit expectation that price will move as real data comes in — and communicating that internally (and sometimes even to early customers) rather than treating the first price point as permanent.

A practical mechanism: grandfather early customers at their signed price for a defined period (12-18 months is common) rather than promising it indefinitely, which preserves goodwill with the customers who took an early bet on you while leaving room to correct pricing as real usage data, retention data, and expanded willingness-to-pay research accumulate. Founders who lock in “customers for life at this price” language in the excitement of early sales conversations frequently regret it once the real value story becomes clear and the original number turns out to have left significant revenue on the table.

Watch Retention and Expansion as the Real Pricing Signal

Once a product is live, even at a rough initial price, the most reliable ongoing pricing signal isn’t competitor movement or founder intuition — it’s the combination of retention rate and expansion behavior at different price points, if you have the ability to run even informal price experiments across cohorts (new customers at a different starting price than existing ones, for instance).

A price that’s too low relative to true value tends to show up as extremely high retention and enthusiastic expansion but slower-than-expected revenue growth relative to the value being delivered — customers stick around and grow usage without much friction, which feels good but is actually a signal of underpricing. A price that’s too high relative to value tends to show up as elevated churn specifically among customers who had a rough first 60-90 days, or unusually high price-objection rates in sales conversations relative to how differentiated your product actually is. Reading these signals requires actually tracking them by cohort and price point, not just watching aggregate revenue, but for a novel product with no external benchmark, this internal signal — how real customers behave at the price you set — ultimately becomes a more trustworthy pricing benchmark than anything an outside comparison could have offered in the first place.

A Worked Example: Calculating a Value-Based Price From Scratch

Take a real calculation to make the value-capture logic concrete. Say your product automates a compliance review that a mid-market customer currently has a compliance analyst do manually, taking roughly 8 hours a week. Fully loaded cost for that analyst role (salary plus benefits and overhead) works out to about $65/hour, so the manual process costs the customer roughly $520/week, or about $27,000 annually. Your product replaces essentially all of that time, so the value created is close to the full $27,000, not a partial estimate, because the automation is direct and unambiguous rather than something the customer still has to partially execute themselves.

At a 15% value-capture rate — reasonable for a product that directly and fully replaces the manual work rather than merely assisting it — that puts a defensible annual price around $4,000, or roughly $335/month. Compare that to a founder anchoring off an adjacent project-management tool priced at $49/month: the gap between $49 and $335 is exactly the kind of value-based repricing opportunity that gets left on the table when pricing starts from “what do similar-sounding tools charge” instead of “what is this actually worth to the person paying for it.” Running the Van Westendorp questions against this same customer segment would likely validate a range clustering somewhere between $250-450/month, giving you both a bottom-up calculation and a top-down survey result converging on a similar number — which is the point where you can commit to a price with real confidence instead of a guess.

The Failure Mode: Averaging Instead of Segmenting

A common mistake once willingness-to-pay data comes in is collapsing it into a single average price across all respondents, when the more useful read is almost always segmented. If your research included both 20-person startups and 2,000-person enterprises, and their acceptable price ranges don’t overlap much, averaging them produces a number that’s wrong for both — too expensive for the small segment, too cheap for the large one, leaving real revenue on the table at the top of the market while still failing to convert efficiently at the bottom.

The fix is treating divergent willingness-to-pay data as a signal to build tiers rather than a single price, with the segments that emerged from the research becoming the basis for tier boundaries — not an arbitrary “Starter/Pro/Enterprise” split invented after the fact. If your Van Westendorp data shows small teams clustering around $150-250/month and larger organizations clustering around $800-1,200/month, that’s close to your tier structure already, before you’ve even decided on specific feature gates between them.

Sequencing the Pricing Work So It Doesn’t Stall Launch

Founders sometimes treat “we don’t have a benchmark” as a reason to delay launch indefinitely while chasing a perfect number. The more productive sequence: set an initial price using the value-metric calculation above (even a rough version, using reasonable estimates rather than perfect data), launch with it, and run the willingness-to-pay research in parallel with the first 10-20 real sales conversations rather than gating launch on completing the research first. Real sales conversations at a real price generate pricing signal faster than a research study does, because you’re watching actual objection patterns and close rates rather than survey responses. The two data sources should converge within the first quarter of selling — if they don’t, that’s the signal to run the price increase already discussed, sooner rather than later, since every month spent underpriced is quantifiable revenue that grandfathered early customers make expensive to recover retroactively.