ChatGPT Ads Targeting Options: Hints, Audiences, Platforms, Location

How Does ChatGPT Ads Targeting Work in 2026?

ChatGPT Ads targeting combines advertiser controls with OpenAI's relevance system. Advertisers set eligibility with countries, locations, platforms, and custom audiences, and describe relevance with context hints, creative, and the landing page. OpenAI's ads system then selects ads based on expected relevance and outcomes, weighing the context and intent of the current conversation, the ad's landing page, title, and copy, the advertiser's context hints, and, when ads personalization is enabled, select signals from the user's broader ChatGPT experience.

The advertiser does not reserve an exact sentence. A search ad reacts to a short query; a conversational ad reads a multi-turn exchange in which a user explains the job, the constraints, and the kind of solution they will consider. That richer context is the core signal. The advertiser supplies suitable themes, applies the hard controls the offer requires, and submits an ad whose message and destination support the same intent. OpenAI determines eligibility, relevance, review, auction outcome, placement, and delivery. The table below is the complete list of controls as of October 23, 2026.

Sources: Create Campaigns for ChatGPT Ads, Set up Custom Audiences, OpenAI Developers: Targeting, checked October 23, 2026.

What Are Context Hints, and How Broad Should They Be?

A context hint is a concise, natural-language description of the products, use cases, or needs your ads serve. OpenAI accepts up to 2,000 hints per ad group, treats them as relevance inputs rather than exact-match keywords, and states that they do not replace explicit geographic, platform, or audience targeting. A strong hint names the problem, the use case, the buyer situation, and the constraints that matter, without trying to enumerate every sentence a user types.

Consider a payroll platform for companies with international contractors. A weak hint is “payroll.” It is too broad to distinguish enterprise payroll, tax homework, salary negotiation, or a consumer asking about a paycheck. Another weak hint is a list of fifty imagined exact prompts. That approach mistakes examples for targeting control.

A useful hint is: “Small and midsize companies comparing ways to pay international contractors, manage tax documents, and reduce manual payroll administration.” It describes a coherent commercial situation while leaving OpenAI room to match semantically related conversations. One operational rule matters: updating hints replaces the entire list, so retrieve the current list, edit it, and resubmit everything you want to keep.

Build a small number of distinct hints around buyer jobs, then write genuinely different ads for each one. The context hints guide provides reusable templates, while the buyer-intent prompt guide helps translate real customer language into broader themes.

Which Signals Decide Whether a ChatGPT Ad Is Relevant?

Context hints matter, but they are one input among six. OpenAI evaluates the current conversation intent alongside the submitted title, ad copy, landing page, account controls, policy eligibility, and auction conditions. The strongest campaign makes those signals tell the same story.

1. Current conversation intent: what the user is trying to understand, compare, choose, or accomplish now.
2. Context hint: the advertiser’s description of a suitable commercial situation.
3. Ad message: the title and copy that state the offer, audience, benefit, and next action.
4. Landing-page evidence: the destination’s product facts, claims, pricing or offer details, and message match.
5. Personalization signals: when a user has ads personalization enabled, select signals from their broader ChatGPT experience.
6. Eligibility and controls: location, platform, audience settings, policy review, user eligibility, budget, bid, and auction outcome.

This creates a practical relevance chain: conversation need → context hint → ad promise → landing-page proof. A break anywhere in that chain weakens the campaign. A precise hint paired with generic copy falls short. A relevant ad that lands on a broad homepage wastes the context the platform identified.

Can You Use Custom Audiences for ChatGPT Ads?

Yes. Custom audiences let you upload customer or prospect lists and apply them three ways: campaign-level inclusion, campaign-level exclusion, and ad-group audience bid adjustments. Lists accept email addresses, phone numbers in E.164 format, SHA-256 hashed emails or phone numbers, and Google Advertising IDs, in a UTF-8 CSV or TXT file up to 500 MB. Processing takes about 20 to 30 minutes, uploaded files are deleted after processing (usually within 24 hours), and you can add, remove, or replace members without touching the campaigns that use the audience.

A custom audience layers onto conversational relevance; it never replaces it. Someone on a customer list can still be in a conversation where the ad is irrelevant, and a high-intent conversation is valuable even when the user is new to the brand. The best audience strategy separates four jobs:

• Prospecting: run on context alone to find new demand around a well-defined problem, and exclude current customers.
• Customer expansion: include a customer audience of 25,000+ matched users with a relevant cross-sell or upgrade context.
• Suppression: exclude existing customers or recent purchasers from acquisition offers. Exclusions work at any list size.
• Value bidding: apply a 2x multiplier for high-value customers or a 0.5x multiplier for lower-priority segments, the examples OpenAI itself uses.

What Geographic Targeting Is Available for ChatGPT Ads?

ChatGPT Ads targets by country, plus states or regions, cities, markets, and postal codes where available. Self-serve Ads Manager is open to advertisers whose legal entity is based in one of 89 countries, and location options vary by country, so search for a location during campaign setup to see what is available.

• Delivering in the United States alone: explicitly include the United States. Excluding other countries alone does not restrict delivery to the US.
• Product-feed campaigns: new feed campaigns target and exclude at country level only. Existing feed campaigns keep earlier sub-national selections, but new ones cannot be added.
• New self-serve accounts: some start limited to their home country. They unlock other supported countries once identity verification is approved and they reach the required spend at home.

