Google will ask you for a number.
Somewhere in setting up automated bidding, a field appears asking what you want to pay per conversion. Most people put in something that feels reasonable. A round number. Something a bit below what they are paying now, because lower is better.
That number is the single most consequential setting in the account. It decides how much of the market you can reach, which auctions you can win, and whether the campaign scales or strangles. And it is almost always a guess.
The number is knowable. It comes out of your own business, not out of Google, and working it out takes about twenty minutes.
An Investment or an Expense. You Pick Which One.
The framing matters before the arithmetic does.
Advertising is an expense when you spend a fixed amount because you feel you should, and you judge it by whether the invoice feels affordable. Advertising is an investment when you know what a customer is worth, you know what you paid to get one, and the gap between those two numbers is the return.
The second version is not more sophisticated. It is just the version where you can answer the question "should I spend more next month?" without guessing.
Everything below is in service of one number: the most you can pay to acquire a customer and still be glad you did. Once you have it, target CPA is not a guess anymore. It is that number, minus the margin you want to keep.
The Four Numbers
You need four things, and you almost certainly have all of them already.
- Average sale value. What a typical first purchase or job is worth. Found in your invoicing or POS over the last 6 to 12 months.
- Gross margin. What is left after the cost of delivering it. Revenue minus direct costs, as a percentage.
- Close rate. The share of leads that become customers. Leads in versus jobs won, over the same period.
- Repeat value. What the same customer spends after the first sale. Your repeat purchase rate, or average contract length.
The first two are usually easy. The third is where most businesses stall, and it is the one that changes the answer most.
Average sale value should be a median rather than a mean if you have a handful of unusually large jobs, because one outlier will flatter the whole calculation. Use the typical job, not the best one.
Gross margin means after the cost of actually delivering the work: materials, the labor hours, the merchant fees. Not after rent and salaries. You are working out what each additional customer contributes, and overhead is there whether you win them or not.
Close rate is the number nobody has to hand, and estimating it is where the whole calculation goes wrong. If you cannot count it properly, count it roughly: how many enquiries did you get last month, and how many turned into paying work? Two numbers off the top of your head beat no number at all, and you can improve the measurement later. What you cannot do is skip it, because it is a multiplier.
Repeat value only counts if you can point at evidence. If you genuinely do not know whether customers come back, treat the first purchase as the whole value and revisit when you do know.
Work It Backwards
The arithmetic runs in one direction: from what a customer is worth, back to what a lead can cost.
Take a service business with a $2,000 average job at 50% gross margin. Every customer contributes $1,000 before overhead.
Say you are willing to spend a third of that contribution to acquire the customer. That is $333 you can pay per customer.
Now the step everyone skips. Google does not deliver customers. It delivers leads, and only some of them buy. If you close one lead in four, four leads cost you one customer, so your maximum cost per lead is $333 divided by 4, which is about $83.
That is your ceiling. A target CPA of $83 is break-even against the split you chose. You would set it somewhere below that to leave room.
Here is the same calculation at three different close rates, holding everything else constant:
- Closing 50% of leads. 2 leads per customer, so you can pay up to $167 per lead.
- Closing 25% of leads. 4 leads per customer, so you can pay up to $83 per lead.
- Closing 10% of leads. 10 leads per customer, so you can pay up to $33 per lead.
Same business, same margin, same appetite. The affordable cost per lead moves by 5x purely on how good you are at closing.
This is why two competitors in the same market can both be right about wildly different bids. The one who closes half their leads can afford to pay double, and will win the auction every time.
The Number Changes If They Come Back
If your customers buy once and never return, the number above is the whole story. Most businesses are not like that.
A chiropractor treating a prenatal patient is not looking at one visit. That patient comes back through the pregnancy and often after it. A coffee equipment supplier is not selling one machine, they are opening an account that reorders. Lifetime value is the difference between those and a single transaction.
Bring it in the same way. If the average customer is worth 3x their first purchase over the relationship, the $1,000 contribution becomes $3,000, and the affordable cost per lead triples.
Two cautions, because this is where the arithmetic gets abused.
