Pricing Architecture: How to Set Prices That Scale
The first price I ever set for a SaaS product was ₱999 per month. I picked it because it sounded reasonable, ended in a 9, and was cheaper than what I thought competitors were charging. It was wrong in three directions simultaneously — too low to sustain operations, too high for the entry-level segment I was actually reaching, and untethered to any value metric that mattered to the customer.
Pricing is the lever that touches every other metric in your business: CAC payback, margin, churn, LTV. Get it wrong and you build a product that works but a business that leaks. The advice you'll find on pricing usually lands in one of two places: "charge what you're worth" (useless) or "do value-based pricing" (correct in principle, silent on mechanics). This piece is about the mechanics. How to derive an actual number from unit economics, positioning, and competitive reality — without needing 2,000 survey respondents or a pricing consultant.
The Floor: What Your Unit Economics Demand
Before you can charge anything, you need to know the minimum price at which you stay solvent. This is not your target price. It is the lowest number you can accept before you are actively destroying the business.
The floor calculation has three inputs: Cost of Goods Sold (COGS) per unit, operating overhead allocated per customer, and minimum acceptable gross margin.
For a SaaS product, COGS per customer typically includes cloud infrastructure costs (compute, storage, bandwidth), third-party API costs that scale with usage, and direct customer support time (measured in hours per month × loaded hourly cost). Overhead allocation takes your fixed costs — team salaries, software subscriptions, legal and compliance — divides them by current or target customer count, and adds that per-customer burden.
If your COGS per customer is $18/month and your overhead allocation is $12/month, your total cost basis is $30/month. If you are targeting 60% gross margin (reasonable for early-stage SaaS), your floor is $30 / (1 - 0.60) = $75/month. Charging anything below $75 means you either absorb losses or compress margin to a point where you cannot invest in growth.
Most founders skip this step and price based on gut or competitive anchoring. The result is a product that feels financially functional at low customer counts but whose unit economics collapse at scale because the margin was never there to begin with.
Calculate your floor first. Write it down. Do not set a price below it for anything except a deliberately loss-led acquisition tier where the economics of the upsell justify the subsidy.
The Ceiling: What Your Best Customer Will Pay Without Hesitation
The ceiling is not what the market will theoretically bear. It is what the most motivated, most capable customer in your target segment will pay without requiring a sales cycle to justify the expense.
The fastest way to find this number is to talk to the people who have the problem your product solves and already spend money solving it — even badly. Ask them three questions:
- What are you currently spending (in money or time) on this problem?
- At what price would this feel like a no-brainer?
- At what price would you start to question whether it's worth evaluating?
Question 1 anchors you to the economic value already in play. Question 2 gives you the ceiling. Question 3 identifies the hesitation threshold — the price above which you start paying an attention tax in the sales process.
The ceiling is also constrained by your value metric. If you price per seat and your product saves one hour per user per week at a $50/hour loaded labor rate, the annual value per seat is roughly $2,600. A seat price above $800/year starts requiring explicit ROI justification. Below $400/year, you are leaving money on the table. The ceiling is not a feeling — it is a ratio of price to delivered value, and customers have an intuitive sense of when that ratio is off.
One important caveat: your ceiling differs by segment. An enterprise buyer in Singapore has a different ceiling than an SMB owner in Cebu, even for identical value delivered. Segment ceilings matter more than a blended ceiling when you are deciding which segment to lead with.
Positioning Anchor: Where You Sit in the Market and Why That Constrains the Range
Positioning is a constraint on pricing, not a target. Once you declare a market position — premium, value, specialist — you are accepting that your price must signal that position consistently.
A product priced below the market median signals "accessible" or "value," regardless of what the landing page says about enterprise capability. A product priced in the top quartile signals "serious investment" — which attracts buyers who expect a commensurate outcome and will churn angrily if they don't get it.
For HavenWizards' portfolio ventures, we have consistently found that mid-market positioning ($X to $3X range of the category median) works well for products that have a genuine differentiator but cannot yet support the enterprise sales motion required to sustain premium pricing. The error to avoid: pricing in the value tier while delivering a premium product, because you attract price-sensitive customers who churn the moment a cheaper competitor emerges.
The positioning anchor also informs your tier structure. If you lead with a freemium or entry tier, that tier must reflect your positioning — it cannot be so stripped down that it communicates "this product is for people who can't afford real software." The entry tier shapes perception of the whole product.
Write out your intended market position explicitly, then check whether your current price range is consistent with it. If they conflict, something needs to change — either the price or the positioning claim.
Three Price-Testing Methods That Don't Require a Large Sample
Most pricing guidance assumes you have 500 survey respondents and six months to run A/B tests. Founders building their first or second venture rarely have either. Here are three methods that produce actionable signal with small samples.
