How Much Should I Spend on Ads? The calculator does the math.

The single most expensive mistake in small-business advertising is valuing a customer at their first transaction. A $300 med spa visit looks like a $300 customer, so when ads cost $325 per new patient, the owner panics and shuts everything down.

But on the calculator's med spa starting values that patient comes back four times a year, for about 1.4 years. That is $1,680 of revenue, and at a 58 percent gross margin, about $974 of gross profit. Paying $325 to acquire that patient is a trade worth taking every day. This number, everything an average customer pays you across the whole relationship, is Customer Lifetime Value (LTV). The calculator then multiplies it by gross margin, because you can only spend profit, and that is the foundation every other number is built on. Six steps, the same six the calculator runs:

Return on Ad Spend (ROAS) is the simplest health check in advertising: how many dollars come back for every dollar you put in. Spend $10,000, generate $30,000, that is a 3× ROAS.

The calculator's ROAS section shows your break-even ROAS on the first sale (1 divided by your gross margin, so 2.0× at a 50 percent margin and 1.72× at the calculator's med spa starting margin of 58 percent), the ROAS the platforms will report at your ceiling, on the first sale and over the customer's lifetime, and a 3× target you can make the ceiling for your own numbers.

Two numbers matter, and neither is a universal standard:

• Break-even ROAS on the first sale, 1 divided by gross margin. At 58 percent margin that is 1.72×, at 50 percent it is 2×, at 25 percent it is 4×. Below it you lose money on every first sale.
• A target. The calculator defaults to 3× on first-sale revenue. That is the starting point PELORA uses, not a published standard, and you can set your own.

The trap almost everyone falls into: calculating ROAS on first-transaction revenue alone. Measure that $300 first visit against a $325 acquisition cost and the campaign looks like a loss, 0.9×. Over the patient's lifetime it returns $1,680 on $325, about 5× on revenue and 3× on gross profit, exactly the 3 to 1 bar from Step 3. Look at both, or you will kill profitable campaigns and never know what you lost.

The businesses that win at paid advertising are not the ones with the biggest budgets. They are the ones who know their numbers, and therefore stay in the game long enough for compounding to work.

This is the part nobody explains, and it is responsible for more wasted ad spend than any targeting mistake ever made.

When you launch a campaign, Meta's algorithm enters a learning phase: it tests audiences, placements, and creative combinations to figure out who your buyers actually are. During this period results are unreliable, costs run high, and delivery is inconsistent. That is not failure, that is the system working as designed.

The 50-conversion rule

Meta says an ad set leaves the learning phase after about 50 results in the week after its last significant edit (Meta Help Center). Google says Target CPA can start with no conversion history, and recommends at least 30 conversions in the last 30 days before you judge it (Google Ads Help, Target CPA). Below those volumes, the algorithm is guessing, not optimizing.

Here is the catch for service businesses: if your conversion event is "became a patient," you may never hit 50 a week. The fix is to track an earlier, higher-volume funnel action as your conversion, a form fill, a click-to-call, a booking page view. More weekly signal means the algorithm learns faster, exits learning sooner, and your cost per result drops. The calculator above checks this automatically: enter your monthly client goal and it tells you whether you clear the threshold or need an earlier tracking event.

When to judge, in each platform's own words

Judging ads in week one is like judging a gym membership after one workout. The owners who lose money on ads are almost never the ones with bad products, they are the ones who quit during the learning phase, right before the machine started working.