How to Set Your Target CPA and Max Bid From LTV

I have watched a lot of beginners pick a bid the way you pick a lottery number. They type in something that feels reasonable, launch, and then wonder why the account either spends nothing or loses money fast. The number you bid is not a vibe. It is a calculation, and once you see the math you stop guessing.

This piece is about that math. Not what CPM, CPC, and CPA mean as bidding models, but how you actually choose your target CPA and your maximum bid starting from what a customer is worth. I will give you the formulas I use, show how conversion rate and fees change the answer, and work a full example in dollars.

The one rule everything else hangs on

Here is the whole game in one sentence. You cannot pay more to acquire a customer than that customer will return to you, or the buy loses money. Obvious written down, but almost every blown budget I have cleaned up violated it in some quiet way.

So before you touch a bid, you need a number for what a customer is worth. The simplest estimate of lifetime value is LTV = ARPU x Lifetime, where ARPU is average revenue per user in a period and Lifetime is how many of those periods a user sticks around. A subscriber who pays 10 dollars a month and stays 8 months has an LTV of 80 dollars. That 80 is the ceiling. Everything below it is where the decisions live.

Keep one thing straight: LTV should be revenue you actually keep. For e-commerce, use gross margin per order, not the sticker price. Paying 40 dollars to acquire a customer who buys 80 dollars of product at 30 percent margin means you earned 24 and spent 40. Always work from money you keep.

From target ROAS or payback to a target CPA

Your target CPA is not the full LTV. If you spent your entire LTV to acquire someone, you would break even at best, with nothing left for product, team, or profit. So you decide up front how much return you need on ad spend, and that sets the cap. There are two common ways to express it, and they are the same idea in different clothes.

• Target ROAS. Want 2 dollars back for every 1 dollar of ad spend over the customer lifetime? Your target lifetime ROAS is 2, and your max acquisition cost is LTV divided by ROAS. With an 80 dollar LTV and a target ROAS of 2, your target CPA is 40 dollars.
• Payback window. Prefer to think in time? Decide how fast you want spend returned. Want it back in the first 3 months on a 10 dollar per month subscriber? That caps your target CPA at about 30 dollars, since 3 months of revenue is 30. The rest of the lifetime is profit.

Pick the frame that matches how your business thinks. Cash-tight operations lean on payback because they cannot wait a year to recover spend. Businesses with patient capital lean on lifetime ROAS. Either way you land on one dollar figure: the most you will pay for a customer. That is your target CPA.

From target CPA to a max bid

Now the platform. On most US channels you can hand the system your target CPA directly through target CPA bidding or value-based bidding and let it optimize. But you still need the mechanics underneath, because that is how you sanity-check the algorithm and set manual caps when you need them.

The bridge between a target CPA and a bid at the click or impression level is your conversion rate. The relationship I keep on a sticky note: CPI = CPC / CR, which rearranges to max CPC = target CPA x CR. CPI is cost per install or acquisition, CPC is cost per click, and CR is the click-to-conversion rate.

In plain English: if only a fraction of clicks convert, each click can only cost a fraction of your target CPA, because you pay for the clicks that did not convert too. A 5 percent conversion rate means 20 clicks per conversion, so each click stays under one twentieth of target CPA. Lower conversion rate, lower safe bid. This is why creative and landing pages matter: they move CR, and CR moves the bid you can afford.

A full worked example in dollars

Let me run one end to end so the steps are concrete. Say I am acquiring subscribers for a US fitness app.

• ARPU: 12 dollars per month.
• Lifetime: 7 months average before churn.
• LTV: 12 x 7 = 84 dollars.
• Target lifetime ROAS: 2 (I want to double ad spend).
• Click-to-install conversion rate: 4 percent.

Step one, target CPA. LTV divided by target ROAS is 84 / 2 = 42 dollars, so I pay at most 42 to acquire a subscriber. Step two, fees. App stores take a cut and payment processing shaves a bit more. If the store keeps 30 percent of that 12 dollar payment, my real ARPU is about 8.40, my real LTV is about 59, and my honest target CPA at 2x ROAS drops to roughly 29 dollars. Skipping this step is the most common way I see people overpay.

Step three, the max bid. With a 4 percent conversion rate and a 29 dollar target CPA, max CPC = 29 x 0.04 = 1.16 dollars per click. So I cap manual CPC near 1.16, or hand the platform a 29 dollar target CPA and let value-based bidding find clicks under it. If my real conversion rate turns out to be 2 percent, the same math gives 0.58 per click, so I would need cheaper clicks or better creative to survive.

Where the number moves after launch

Your first bid is a hypothesis, not a verdict. Once data comes in, the same formulas tell you what to change. If conversion rate climbs because you tightened the landing page, your affordable bid climbs with it and you can compete for better inventory. If churn is worse than you assumed, your real LTV falls and your target CPA has to come down before you keep spending.

Audience value matters too. Higher-value segments cost more because everyone bids for them, so bids rise with the value of who you chase. A few targetings that lift realistic LTV in the US: adults 25 and up who tend to be employed, Tier 1 geographies, large metros with higher wages, newer devices that signal disposable income, and people who have already paid on a platform before. A higher CPA for those users can still be profitable if their LTV justifies it.

One habit keeps you honest: recompute your target CPA whenever a real input changes, rather than defending the number you launched with. The bid is downstream of LTV, ROAS, and conversion rate. Update the inputs, and let the output tell you what to do.