The Five Numbers Solo Founders Should Run Before Pricing a SaaS Product

Pricing is the most consequential number a SaaS founder sets and the one most often set by gut. A 10% pricing miss compounds across every customer, every month, forever. A 10% feature miss can be patched in a sprint. The asymmetry is the reason to do the math before launch, not after the first refund request.

The five numbers below are not benchmarks to compare yourself against. They are inputs to a single decision: the minimum price below which the unit economics break, and the maximum price above which the market will not buy. Anything in between is a pricing choice. Anything outside is a pricing mistake.

Why these five

Pricing models in textbooks list dozens of inputs. Most are second-order. The five here have non-substitutable signal: cut any one and the price you set is structurally unsound, not just suboptimal. COGS sets the floor. CAC sets the marketing budget. Sensitivity sets the safety margin. Willingness-to-pay sets the ceiling. Payback sets the cash plan. Outside the range these inputs bound, no amount of positioning fixes the math.

1. COGS-per-active-user

Cost of goods sold per active user is what one paying customer costs you to serve in a month, fully loaded. The formula:

COGS-per-user = (Hosting + Third-party APIs + Payment fees + Support cost + Variable infra) / Active paying users

Each component, concretely:

• Hosting and infra. Compute, database, storage, CDN bandwidth attributable to customer workload. Pull the actual cloud bill, not the projected one.
• Third-party APIs. The line item that ruins AI-product margin: token costs, embeddings, vector DB, model-inference fees. Usage-weighted, not flat.
• Payment processing. 2.5–3.5% on card, 0.5–1.0% on ACH for US-based products. International cards run higher.
• Support cost. Your own time at a loaded hourly rate counts. Solo founders skip this and underprice support-heavy products by a wide margin.
• Variable infra. Per-user files, per-user logs, per-user backups. Not the platform fees that exist whether you have one user or one thousand.

Worked example. A solo AI-writing SaaS at 200 paying users: hosting $80, OpenAI tokens $620, vector DB $40, Stripe fees on $50 ARPU = ~$310 across 200 users, founder support ~10 hours at a $75 loaded rate = $750. COGS-per-user = ($80 + $620 + $40 + $310 + $750) / 200 = $9.00/user/month. With a $50 ARPU, gross margin is ($50 − $9) / $50 = 82%.

Where founders get this wrong: they include the wrong things and exclude the right ones. Hosting platform fees that do not vary with users belong in operating expense, not COGS. Founder time building the product is OpEx. Founder time answering tickets is COGS, and ignoring it makes gross margin look 10 to 15 points better than it is. The Profit Margin Calculator handles the cost-stack arithmetic; the AI Product Margin Calculator is the version with token-cost inputs already wired up.

2. Customer acquisition cost

CAC is what you spend to get one paying customer, fully loaded. Most founder-reported CAC numbers exclude two large costs: the founder's own time and the long-tail spend that has not yet attributed.

The honest formula:

CAC = (Paid spend + Tools + Founder marketing time × loaded rate + Content production cost) / New paying customers in window

The window matters. CAC in month one of a campaign is uninformative. Use a 90-day rolling window, or longer if your sales cycle exceeds 30 days.

Common errors:

• Counting organic as zero CAC. If "organic" includes content marketing, SEO writing, podcast appearances, or community work, those have a real cost. Either pay yourself for the time, or pay a contractor and use that rate. Skipping this makes content channels look infinitely profitable when they are usually 12–24 month payback investments.
• Excluding free-trial servicing cost. If a free trial costs you $4 in API and $20 in support time per user and only 8% convert, the loaded CAC is $300 per converted customer before you spent a dollar on ads.
• Mixing channel CAC with blended CAC. Blended CAC averages cheap referrals with expensive paid search and produces a number that does not match any actual decision. Always run blended and per-channel side by side.
• Excluding tooling. Email platform, CRM, landing-page tool, ad-management subscriptions all belong in CAC. They are not free because they are recurring.

Worked example. A founder running Google Ads spends $600/month on ad cost, $80 on the email tool, $40 on the landing-page tool, and 25 hours of their own time at $75/hr loaded ($1,875). New paying customers in the 90-day window: 18 across the period, so 6/month. CAC = ($600 + $80 + $40 + $1,875) / 6 = $432/customer. The headline ad CAC of $600/6 = $100 looks healthy. The loaded CAC is 4x that. Both are real, the second is the one to plan from. The CAC Calculator walks through the components for one channel; the CAC Payback Calculator follows it with the months-to-recovery view.

