The numbers

Every planning estimate says how it was setCopy Anchor Link

A planning estimate is a number we set so the math can run before you have your own. Each one below names how it was set and the reading that replaces it. Tables across this site print the label once, in their footer, and link here.

Third-party figures are not planning estimates: each prints its sample and year, such as search ads at $5.26 a click (industry average, 16,000 advertisers, 2024 to 2025) or LinkedIn acceptance across 13.2 million requests.

One rule sets what a customer may costCopy Anchor Link

A customer may cost the lower of two numbers. The first is the margin it earns in its first 12 months after churn: m × (1 − (1 − c)^12) ÷ c, where m is monthly margin and c is monthly churn. The second is a third of its lifetime value: m ÷ c ÷ 3. Every table on this site reads this rule, and a loaded CAC (cash plus your hours at $100) is held to it.

At 5% churn a year earns 9.19 months of margin, not 12, and a third of lifetime value is 6.67 months. So the lifetime cap binds: a $49 product may spend $261 a customer, though $360 would pay back in a year. Over $5,000 a year the deal is a 12-month term, nobody leaves in year one, and the full 12 months of margin binds.

Cold email costs $71 cash and 8.5 of your hours a customer, $921 loaded. That clears the rule only from $173 a month. Below it, Hands-on buys learning, not revenue, until a rewritten offer lifts replies: at 6%, loaded CAC falls to $739. Paid search at the average click costs $1,753 self-serve and clears from $329 a month; at $299 it pays back in 8.9 months after churn, not the 7.3 months that leaving churn out suggests.

Planning estimate unless a source is named: how each was set.

Computed by one rule: the lower of a customer's first 12 months of margin after churn and a third of its lifetime value, at 80% margin and 5% monthly churn, or a 12-month term over $5,000 a year. Each row's break-even click uses that row's share of visits that pay (how each price converts). Hands-on is cold email at $921 a customer with your hours at $100. Planning estimate unless a source is named: how each was set.

Each price converts its own way, so each has its own break-even clickCopy Anchor Link

A click is worth buying only when the customer it brings pays the click back under the rule. The line is max cost per click = (CAC you can afford − your hours a won deal × $100) × visit to paid, and visit to paid depends on how the price sells.

• Under $300 a year, self-serve: visit to signup × signup to paid = 2.0% × 15% = 0.3%. At $19: $101 × 0.3% = $0.30.
• $300 to $5,000 a year, a small team: visit to trial or demo × trial or demo to paid = 2.5% × 20% = 0.5% (0.4% to 0.8%). At $49: $261 × 0.5% = $1.31. At $299: $1,595 × 0.5% = $7.97.
• Over $5,000 a year, a company deal: demo request × qualified × won = 1.0% × 50% × 25% = 0.125%, with 6 of your hours a won deal. At $1,000 a month: ($9,600 − $600) × 0.125% = $11.25, and $9.00 to $13.50 as the win rate runs 20% to 30%.

Search ads average $5.26 a click (industry average, 16,000 advertisers, 2024 to 2025). That buys a self-serve customer for $1,753, a small-team customer for $1,052 and a company deal for $4,208 plus your hours. So a $49 product earns its clicks through Launch and search; a $299 product can buy them through Paid on terms under $7.97. Keyword opportunity map prices every term against this line before a dollar moves.

Planning estimate unless a source is named: how each was set.

Payback counts your hours and your churnCopy Anchor Link

payback months = the n where m × (1 − (1 − c)^n) ÷ c reaches the loaded CAC; it never comes when the loaded CAC passes m ÷ c, a whole lifetime of margin.

Loaded CAC is cash plus your hours at $100. Hours that stay fixed as customers grow (a page a week) are spread over the customers a month at scale; hours that grow with customers (an hour a call) are counted per customer. A launch looks free in cash, but 6 hours a week for 2 customers a month is 13 hours, $1,301 a customer, close to paid search.

At $49 a month, 4 of the 11 ways clear the rule on loaded CAC: SEO and content, AI answers, Directories and backlinks and Referrals. The last column says the lowest price at which each one does.

Planning estimate unless a source is named: how each was set.

At scale, each way runs into its own ceilingCopy Anchor Link

A way that works at 400 contacts or 1,000 visits stops growing somewhere. The table says how many customers a month each way brings at scale, what that takes, and what caps it.

Pick the corner whose month-one cash plus hours fit inside what you can spend before month 6, then plan the second way before the first one hits its ceiling.

Planning estimate unless a source is named: how each was set.

One rule stops, reads and scales a paid testCopy Anchor Link

• Stop at 300 clicks with zero signups. By the rule of three, zero in 300 means the rate is likely under 3 ÷ 300 = 1%, the visit-to-signup line. A true 1% gives zero that often only 0.049 of the time.
• Read at 500 clicks, as an interval: stop when CAC at the optimistic bound fails the rule, hold and rewrite when only the pessimistic bound fails.
• Scale only on paid customers you have seen: 30 or more signups and at least 3 of them paid, or one full trial cycle read. Until then spend holds at the test tier, however good the signup rate looks.

Paid, Message test and Weekly readout apply this rule as written; the agent sets spend to zero at the stop line without waiting for you.

Read a rate only once you have enough of itCopy Anchor Link

A rate from a small sample is a range, not a number. 10 signups from 500 clicks reads 2%, and the 95% interval runs 1.1% to 3.6%: a 3.3 times spread in CAC. So the test rule stops on the optimistic bound and scales on the pessimistic one.

Comparing two headlines on clicks at a 1% click rate takes about 7,700 impressions each to see a 50% lift. A 30% lift takes about 19,800. Two pages at 2% against 3% signup need about 3,800 visits each, which no founder buys on search. Test the message on clicks with Message test, and read signups pooled.

An answer share of 10% read from 200 answers moves ±4.2 points by chance. AI answer audit runs each question 3 times and flags only moves of 5 points or more. Cold email is judged at 300 contacts: 1% of 300 is 3 positive replies, enough to tell 1 per 100 from 4.

Churn compounds, so it caps what a customer may costCopy Anchor Link

Monthly logo churn c leaves (1 − c)^12 of your customers after a year. At 5% a month you keep 54%. At 3% you keep 69%. At 10% you keep 28%, which no acquisition rescues.

Lifetime value is about monthly margin ÷ monthly churn. A $49 product at 5% churn is worth $784; a third of that, $261, is the most it may cost. At 3% churn the cap rises to $436 and the year's margin, $400, binds instead.

That is why Keep and grow starts once the first 10 pay, and why Silence watch scores every customer daily.

Price your own hours, or the cheap ways look freeCopy Anchor Link

An hour of yours is $100 here. Replace it with what an hour of building is worth to you.

Priced, a cold-email customer costs $71 cash and 8.5 hours: $921. Its month is 400 contacts, 2.9 calls and 0.6 customers, so the calls are most of the hours. At a 10% close rate it takes about 1,400 contacts a customer and the loaded CAC doubles to $1,843.

A launch week of 15 to 20 hours, about 26 across the month, that brings 2 customers costs about $1,300 each in hours. Neither is free; each is cheap in cash. Your agents take the list building, the drafting and the reading; your hours go to calls, thread answers and launch day. Weekly readout prints your minutes beside each Play, and the demand budget holds the hours you said you can spare.

Planning estimate unless a source is named: how each was set.