The formula, and why it works

The logic runs in one line. Your ad cost per conversion at break-even must equal your net value per conversion, and since cost per conversion is CPC divided by conversion rate, the formula rearranges to: break-even CPC equals CVR times net value per conversion. A 2 percent conversion rate on clicks worth $50 each in net value means you can pay up to $1.00 per click before the traffic loses money. Two inputs, one multiplication, and every bid in the account suddenly has a reference line.

The formula’s honesty depends entirely on what goes into “net value per conversion,” and this is where most versions of the calculation flatter themselves. Net value means profit, not topline. Start with revenue per conversion, multiply by real gross margin, then subtract every variable cost attached to that conversion, fulfillment, payment fees, commissions, onboarding, and expected refunds. For lead gen and SaaS, the conversion in the formula is the monetization event, the closed deal or paid subscriber, not the form fill, which means the CVR in the formula is the full click-to-sale rate, and the value is contribution margin within your payback window.

Getting the inputs honest

Map the money before doing the math. Identify the revenue event, its average value including recurring and upsell components you’re willing to count, and convert topline to gross profit using the margin your books actually show rather than the one in the pitch deck. Then list the variable costs that ride along with each conversion. For ecommerce that’s COGS, shipping and 3PL, packaging, processor fees, platform fees, discounts, and expected refunds with their return shipping. For SaaS it’s hosting, per-seat support, onboarding, and commissions. The inclusion test is simple: if the cost scales with volume, it belongs in the calculation, and leaving it out doesn’t make the traffic profitable, it makes the spreadsheet lie.

The time-horizon decision deserves its own paragraph because it changes the answer more than any other input. If the business needs payback within 30 or 90 days, use contribution within that window, not lifetime value, since LTV-justified bids with twelve-month payback are how growth-stage brands run out of cash while technically profitable. If longer payback is genuinely affordable, use expected LTV minus variable servicing costs, but write the payback tolerance down as a decision someone owns rather than an assumption nobody made. This is the same margin-first discipline behind optimizing for profit instead of ROAS, applied one level down at the click.

The math, worked twice

Ecommerce first. AOV of $120 at 55 percent gross margin gives $66 of gross profit. Subtract $10 shipping, $3 payment fees, and refunds at 5 percent of orders costing their lost margin, about $3.30, and net value per conversion lands at $49.70. At a sitewide CVR of 2.5 percent, break-even CPC is 0.025 times 49.70, roughly $1.24. Every click above that price loses money on this funnel, whatever the platform’s ROAS column says about it.

Lead gen second, where the chain is longer. A 12 percent click-to-lead rate times a 20 percent lead-to-close rate gives an overall click-to-sale CVR of 2.4 percent. With net profit per closed sale of $600 after variable costs, break-even CPC is 0.024 times 600, or $14.40. And since break-even means zero profit, the working bid sits below it. Requiring a 30 percent margin on ad spend caps the working CPC at 70 percent of break-even, $10.08, which is the number that actually goes into the bid strategy.

Audit the inputs, stress-test the edges, then scale

The formula is only as good as its two inputs, and both come from tracking that deserves suspicion before trust. Deduplicate conversion events, align attribution windows across platforms, account for the iOS and browser signal loss with server-side or CRM-based tracking, and reconcile platform-reported conversions against backend orders, because a CVR inflated by double-counting produces a break-even CPC that grants permission to overpay, the exact failure we detailed in how to know if you’re overpaying for every click.

Then stress-test, because the inputs are estimates and estimates move. Build best, base, and worst scenarios by flexing CVR, margin, AOV, refunds, and close rates by 20 to 30 percent, and watch what happens to the break-even line. A funnel where a modest CVR dip flips the economics negative needs a bigger bidding buffer than one that stays profitable across the range, which is why the working guardrail is bidding 70 to 85 percent of break-even, with the gap sized to your input volatility, and raises granted only when cohort data confirms the margins held.

Finally, compute it where the decisions actually happen. Break-even CPC differs by device, geo, audience, and funnel stage because CVR and value differ there, so a single blended number sets caps too high in your weak segments and too low in your strong ones. Segment the calculation, set bid caps accordingly, watch cohort economics (refunds, churn, upsell) rather than day-one revenue, and update the inputs on a schedule, because a rising CVR or improving margin raises the line, and a raised line is quantified permission to scale, the trigger logic we laid out in knowing when to increase or pause ad spend. Bidding against a number you derived, audited, and stress-tested isn’t conviction as a mood. It’s conviction as a calculation, which is the only kind that survives contact with the auction.