Your European CAC is not one number. It is one per country
We set out to write the usual piece: a table of cost-per-click by European market, so you could budget. Then we checked the sources against each other. The same country turns up at €0.43, at about a dollar, and at €1.67, depending on whether someone measured actual clicks, read bid estimates out of a planner, modelled it, or averaged nineteen of their own accounts. So this page does something else instead.
Four methods, one label
The spread is not sloppiness. It is that “average CPC in Germany” describes at least four different quantities, and almost nobody says which one they are quoting.
The strongest basis, and the slowest to appear. The measured figures we found for European markets date from December 2024. By the time click data is aggregated, published and cited, the auction it describes has moved.
Accurate, and behindWidely quoted and frequently mistaken for measurement. A planner reports what the top-of-page bid range looks like for a keyword set, not what anyone paid. One of the comparisons we found derives whole-country ratios this way, from a single account.
A forecast of a price, not a priceOne source gives tidy 2026 values for sixty-odd countries and states plainly that they are modelled rather than measured, checked only against the ranking of an older measured set. The table looks exactly like measurement. The caveat is one line.
Read the methodology note, alwaysOften the most honest of the four, because practitioners tend to say what their sample is. One German-language benchmark is drawn from nineteen accounts, and its author writes that it is not representative of the German market. The figure still gets cited as if it were.
Honest about itself, quoted dishonestlyThere is a fifth problem on top of the four. Google and Meta figures circulate side by side and are routinely mixed, although they price different inventory for different intent. A page quoting one platform’s cost per click against another’s conversion rate is producing a number that corresponds to nothing.
The ordering, and not much else
If you read enough of these sources the absolute values stay stubbornly inconsistent, but one thing holds across methods: the relative positions of the markets.
- Germany is consistently more expensive than France, and France than Portugal
- Relative positions are stable enough to form a hypothesis with
- Sector matters more than country: within one language, ecommerce and legal differ several-fold
- Your own account data, which is measured, current, and about your keywords
- Any single absolute figure, which moves by roughly a factor of two across sources
- A country average applied to your sector
- Google numbers mixed with Meta numbers
- Modelled tables read as if they were measurements
- Anything cited without its method, which is most of what circulates
Which gives the only defensible way to build a European acquisition model: anchor it to a number you actually measured, in a market you already sell in, and apply a relative adjustment you are willing to defend. That is what the calculator below does. It contains no benchmark from us, because we do not have one worth giving you.
And now the part that actually dominates
Even with a perfect click price, the cost of acquiring a European customer is mostly not advertising. It is the one-off cost of being able to sell in that market at all, divided by however few customers you win in the first year.
Drag the volume down to twenty and watch the bar. For most companies entering a European market for the first time, the honest first-year number is closer to twenty than to two hundred, and at that volume the entry cost is the whole conversation. Optimising click prices in that situation is optimising the smaller half of a number you have not yet made small.
What “opening a market” actually consists of
This is the line item most plans leave out, because it does not look like marketing and does not arrive as an invoice from an agency.
Translating a site is the visible cost and the smaller one. Keeping it current, supporting customers in that language, and handling the inbound that follows are recurring. The EU has twenty-four official languages across twenty-seven member states, which is why “sell across Europe” is not one project.
Card penetration, bank transfer habits and local schemes differ sharply between European markets. A checkout that offers only what works at home converts worse, and that shows up as acquisition cost rather than as a payments problem.
A local address, a local phone number, invoices that look right, terms in the local language, and some evidence that other local customers exist. None of it is expensive; all of it is work, and its absence raises the cost of every click you buy.
Beyond the €10,000 EU-wide threshold for consumer sales you charge the customer’s national VAT rate, which the One Stop Shop lets you handle through a single registration. That is an administrative saving rather than a commercial one, but it is the difference between one registration and several.
What we could not find out
We looked. The measured data we found is from 2024; the 2026 tables are modelled or drawn from single accounts. There may be a good proprietary dataset behind a paywall, and if you have access to one, use it in preference to anything on this page.
