Your Google Ads account shows a 4x ROAS. Your Meta account shows a 12 EUR cost per lead. Both numbers are green, and you still cannot say whether your acquisition makes money. That is expected: both platforms measure what happens on your website, never what happens inside your business.

Answering "am I profitable" means leaving the interface and rebuilding a chain: what you pay for a click, what that click becomes on your side, and what is left once your costs are paid. Here is the order I follow when I open an account, and the places where the chain breaks.

What the platform counts is not what you bank

A conversion in Google Ads or Meta is an event fired on your site: a form sent, a call started, an order placed. The platform counts it to the second. What it cannot know is whether that form became a customer, whether that order came back as a return, and whether your margin survived shipping.

I manage three accounts of the same B2B SaaS group where the full sales funnel flows back from the CRM into Google Ads, stage by stage, up to the closed deal. That lets me compare two things nobody usually sees side by side, over the last 90 days and on the same clicks:

  • The recorded lead (326 over the period) lands 95 % of the time within 24 hours of the click.
  • The signed deal (26 over the period) lands in close to four cases out of ten more than two weeks after that same click.

Two realities, one account. And the second number is a floor, not a full measurement: Google's tracking window stops at 90 days, so anything that closes later appears nowhere. If your sales cycle runs longer than three months, part of your revenue is structurally missing from the interface where you make your decisions.

This is the first thing to accept: the platform is an event counter, not a profit and loss statement.

Your CPA means nothing until you compare it to your ceiling

Cost per acquisition on its own says nothing. A 24 EUR CPA is excellent for a mattress seller and unworkable for a sock seller. The only question that matters: what is the highest amount you can pay for a sale without losing money?

For online sales, the available margin comes in two steps.

  • Your average order value multiplied by your gross margin gives what is left after product cost. An 80 EUR order at 40 % margin leaves 32 EUR.
  • You then remove the share you want to keep as profit. If you want to keep a quarter of that margin, your maximum CPA drops to 24 EUR.

Above 24 EUR, every extra sale costs you money, even with a ROAS the platform shows in green. Below it, you have room to raise bids and take volume. That ceiling is the first number to set, before any optimisation: it turns a reporting column into a decision. The ideal CPA calculator runs the maths on your own numbers.

The same logic works in ROAS instead of euros: your break-even point is 1 divided by your margin. A 40 % margin sets a minimum ROAS of 2.5. It is pure maths, and it is the kind of marker that stops you celebrating a 2x ROAS on a product that needed 2.5. The break-even ROAS calculator covers that case.

Between the lead and the customer sits your close rate

If you sell a service, you are not buying customers, you are buying conversations. And not all of them convert.

Customer acquisition cost (CAC) is what you spend on ads to sign a customer, not to receive a form. That is the difference most dashboards never make.

Moving from one to the other is a single division: your cost per lead divided by your close rate. A 40 EUR CPL with 20 % of leads signing gives a 200 EUR CAC. The number you were watching was five times too optimistic, and nothing in the interface would have told you.

That close rate is also the cheapest lever you own, because it costs nothing in media spend:

  • Moving from 20 % to 25 % close rate brings the same CAC down from 200 to 160 EUR, without touching a single bid.
  • The other way round, a form that is too easy to fill can lower your CPL and your close rate at the same time. Your reporting improves while your acquisition gets worse.

This is why I always ask for the close rate before commenting on a cost per lead. Without it, a CPL is not a performance figure, it is a measure of how fast forms get filled.

The trap that breaks almost every calculation

Here is the mistake I see most often, including in teams with good data: dividing this month's sales by this month's leads.

If your cycle runs six weeks, the customers signed in March came from leads generated in January and February. Comparing the two mixes different populations. The result is not roughly right, it is wrong, and it moves mostly with your budget growth: a month where you spend more mechanically lowers your apparent close rate, even though nothing changed in your sales quality.

The fix is easy to state and takes some discipline: track cohorts. You take the leads of a given month, follow them to their outcome, and only compare months that have matured. On a six-week cycle, your most recent usable month is two months old. That is frustrating, and it is the only reading that holds.

Direct consequence: stopping a campaign after three weeks on a two-month sales cycle is judging a harvest during germination.

A customer is worth more than an order

So far we have counted a single purchase. Yet most businesses do not make money on the first sale, they make money on the second.

Lifetime value (LTV) is the total a customer brings you across the whole relationship, margin included. A 29 EUR monthly subscription kept for fourteen months is not worth 29 EUR, it is worth 406. A 200 EUR CAC reads very differently depending on which of the two you look at.

Two warnings before celebrating:

  • Work in margin, not in revenue. A gross LTV is a sales number, an LTV net of costs is an owner's number.
  • Measure the real lifespan, do not assume it. This is the variable where optimism does the most damage, because it multiplies everything else.

The LTV to CAC ratio, and the payback period

Once both numbers are set, their ratio decides. The LTV to CAC ratio divides what a customer brings over their whole life by what they cost to acquire. A 600 EUR LTV against a 200 EUR CAC gives a ratio of 3.

  • Below 1, you are paying to work: every new customer widens the gap instead of closing it.
  • Around 3, the zone most businesses aim for: enough to pay the structure and keep investing.
  • Far above 5, the question flips: you may be too cautious, and leaving the market to someone else.

These levels are markers, not a law. A shop with monthly repeat orders and a firm that signs two clients a year are not judged on the same scale. I covered the full calculation, channel by channel, in the article on the CLV to CAC ratio, and the short definitions live in the glossary.