Use a lifetime value you can evidence, from your own repeat-purchase data, not from an optimistic assumption about how long customers stay. A number you made up will justify any bid you like, which is exactly what makes it dangerous.
Remember when the money arrives. Lifetime value pays out over months or years. Your ad spend is due now. A business with a healthy LTV and no working capital can still bid itself into a cash crisis while being perfectly profitable on paper.
Not Every Service Deserves the Same Budget
The calculation is not one number for the business. It is one number per thing you sell, and they are rarely close.
We ask every client the same question early: of everything you offer, which has the highest profit potential over time? Not which is most popular, and not which you enjoy most. Which one, per customer acquired, makes you the most money.
For one chiropractic practice the answer was prenatal and pregnancy care, because those patients return on a schedule for months. That single answer reshaped the whole account: the keywords, the budget split, the bids.
Most accounts we audit spread budget evenly across services as though every customer were worth the same. They aren't. Once you have a max cost per lead per service, the split stops being a matter of taste.
It also changes what you are willing to lose. A service with a high lifetime value can justify bidding to the top of the page on expensive terms, because you will make it back over the relationship. A one-off transaction with thin margin cannot, and pretending otherwise is how accounts end up with their most expensive clicks pointed at their least profitable work.
The practical version: rank what you sell by profit per customer acquired, then check where your budget actually goes. If the ranking and the spending do not match, you have found something worth more than any bid adjustment.
What To Do With the Number
You have a ceiling. Here is how it gets used, in order.
1. Sanity-check what you already pay. Pull your actual cost per conversion for the last 90 days and compare. If you are above your ceiling, you now know it and can act. If you are comfortably below, you have room to bid up and take more of the market, which is usually the bigger opportunity.
2. Set target CPA near your actual, not your hoped-for. This is the most common way accounts get strangled. We manage an account where the true cost per acquisition had settled around $250 while the target sat at $200. A target meaningfully below what the account actually achieves throttles delivery and can stall it entirely. Raising the target to $250 fixed it. Start at what you are really achieving and walk it down as performance improves.
3. Check the budget can support it. A daily budget should be able to buy multiple conversions, not a fraction of one. If your cost per lead is $83 and your daily budget is $50, you cannot buy a lead a day, and the campaign will never accumulate enough data to optimize. We do not start accounts below about $1,500 a month for this reason, and it is arithmetic rather than preference.
4. Revisit when the business changes. Prices go up, margins move, close rates improve. The number is not permanent.
When the Math Says No
Sometimes you do the calculation and the affordable cost per lead comes out at $20, and the going rate for a click in your market is $12 with a 5% conversion rate, which means leads cost $240.
That is not a bidding problem, and no amount of optimization closes a 12x gap. It means one of four things has to change: the price, the margin, the close rate, or the channel.
That is a genuinely useful answer, and it is one you can only get by doing the arithmetic first. Plenty of businesses spend a year and a lot of money discovering it the slow way.
Of those four levers, close rate is usually the cheapest to move. Going from closing one lead in ten to one in five doubles what you can afford to bid without changing your prices, your product, or your ads. For a lot of businesses the fastest route to a workable Google Ads account runs through how quickly they answer the phone.
The second cheapest is narrowing what you advertise. A calculation that fails across everything you sell often works comfortably for your best two services, because the average was hiding them.
If the numbers do work, Google Ads becomes something you can scale deliberately, because you know the ceiling and can see how much room is under it.
The Twenty Minute Version
- Pull your average sale value and gross margin from your own records.
- Work out your close rate: leads in, jobs won, same period.
- Decide what share of your gross profit you are willing to spend acquiring a customer.
- Divide by your close rate. That is your maximum cost per lead.
- If customers repeat, multiply by an evidenced lifetime multiple, not a hopeful one.
- Compare it to what you are actually paying today.
- Do it per service, not for the business as a whole.
Then, and only then, put a number in the target CPA field. If you are not yet at the conversion volume where automated bidding works at all, that is a separate question and it comes first.
We do this with every client before we touch a bid, and it is usually the most useful twenty minutes of the engagement. If you would rather work it through with someone, we are glad to help. If you would rather do it yourself, do it on the back of an envelope tonight. The envelope is enough.
