Google Ads language targeting does not do what you think
Two settings sit near the top of every campaign, both look self-explanatory, and both do something other than what their names suggest. Between them they decide who sees your advertising — and most accounts have never had them examined.
This matters in two directions, and it is the same mechanism both times.
If you are a clinic in Mexico or Costa Rica whose patients are American, you need your advertising to reach people searching in English. And if you are a contractor in Texas, Florida or California, a large share of the homeowners in your service area search in Spanish — and if you have never thought about that, you are invisible to them while paying to compete for everyone else.
Same two settings. Opposite directions. Both usually wrong out of the box.
The language setting does not read the search
The intuition is that setting a campaign to English means it shows up for searches typed in English. That is not what it does.
Language targeting works from signals about the person: what language their Google account is set to, and the language of the content they generally browse. It does not inspect the query you just matched.
A person whose Google is set to English can type a Spanish query and still be eligible to see your English campaign. The language setting never looked at what they typed.
This has a practical consequence that catches people out constantly: you cannot use the language setting to control which queries you appear for. That is what keywords and negative keywords are for. The language setting narrows the audience pool; your keyword list decides the searches.
It also explains an experience almost every bilingual market advertiser has had — running English ads and getting Spanish enquiries, or the reverse — and concluding that Google is broken. Google is not broken. The setting is doing exactly what it was designed to do, which is not what its name implies.
The location setting spends money outside your map
The second setting is quieter and usually more expensive.
When you type a city into location targeting, Google's default behaviour is broader than the pin you dropped: it will serve your ads to people in that place and to people elsewhere who look interested in it. For an online business, that is often welcome. For a business that requires a body to physically arrive, it is a slow leak.
A roofing contractor in Houston does not benefit from being shown to someone in Ohio reading about Houston. A clinic in Cancún genuinely might — an American researching treatment abroad is a real prospect. Whether the default helps or hurts depends entirely on whether your customer travels to you.
There is a stricter option that limits delivery to people who are in or regularly in the targeted area. Google has renamed these choices more than once over the years, so do not go looking for a specific label — go looking for the option that talks about presence rather than interest, and read what it says before you accept it.
Why one bilingual campaign is a bad idea
Once people discover that both languages are reaching them, the tempting fix is a single campaign that covers both. It saves setup time and costs you clarity.
In a shared campaign the two languages compete for the same budget, and the cheaper one wins — not the more profitable one. You also lose the ability to answer the only question that matters: which language is producing paid work, and at what cost. Everything blends into one average that describes neither.
Two campaigns cost nothing extra to run. They let you set different budgets, bid differently, and send each to a landing page written in the language the reader actually wants — which is not a detail. Sending a Spanish-speaking homeowner to an English landing page after they clicked a Spanish ad is how you pay for the click and lose the customer at the door.
What to check this week
None of this requires an audit to look at. Open the account and check four things.
Your location setting, and specifically whether it is on the broad default. If your customer has to physically show up, tighten it and watch what happens to cost per enquiry over the following fortnight.
Your search terms report — the actual queries people typed, not your keyword list. This is where you discover the other language is already reaching you, and whether those enquiries are any good.
Whether your landing page matches the ad's language. A mismatch here wastes every click that got that far.
Whether you can tell the two apart in your reporting at all. If both languages run in one campaign, you cannot, and that is the first thing to fix — before optimising anything else, because until then you are optimising blind.
The pattern underneath all four is the same one I keep running into: the settings are not broken, the assumptions about them are. And an assumption that costs you a percentage of every month's budget compounds quietly for years, because nothing in the interface ever flags it as a problem.
Want someone to actually look?
The acquisition audit goes through the settings, the search terms and the tracking, and tells you where the budget is going that should not be — each finding with its source attached. Audit · $200, credited in full if we work together.