CPA calculator (cost per acquisition and CAC)

1. Pick what to work out. CPA or CAC takes cost and conversions. Conversions tells you how many a budget should bring at a given CPA. Budget needed tells you what a conversion goal will cost.
2. For CPA, enter ad spend and the conversions your ads manager reports for the same campaigns and dates.
3. For CAC, enter all the costs of winning customers in the period (ad spend plus agency fees, creative production, influencer costs and any sales costs) and divide by new customers only, not repeat orders.
4. Add order value and margin (optional) to see ROAS on the first order, your break-even CPA and whether each new order makes or loses money.

CPA = ad spend ÷ conversions

CPA example. A store spends $3,000 on ads in a month and the ads bring 75 purchases. CPA = $3,000 ÷ 75 = $40.

CAC = total acquisition costs ÷ new customers

CAC example. The same month, the store also pays a $1,000 agency fee and $500 for creative. Of the 75 orders, 60 came from new customers. CAC = ($3,000 + $1,000 + $500) ÷ 60 = $75. The CAC is higher because it counts every cost and only new customers.

Planning with CPA

• Budget needed = conversions × CPA. 120 orders at a $40 CPA need a $4,800 budget.
• Conversions = budget ÷ CPA. $2,000 at a $50 CPA should bring about 40 conversions.

Break-even CPA

Break-even CPA = average order value × margin

With a $90 average order and a 50% margin, each order leaves $45 before ads, so a CPA above $45 loses money on the first order. Use the break-even ROAS calculator to include shipping, fees and returns.

A good CPA is below what a customer is worth to you, so there is no single number that fits every store. A $40 CPA is excellent for a $300 jacket and ruinous for a $25 T-shirt. Start from your own economics:

• First-order view. CPA should sit below your break-even CPA (order value × margin), and below it by enough to leave a profit.
• Lifetime view. If customers reliably buy again, you can pay more to win them. Only count repeat purchases you can see in your own data.
• The conversion you count. A sign-up or add-to-cart is cheaper than a purchase and worth less. Compare like with like.

Published benchmarks mostly cover leads, not purchases. The median cost per lead across US Google search campaigns, for example, is about $66.69. That is a useful sense of scale for lead generation, but an ecommerce CPA depends mainly on your price and margin.

What moves CPA: click-through rate and CPM (the cost of getting visitors), conversion rate on your site, audience and season, the offer, and whether the campaign has enough conversions to leave the learning phase.

CPA is the cost of a visitor divided by the share of visitors who convert, so there are two sides to work on:

• Cheaper visitors. Stronger creative raises click-through rate, which lowers cost per click. See the CTR calculator.
• More visitors who buy. A faster page, clear product photos, visible shipping costs and a short checkout raise conversion rate. See the conversion rate calculator.
• Give the algorithm enough data. On Meta, an ad set usually leaves the learning phase after about 50 results in the week after its last significant edit. A budget too small to reach that, or frequent edits, keeps costs unstable.
• Cut what does not convert once it has spent enough to judge, for example one and a half times your target CPA with no conversions.
• Check tracking. Purchases that do not reach the ad platform make CPA look worse than it is and train the algorithm on bad data.

• A ceiling from your margins. Ad Autopilot sets the target CPA for a sales campaign as average order value ÷ target ROAS, calculated in code from the order value and gross margin you enter. For a lead or sign-up it uses the value you give per result, or states an assumed value, so a lead is never allowed a whole order's budget.
• A ceiling, not a forecast. The plan keeps the most you can afford apart from what the market is likely to charge. The expected cost comes from your own account's history on the same offer (at least 10 results) or a grounded benchmark, and when neither exists the plan says so.
• Budget checks. It compares your daily budget with 50 ÷ 7 × target CPA, the spend needed for about 50 conversions a week, and flags a budget too small to learn.
• Losses need your approval to stop. An ad needs at least 2,000 impressions before the agent acts on it. If a campaign's CPA stays at twice your maximum or more for 3 settled days in a row, the agent can propose stopping it, and that proposal always waits for your approval.