3. Gross-margin runway sensitivity

This is the number nobody runs. It tells you how many months of runway you lose at current burn if COGS-per-user moves 20% in the wrong direction. Most solo founders treat COGS as static. It is not. AI-product token costs swing with provider pricing. Hosting bills swing with usage spikes. Support time swings with feature complexity.

The formula:

Runway sensitivity = ΔCOGS-per-user × Active users × Months of runway / Net cash position

Plain version: a 20% COGS increase on 200 users at $9 baseline COGS adds $360/month to burn. If you have 14 months of runway at current burn and $80,000 in the bank, that 20% jump cuts your runway by $360 × 14 / $80,000 ≈ 6%, or about 0.8 months. Modest sounding. Now run it again at 1,000 users: $1,800/month, a 31% runway cut, around 4.4 months gone. The same percentage shock has very different consequences at different scales.

Why this matters for pricing: a price set at 70% gross margin at launch, when you have 100 users and small absolute COGS, can survive a token-cost shock. The same 70% gross margin price at 2,000 users cannot. Pricing decisions made at low scale need a margin cushion that low-scale data does not visibly require. The Startup Runway Calculator and Profit Margin Calculator together stress-test the COGS shock scenario before it happens.

Run a 20% COGS-up scenario at the user count you expect 12 months out, not the count you have today. Price for the scaled cost stack. Founders who skip this either cut prices in panic at scale or eat the margin loss for a year while replatforming.

4. Willingness-to-pay anchor

Cost tells you the floor. Willingness-to-pay tells you whether the ceiling has any room above the floor. Without it, pricing degenerates into cost-plus, which Nagle calls "the most common error in pricing decision making"^[3] for a reason: it ignores what the buyer would actually pay.

The cheapest credible willingness-to-pay measurement for a small audience is the Van Westendorp Price Sensitivity Meter^[2]. Four questions, asked of 30+ target buyers:

• At what price would you consider this product so expensive you would not buy it?
• At what price would you consider this product expensive but still consider buying it?
• At what price would you consider this product a bargain, a great buy for the money?
• At what price would you consider this product so cheap that you would question its quality?

Plot the four cumulative distributions. The intersections give you a Range of Acceptable Prices (where the "too cheap" and "too expensive" curves cross), an Optimum Price Point, and an Indifference Price Point. The Range of Acceptable Prices is the only output that matters. It is the band inside which the market will buy without rejecting on price.

Worked example. A B2B productivity tool surveys 40 target users. Range of Acceptable Prices comes out as $25–$95/month. Optimum Price Point falls at $48. The founder was about to launch at $19, anchored on the cheapest competitor. The data says $19 is below the lower bound of acceptability: buyers question quality below $25. Launch price moves to $39 (inside the range, below optimum, leaves room for the second-tier $69 plan). That single decision routes a meaningful percentage of recoverable margin to the bottom line for the life of the product.

Position inside the range determines elastic vs inelastic strategy. Closer to the lower bound favors growth via volume, lower margin per customer, larger total addressable buyers. Closer to the upper bound favors expansion and tier upgrades, higher margin per customer, narrower addressable buyers. The right one depends on whether retention and expansion mechanics favor depth or breadth. The SaaS Pricing Strategy Calculator takes COGS and target margin as inputs and returns a price floor; the Price Elasticity Calculator answers whether a price change increases or decreases revenue at your current elasticity.

5. Payback period

CAC payback in months tells you how long until a customer pays back their acquisition cost out of gross profit. The formula:

CAC Payback (months) = CAC / (ARPU × Gross Margin)

OpenView 2024 data shows median payback of 18 to 24 months for mid-market B2B SaaS, 12 to 18 for SMB^[1]. ChartMogul's 2024 cohort data shows median SMB SaaS payback moved from 14 to 16 months between 2022 and 2024^[4]. For a bootstrapped solo founder, anything over 12 months is a cash-flow problem regardless of how the LTV:CAC ratio looks on paper.

Worked example using earlier numbers. ARPU = $50, gross margin 82%, CAC = $432. Payback = $432 / ($50 × 0.82) = $432 / $41 = 10.5 months. A 10.5-month payback is acceptable for a venture-backed firm and tight for a bootstrapped one. Cut CAC to $300 by killing the worst-performing channel, payback drops to 7.3 months: a cash-flow profile that lets you reinvest acquisition spend twice in the same year.