One more number matters as much as the ratio: the payback period, meaning how many months before a customer has repaid their acquisition cost. A ratio of 4 with eighteen months of payback is an excellent deal for a business with cash, and a serious problem for one without. Profitability and solvency are not the same question.

Put side by side, these metrics do not look at the same thing. Each one stops at a different point of the chain, and that is what lets two green numbers tell two opposite stories.

The platform Your business Click Conversion Real customer Margin Whole life CPA counts conversions, not customers ROAS revenue, never margin CAC real customers, once closing has happened POAS profit, once product costs are paid LTV/CAC the whole relationship, repeat orders included

Google and Meta answer different questions

Comparing both platforms by putting their interfaces side by side always makes Google win, and part of that win is a measurement artefact.

  • Search captures intent that already existed. Someone looks for your product, you show up, they buy. The link between click and sale is short and visible.
  • Social creates intent that did not exist. The person sees, remembers, and sometimes comes back through a brand search several days later. That sale gets counted by Google.

The result is that Meta is regularly undervalued by attribution models, while Search collects sales it did not create on its own. Adding up the conversions each platform reports almost always gives a total higher than the orders actually recorded, since both claim the same sale.

The marker that does not lie is rougher and more honest: your total revenue divided by your total ad spend, all channels together. It does not tell you which channel works, but it tells you whether the whole thing holds. When that global ratio degrades while every platform reports good results, your measurement is wrong somewhere. That is the moment for an account audit rather than more budget.

What stays invisible, and how to handle it

Everything above assumes a number exists. Part of your profitability escapes measurement entirely, and I would rather say it plainly than sell a precision that does not exist.

Word of mouth comes first. A happy customer brings another one, who will arrive through a brand search or a direct visit. That second sale gets credited to direct traffic, or to nothing at all, when it was in fact paid for by the campaign that acquired the first customer. Your advertising keeps producing customers long after the budget is spent, and no dashboard will show you that.

Brand awareness comes next. Someone who has crossed your name three times over six months does not convert on the third contact, they convert the day their need appears. On that day, they type your name into Google and cost you a few cents. Everything built before that is counted nowhere, and it is exactly why that click was so cheap.

Three approaches get you close to that territory without pretending to measure it:

  • The open question at signup. "How did you hear about us?" as a free text field is imperfect and self-reported, but it captures paths no tool can see.
  • Tracking your brand search volume. If it grows while you invest in awareness, something is building, even if attribution files it nowhere.
  • The geo test. Cut or strongly raise one channel in a comparable region, then look at that region's total revenue. It is the only method that measures a real contribution instead of a click path.

The position I defend: this grey zone justifies keeping a share of budget on channels that never look perfect in a spreadsheet, but it never justifies covering losses. An unmeasurable channel whose removal changes nothing in your total revenue was not invisible, it was ineffective. The geo test is what separates the two.

The 10 ratios that tell you if you are profitable

These are the ones I look at, in the order they build on each other: each needs the previous one to mean anything.

  • Gross margin rate: (selling price minus product cost) divided by selling price. It governs all the others, and it is the only one that does not come from marketing.
  • Break-even ROAS: 1 divided by your gross margin. The floor below which one more sale costs you money.
  • ROAS: revenue generated divided by ad spend. Useful for daily steering, blind to your margin.
  • POAS: profit divided by ad spend. The same calculation, once product costs are paid. Two products at the same ROAS can have opposite POAS.
  • MER: total revenue divided by total ad spend. The only one that depends on no attribution model, so the only one no platform can flatter.
  • CAC: acquisition spend divided by the number of new customers. Not to be confused with CPA, which counts every conversion, repeat purchases included.
  • Close rate: signed customers divided by leads received. This is what turns a cost per lead into a cost per customer.
  • LTV to CAC ratio: customer lifetime value divided by acquisition cost. The verdict on everything above.
  • CAC payback period: the number of months before a customer has repaid their acquisition. A duration rather than a ratio, and it decides whether you can fund your growth.
  • Lin Rodnitzky ratio: the CPA of all your search queries divided by the CPA of the queries that converted. Around 1.5 the account is healthy; above 2, part of the budget goes to traffic that never converts.

Only two of these can be read inside an ad platform: ROAS, and the Lin Rodnitzky ratio if you dig into the search terms report. The other eight need a number that lives in your accounting or your CRM, and that is exactly what makes them hard to dress up.

Where to start

  • Set your ceiling before looking at performance: maximum CPA if you sell online, maximum CAC if you sell a service. Until that number exists, no campaign figure can be read.
  • Ask your sales team for their close rate, even a rough one. It turns a cost per lead into a cost per customer, and it often changes the conclusion.
  • Check your sales cycle length before judging a month. A cycle longer than your analysis window produces inverted conclusions.
  • Look at your global revenue to ad spend ratio once a month, on top of per-platform numbers. When the two readings disagree, measurement is the suspect.
  • Keep a line for what cannot be measured, and test it by region once a year instead of arguing about it.

Rebuilding this chain takes half a day and beats six months of bid adjustments on numbers that mean nothing. That is exactly the work I do when opening an account in a Google Ads audit or a Meta Ads audit.

Frequently asked questions

Are Google Ads worth doing?

Google Ads is worth doing when your cost per acquisition stays under the margin you can spend on a customer. The platform knows neither your margin, nor your close rate, nor your lifetime value, so it cannot answer for you. The calculation uses your own numbers, once a quarter.

Do Google Ads pay off?

They pay off when people already search for what you sell and your margin absorbs the cost of a click in your market. For a need nobody puts into words yet, search has little volume to give you and a discovery channel like Meta works better. Check the search volume before the budget.