The Van Westendorp Price Sensitivity Meter (5-15 respondents). Ask four questions: At what price is this too cheap to trust? At what price is this a bargain? At what price does this start to feel expensive? At what price is this too expensive to consider? Map the responses. The acceptable price range is between the intersection of "too cheap" and "too expensive" curves. The optimal price point is roughly where "bargain" and "expensive" intersect. You do not need statistical significance — you need directional clarity.
Concierge pricing (3-5 customers). Before you build pricing tiers, offer the product at two or three different price points manually to different customer conversations. Not on the website — in sales conversations. "We're currently working through our pricing, here's what I'd offer you today." Observe willingness to continue without pushback. This produces real behavioral data, not stated preferences. The difference matters: people who say they would pay $200/month and people who actually hand over $200/month are not the same population.
The price-objection inventory (ongoing, 5+ sales calls). In every sales call or trial-conversion attempt, ask directly: "Is price a factor in your decision?" Capture the exact objection. There are only three meaningful responses: "No, we'd pay more if the product did X" (underpriced or wrong feature set), "Yes, we need a lower entry point" (mismatched tier structure), and "Yes, this is beyond our budget entirely" (wrong segment). Each points to a different fix.
The Price Increase Decision Framework
You are underpriced if any of the following is true: your trial-to-paid conversion exceeds 30% without a sales process, customers rarely negotiate on price, your customer success team reports that churn reasons cluster around product gaps rather than price, and no customer has ever complained that your pricing is too high.
When you are underpriced, the instinct is to raise prices dramatically. The better approach is graduated repricing — increase prices for new customers first, observe conversion impact for 90 days, then migrate existing customers to a new tier with a grandfather period of 6-12 months.
The specific sequence: (1) Set new pricing for all new signups. (2) At 30 days, measure whether new-customer conversion changed significantly. Significant means more than 15% decline. If conversion held, proceed. (3) At 60 days, send existing customers notice of upcoming price change with the grandfather deadline clearly stated. (4) At 90 days, verify that churned-due-to-price is under 5% of the existing customer base. If it exceeds 5%, you either moved too fast or misjudged the ceiling for that segment.
The thing most founders fear — that a price increase will cause mass churn — almost never happens if the product is delivering value. What causes churn is raising prices on customers who were already wondering whether the product was worth it.
The Pricing Decision Made Wrong and Then Corrected
One venture in the portfolio launched an operations workflow tool at $49/month flat, chosen because it matched the price of a competitor we'd benchmarked. The pricing held for four months. Conversion was reasonable (around 18% trial-to-paid) but expansion revenue was zero — customers signed up, got value, and stayed on the same plan forever. MRR grew only with new customers. NRR was 92%, meaning we were losing 8% of revenue annually even with no churn in the traditional sense.
The diagnosis: the flat pricing structure had no expansion mechanic. Customers had no reason to upgrade because everything was in the base tier. The fix required restructuring the product rather than just changing a number. We separated the product into a base tier (core workflows, $49/month, limit of 3 active automations) and a growth tier ($149/month, unlimited automations plus the analytics dashboard). Existing customers were offered the growth tier at $99/month for 12 months.
Within 90 days of the restructure, 34% of existing customers upgraded to the growth tier. NRR moved from 92% to 118%. The insight was not that $49 was wrong — it was that a single flat price destroyed the expansion motion before it started. The price architecture matters as much as the price itself.
What to Do When Pricing Architecture Is Not the Problem
There is a category of pricing failure that looks like bad pricing but is actually a bad product signal. If you have calculated the floor, tested the ceiling, aligned with your positioning, and used at least one of the three testing methods — and conversion is still poor — pricing is probably not the root cause.
The specific indicators that the problem is not price: customers complete demos or trials but do not convert, discount offers produce minimal uplift in conversion, churned customers cite reasons other than cost, and freemium users do not upgrade even when the upgrade path is obvious.
When you see these patterns, the correct response is not to change the price. It is to investigate whether customers are experiencing the value that the price assumes. In our portfolio, we have seen products priced correctly on paper where conversion failed because the onboarding flow did not get customers to the value moment within the trial window. The price was right; the path to experiencing the value was broken. Dropping the price in that scenario would have reduced revenue without fixing the conversion problem.
Pricing architecture works when the value is real and the customer can experience it. If either condition is absent, no price architecture repairs the underlying problem.
Putting It Together
Pricing is not a number you set once. It is an architecture you build and refine as you gather evidence about what your customers value and what they will pay to get it.
Start with the floor (your unit economics demand it), find the ceiling through direct customer conversations, anchor it to a positioning claim you can sustain, and test at small scale before committing. When you find yourself underpriced — and most early-stage products are — increase gradually, lead with new customers, and watch the data.
The goal is a price that reflects real value, covers real costs, and creates a real expansion mechanic. Everything else is cosmetic.