Every other lever moves payback. Raise ARPU 10%, payback drops 9%. Cut COGS 10% (gross margin goes from 82% to 84%), payback drops 2%. Cut CAC 10%, payback drops 10%. The math collapses every input into months-to-recovery, which is the metric that maps directly onto runway and reinvestment cadence. The CAC Payback Calculator handles the arithmetic with a sanity-check against LTV:CAC.

Putting it together: a worked example

One product, all five numbers, one pricing decision. A solo founder building an AI meeting-notes SaaS, 6 months pre-launch, 35-person beta cohort.

Step 1: COGS-per-user (projected at 250 paying users)
  Hosting + DB           $0.40/user
  OpenAI transcription   $4.20/user (audio length × token rate)
  Vector DB + storage    $0.30/user
  Stripe fees (3%)       depends on price, computed below
  Support (8 min/user)   $1.50/user at $75/hr loaded
  Subtotal pre-Stripe    $6.40/user

Step 2: CAC (planned)
  Paid + tooling         $400/month
  Founder time (20 hr)   $1,500/month
  Expected new paid/mo   8
  CAC                    $237/customer (90-day blended)

Step 3: Runway sensitivity
  Cash on hand           $40,000
  Burn at 250 users      $4,200/month
  Runway                 9.5 months
  20% COGS shock at 250  +$320/month → 8.8 months runway
  20% COGS shock at 1k   +$1,280/month → 7.0 months runway

Step 4: Willingness-to-pay (Van Westendorp, n=35 beta)
  Range of acceptable    $19 – $79/month
  Optimum Price Point    $34/month
  Indifference Point     $42/month

Step 5: Payback at three candidate prices
  Price $24:  COGS $7.12, GM 70%, payback = $237 / ($24 × 0.70) = 14.1 months
  Price $34:  COGS $7.42, GM 78%, payback = $237 / ($34 × 0.78) =  8.9 months
  Price $49:  COGS $7.87, GM 84%, payback = $237 / ($49 × 0.84) =  5.8 months

What the math says. $24 is inside the range but payback is too long for bootstrapped runway given the 9.5-month cushion. $34 lands at the optimum price point with sub-9-month payback, the safe default. $49 sits at the high end of the acceptable range, delivers a 5.8-month payback, but risks elasticity loss in conversion rate that the n=35 survey is too small to estimate confidently.

Decision: launch at $34/month, add a $69 second tier with volume features, revisit pricing at the 200-paying-user mark with a larger willingness-to-pay sample. The Break-Even Units Calculator closes the loop by telling you how many paying customers at $34 cover fixed costs at the chosen burn rate.

Common founder mistakes

• Pricing from competitor screenshots. Competitor pricing tells you what someone else charges, not what your buyer will pay you. Buyers anchor to category norms but pay based on perceived value of the specific product. Use competitive data as a sanity check on the upper bound, not as the primary input.
• Single-tier launch with no elasticity test. Launching at one price makes elasticity invisible. A two-tier launch (e.g. $24 / $49) gives early conversion-rate signal across the band. Three tiers is overkill pre-product-market-fit.
• Underestimating support COGS at scale. A founder who handles 100% of support at 50 users drowns at 500 users. Price in the support cost from day one, or deflect with self-serve onboarding before user count makes it impossible.
• Ignoring Stripe fees in the COGS stack. 3% on a $30 product is $0.90, which matters on a tight gross margin. International cards, currency conversion, and chargebacks each add 0.5 to 1.5%, relevant for cross-border SaaS.
• CLV anchoring on aggregate churn. Aggregate churn double-counts early-cohort drop-off and inflates LTV. Cohort-weighted churn produces an LTV roughly 30% lower in most SaaS data. Full breakdown in the Customer Lifetime Value article.
• Skipping the willingness-to-pay survey because it feels unscientific. A 30-person Van Westendorp survey on warm-list users produces a defensible price range in a week. The alternative is no data at all.
• Treating the launch price as permanent. The first price is a hypothesis. Re-run the five numbers every 6 months until the product reaches steady-state retention, then annually. Most pricing failures are good initial decisions that were never revisited.

The five numbers compress months of pricing analysis into a single afternoon of arithmetic. Run them before launch, re-run at 100 paying users, re-run at 500. Each pass refines the band you can defensibly price inside. Skipping the math postpones it to the moment you have to cut prices in front of churned customers, which is the most expensive time to do pricing analysis.

The five tools that automate the arithmetic: SaaS Pricing Strategy Calculator for the price floor, CAC Payback Calculator for months-to-recovery, Profit Margin Calculator for gross-margin sanity, CLV Calculator for the LTV:CAC ratio, and Break-Even Units Calculator for the customer count that covers fixed cost.