Running ChatGPT ads in the US, the EU and Asia: what changes by market

The OpenAI Ads platform is one set of rules applied in many markets, and a few of those rules change with geography: which targeting tools are available, what consent a conversion pixel needs, and which currency and timezone an account runs in. This guide sets out what changes when you run ads in ChatGPT in the United States, the European Union and Asia, and how Adsonomy structures an account that spans them.

Where can I run ads in ChatGPT?

Wherever the geo lookup in your ad account returns a location. Campaigns target locations by id, of type region, market or DMA, resolved through a search endpoint or the platform's geotargets CSV, with up to 2,500 location ids per campaign and a separate list of excluded locations. Availability is a property of the account and the platform's current market list, so check the lookup rather than a guide.

A country is a location like any other, and a region or market inside it is another id. Platforms are targeted separately: iOS app, Android app, web, desktop web, iOS web and Android web. Reporting can be segmented by country, so a campaign that spans several countries still reports per country, but budgets and bids live at campaign and ad group level, which is why Adsonomy runs one campaign per market.

What changes in the EU?

Two things. Custom audiences are not supported for campaigns targeting the EEA or Switzerland, where the documentation says personalised ads are not yet available, so targeting rests entirely on context hints, product feeds, locations and platforms. And conversion measurement is expected to respect consent: the pixel accepts a consent call, sends nothing while consent is false, does not replay blocked events, and the server side should stop sending the browser reference once consent is revoked.

In practice an EU site initialises the pixel with consent set to false, sets it to true after the visitor accepts measurement, and sends server-side events with hashed identifiers only: SHA-256 of the normalised email, phone number and name, with country, city, region and postal code in clear. Adsonomy hosts its data in Amsterdam, hashes before sending, and never depended on audiences, so an EU account is the same product as a US one with one tab fewer. The conversion tracking guide has the exact calls.

What changes in the US?

Custom audiences become available. The documentation restricts them only for the EEA and Switzerland, so a US campaign can include or exclude uploaded lists of emails, phone numbers or Android advertising ids, raw or hashed, and an ad group can apply a bid multiplier from 0.1x to 10x per audience. Inclusion and bid multipliers need a matched-user minimum with a public planning threshold of 25,000; exclusion works with any ready audience, even an empty one.

The data must be first-party and used with the rights, notices and consents it was collected under; broker data is banned. Adsonomy records an attestation on every audience upload and shows it in the change log. Audiences are a refinement, typically to exclude current customers or to lift bids on a high-value list; the contextual match remains the base layer, and exact matched counts are never exposed by the platform.

What changes in Asia?

Nothing in the documented rules is specific to Asia. Custom audiences are restricted only for the EEA and Switzerland, so for markets outside those the same audience tooling as in the US applies. Which countries an account can target is, again, whatever the geo lookup returns. The ad account's currency and timezone decide how limits and reports are read, and platform targeting lets a campaign concentrate on iOS and Android where that is where the traffic is.

Consent rules are a matter of local law rather than of the platform, and the pixel's consent switch works the same way everywhere. Language is the practical difference, and it is handled the same way in every region, as described below.

How do currencies and timezones work?

One API key is one ad account, and an ad account has one currency and one recorded timezone. All money is in micros of that currency, and daily budget minimums depend on the currency. The account daily spend limit repeats each account-local day. Insight queries take a timezone parameter, so a report can be read in the account's day or in yours.

Adsonomy keeps a plan's currency equal to the ad account currency and runs the plan period in the workspace timezone. Targets such as CPA or CPO are entered in account currency units and ROAS as a ratio, so a target set for a US account is in dollars and one for a Dutch account in euro. Reporting in the product uses the same daily insight rows the platform serves, read in the workspace timezone.

How is language handled?

The OpenAI documentation has no language targeting field and says nothing about language; targeting is locations and platforms. Adsonomy's operating assumption, marked as an assumption in the product and to be confirmed with OpenAI, is that the platform's matching is multilingual but a creative is not, so language lives at the ad group level: one ad group per narrative and language, with hints written in that language and an English mirror kept alongside that can be switched off.

For a market with one dominant language this is one set of ad groups. For a market with several, such as Switzerland or Belgium, or for an English-speaking audience inside an Asian market, it means separate ad groups per language under the same market campaign, each measured on its own.

How should a multi-market account be structured?

One campaign per market, carrying the location targeting, the budget and the conversion event settings; one ad group per narrative, an angle in a language, carrying the hints and the bid; one or more ads per group. Performance is then readable per market at campaign level and per angle and language at ad group level, and a spend plan can be split by campaign so that each market has its own line and its own pace.

Adsonomy serves all three regions with this structure. The EU account has no audiences tab; the US and Asian accounts have one, behind an attestation. Hints, creative, bids and budgets are managed the same way everywhere, and Nomy's guardrails, change log and kill switch do not vary by market. When an account with campaigns in several regions is connected, Nomy proposes narratives per market and language, and the plan can give each campaign its own budget line.

Frequently asked questions