What to Measure Before Buying Promotion

A founder should not buy promotion until three figures are on one sheet: cost per acquisition, payback period and contribution per sale. If cost per acquisition is above contribution per sale, the campaign loses money on every booking. If payback runs past the cash the company holds, the campaign is a loan the founder cannot repay.

What does cost per acquisition measure?

Cost per acquisition is total campaign spend divided by the number of sales the campaign produced. Spend covers the mailer, the list, the print, the postage, the phone line and the labor to work the replies. Sales are counted only when money is collected, not when a lead raises a hand.

A worked example. A test cell of 5,000 names costs $2,400 to build and mail. The cell returns 40 bookings at an average of $310. Revenue is $12,400. Cost per acquisition is $2,400 divided by 40, or $60. Contribution per booking, after the cost of delivering the week, is $95. The campaign pays.

Change one number and the answer changes. If the cell returns 20 bookings, cost per acquisition is $120 and the campaign loses $25 on every sale. The mailer, the list and the segment are the same. Only the response rate moved.

Cost per acquisition is read from the campaign record for the period the campaign ran. It is not read from a forecast. A founder who has never run a paid campaign has no cost per acquisition, only an estimate, and the estimate should be labeled as one.

Why payback period decides whether a founder can afford the campaign

Payback period is the number of months it takes for the contribution from a customer to return the money spent to acquire that customer. A campaign with a $60 cost per acquisition and $95 contribution per booking pays back inside the first month. A campaign with a $300 cost per acquisition and $50 contribution per month pays back in six months.

Six months is not a problem if the company holds twelve months of cash. It is a problem if the company holds three. The campaign is then funded by money that was reserved for payroll, rent or the next build.

Founders who are still choosing between debt, equity and grants can read the trade-offs at startup financing guides before committing cash to a first campaign. The choice of capital sets how long the payback window can be. A grant or a priced round buys time. A short-term loan does not.

Run the arithmetic in this order. First, contribution per sale. Second, cost per acquisition from the test cell. Third, months of payback. Fourth, months of cash on hand. If the fourth number is smaller than the third, the campaign waits.

How is the break-even response rate calculated?

Break-even response rate is the response rate at which campaign revenue equals campaign cost. It is the floor, not the target.

Take fixed campaign cost of $2,400. Take contribution per sale of $95. Break-even sales are $2,400 divided by $95, or 26 sales. On a cell of 5,000 names, break-even response is 26 divided by 5,000, or 0.52 percent.

A founder who does not know this number cannot read a test result. A cell that returns 0.4 percent looks like a failure only if the floor is known to be 0.52 percent. A cell that returns 0.6 percent looks like a win only against the same floor.

Break-even is read from the same campaign record as cost per acquisition, over the same period. It moves when postage moves, when the list price moves and when the average sale moves.

What should be tested before the full rollout?

One variable per test cell. The list, the offer, the mailer format and the reply channel are four variables. Testing all four at once produces a result no one can read.

The protocol is announced before the results. A founder writes down the cell size, the variable under test, the metric that settles the question and the date the result is read. The sheet is dated. Nothing is added after the replies arrive.

A first test is usually the list. Two segments of the same size, the same mailer, the same offer, the same reply channel. The segment with the lower cost per acquisition wins, and the difference is the finding.

A second test is the offer. Same list, same mailer, two expiry dates or two premiums. The offer with the higher contribution per sale wins, even if its response rate is lower.

Which figures belong on the sheet before money is spent?

Six figures. Campaign cost. Expected response rate from the closest comparable record. Average sale. Contribution per sale. Cost per acquisition. Months of payback.

Two more figures belong beside them. Cash on hand in months. Break-even response rate. A sheet with these eight numbers can be argued with. A sheet without them cannot.

Each figure is named with the record it was read from and the period it covers. A response rate from a 2023 mailing to a rented list is not the same figure as a response rate from a 2024 mailing to the house list. The record says which.

Where a rule differs by jurisdiction, the sheet names the jurisdiction. Postage rates, toll-free windows and consumer consent rules for outbound calls are set at the federal level in the United States and, in some cases, at the state level as well. A campaign that runs in two states runs under two records.

What a founder should not pay for

A founder should not pay for promotion before the contribution per sale is known. Without it, cost per acquisition has no ceiling and payback has no floor.

A founder should not pay for a full rollout before a test cell has run. The test cell is the cheapest part of the campaign and the only part that produces a number.

A founder should not pay for a list that has not been sampled. A sample of 500 names at the same offer and the same mailer gives a first read on response at a fraction of the cost.

A founder should not treat a lead as a sale. A lead is a reply. A sale is a collected payment. Cost per acquisition is built on the second, not the first.

The arithmetic, in one place

Contribution per sale equals average sale minus the cost of delivering the product. Cost per acquisition equals campaign cost divided by sales. Payback in months equals cost per acquisition divided by monthly contribution. Break-even sales equal campaign cost divided by contribution per sale. Break-even response equals break-even sales divided by cell size.

Five lines. A founder who can fill them in has the answer to whether the campaign should be bought. A founder who cannot fill them in has a proposal, not a plan.

The bench writes what a campaign costs and what it returns. The figures above are the ones that decide it.