If it exists, it is not publicIt depends on trust, on the incumbent, on whether your category is bought the same way there, and on whether anyone has heard of you. No dataset contains this, and it moves CAC more than click price does.
Find out with a small test, not a modelTranslation and payments can be quoted. Support load, the trust work and the long tail of small adaptations cannot be, before you start. Budget the known half and expect the unknown half to arrive.
Quote what is quotable, reserve for the restMeasured, current, and about your actual customers. It is the only solid number in the whole exercise, which is why this page is built around it rather than around anyone’s benchmark.
Start here, alwaysWhat we read, and why we did not use it
- Measured click data — Statista and Semrush figures for European markets, dated December 2024. Germany around $1.04 and France around $0.68. The strongest basis we found, and nearly two years old.
- Modelled 2026 tables — a sixty-country table giving Germany €1.67, France €1.36 and Portugal €0.87, stated by its own author to be modelled rather than measured and checked only against the ranking of the older measured set.
- Keyword Planner estimates — country ratios derived from bid ranges rather than from clicks paid for, including a September 2026 comparison built from a single account.
- Agency account samples — a German-language benchmark from nineteen accounts, ranging from about €0.90 for ecommerce to €5.50 for legal, with its author stating it is not representative of the German market.
- European ecommerce median — around €0.43 for search, reported September 2026, with the authors noting it does not imply identical costs across countries.
- Meta benchmarks — a separate set entirely, quoted here only to note that they are routinely and wrongly mixed with Google figures.
- One Stop Shop and the €10,000 EU-wide threshold for distance sales to consumers, which is a rule rather than an estimate.
We have deliberately not linked these as recommendations. Several are useful if you read the methodology note; all of them are harmful if you lift a figure out of the table. The point of listing them is to show the spread, not to send you to one.
What people ask once the benchmark falls apart
Why will you not publish European CPC benchmarks?
Because we checked them and they disagree by multiples on the same country. Measured click data, Keyword Planner bid estimates, modelled figures and single-agency account data are four different things reported under one label. Publishing a number we could not stand behind would make this page look authoritative and make your budget wrong.
What does survive the disagreement?
The ordering, roughly. Across the sources we looked at, Germany is consistently more expensive than France, and France than Portugal. The relative positions are stable; the absolute values move by around a factor of two depending on who is counting and how. So use the ordering to form a hypothesis and your own account data to put numbers on it.
Is CPC the same as CAC?
No, and conflating them is the more expensive mistake. Cost per click is a price for attention. Cost of acquiring a customer includes everything between the click and the payment: localisation, support in the language, payment methods people trust, and whatever credibility work a foreign supplier has to do. In a new European market those are largely fixed costs, which means they fall on however few customers you win.
So what actually dominates European CAC?
At the volumes most companies enter a market with, the one-off cost of opening it. Translating and maintaining a site, supporting a language, adding local payment methods and building enough trust to be chosen over an incumbent are paid once and divided by your first-year customers. At twenty customers that division is brutal; at two thousand it disappears.
How many European markets should I enter at once?
Fewer than the map suggests. Each additional language is a fixed cost that recurs annually in support and content, and the EU has twenty-four official languages across twenty-seven member states. One market entered properly usually beats four entered thinly, because the fixed costs multiply while the volume does not.
Where does a Portuguese entity help with this?
With the mechanics rather than with the acquisition cost itself. A single VAT registration and the One Stop Shop let you sell to consumers across the EU without registering in each country, once you pass the €10,000 EU-wide threshold. That removes an administrative cost; it does not reduce what it costs to be chosen.
Send the model, we will attack the volume
Put your own CAC in, choose a volume you would defend, and send us what came out. We will tell you which assumption we think is optimistic, what the entry cost realistically contains for your category, and whether one market done properly beats the three you were considering. The honest answer is sometimes to sell from where you are for another year.