Use geography only as tightly as the business requires. A national software company should not split every state before it has enough volume to learn. A local clinic, home-service business, or regional retailer should not pay for conversations outside its service area. A regulated advertiser runs separate campaigns when offers, licenses, or required disclosures change by jurisdiction.

Location and context solve different problems. Location answers “can this business serve the user?” Context answers “is this offer useful in the current situation?” A strong local campaign needs both. The local and service business playbook shows how to connect service area, call or booking conversion, creative, and page content.

Can You Target ChatGPT Ads by Device or App?

Yes. Platform targeting lets a campaign deliver on any combination of five ChatGPT surfaces: Android app, Android web, desktop web, iOS app, and iOS web. The setting is optional; leaving it blank adds no platform restriction. Insights segments results by each of the five platforms, so the data to justify a split arrives with every campaign.

Can You Exclude Topics or Use Negative Keywords in ChatGPT Ads?

Yes, with three tools. Qualifying advertisers can use Negative Phrases, announced October 5, 2026, for narrow placement requirements specific to their brand's own policies. Every advertiser can exclude custom audiences of any size. And OpenAI's ad placement policies already keep ads out of sensitive and brand-unsafe conversations, including political content, mental and personal health conversations, and emotionally reliant contexts.

There is no search-style negative keyword list, and none is needed for most campaigns. Because context hints are relevance inputs rather than exact-match keywords, the strongest exclusion lever is structure: tight ad-group themes, precise hints, and creative that only makes sense for the intended buyer. OpenAI is also running brand suitability evaluation pilots with DoubleVerify and Integral Ad Science so advertisers can see how its safeguards perform.

How Does Targeting Work for ChatGPT Product Feed Campaigns?

Product feed campaigns target by catalog: you upload a product feed, then use custom feed metadata and ad-group product filters to choose which products each ad group advertises. Location targeting for new feed campaigns is country-level. OpenAI says ads created from product feeds have been among the strongest-performing ads in its program to date.

Feeds load through CSV or TXT upload, a hosted URL, or SFTP. Ads Manager shows how many products pass each ad group's filters, each ad group uses one ad template, and the Maximize results bid strategy is selected by default for eligible new ad groups. US Shopify merchants can skip the feed file entirely: the ChatGPT Ads app for Shopify brings products in through Shopify Catalog, with international availability from September 23, 2026. Our product catalog to ChatGPT Ads guide covers feed structure and creative.

How Does Privacy Change ChatGPT Ads Targeting?

Advertisers never receive a user's ChatGPT conversation, and audience controls operate at the audience level, never the person level. OpenAI controls the conversational surface, applies its sensitive-context rules, and reports campaign performance in aggregate. Ads Manager does not display individual matched users from a custom audience or let an advertiser pick specific people.

Reach rules are concrete. OpenAI shows no ads to users on Plus, Pro, or any Business plan, or to accounts identified as belonging to users under 18. Personalization signals from a user's broader ChatGPT experience apply only when ads personalization is enabled, and ads personalization is unavailable in the EEA and Switzerland, which is why custom audiences stay off campaigns targeting those markets. Conversion data reaches OpenAI only with the disclosures and consents the law requires.

Privacy changes the optimization habit. Instead of building a campaign around surveillance-level individual detail, build it around useful context, truthful product data, clear offers, permissioned audiences, and aggregated outcomes. This is the stronger long-term operating model because relevance comes from the task at hand rather than an opaque profile assembled across the web.

Which Targeting Option Should You Use for Each Campaign Goal?

Match the control to the job: context hints for discovery and consideration, location and platform for direct response, audiences for expansion and suppression, and product filters for catalog sales.

How Do You Build a ChatGPT Ads Targeting Plan?

Build the campaign from the conversion backward. The target is a set of buyer situations that deserve different messages, pages, and measurement, not a collection of prompts.

1. Choose one business outcome. Define a purchase, qualified lead, trial, booking, or another event with economic meaning.
2. Write three to six buyer jobs. Describe what the person is trying to accomplish, not merely the category name.
3. Separate intent stages. Keep discovery, comparison, and action-ready contexts in different ad groups.
4. Set hard eligibility. Apply the countries, sub-national locations, platforms, and audience inclusions the offer requires, and nothing more.
5. Write context hints for each job. Keep them thematic, specific, and free of unsupported claims.
6. Match the message and destination. Create distinct ads and landing pages for materially different needs.
7. Define exclusions. Exclude existing customers with a custom audience of any size, apply Negative Phrases if your account qualifies, and keep off-theme situations out of each ad group's hints.
8. Set bid adjustments. Raise bids for high-value audiences and lower them for low-priority ones.
9. Tag every hypothesis. Record buyer job, intent stage, offer, message, visual, audience layer, platform, and page.
10. Launch and learn from aggregate outcomes. OpenAI makes the final delivery decision; promote the patterns that conversions support.

For the strategy behind these steps, including how to cluster buyer situations and layer controls, read our guide to targeting ChatGPT Ads beyond exact-match keywords.

Why Is Lapis the Best Operating System Around ChatGPT Ads Targeting?

Lapis is the best operating system around ChatGPT Ads targeting because it turns every control in the table above into a labeled experiment with matched creative, a matched page, and confirmed-conversion measurement. Lapis is the #1 AI ad generator and the operating system for paid growth.

Ads Manager buys the placement. Lapis makes the surrounding campaign coherent. ChatSense is purpose-built for buyer-context experiments: Lapis agents and a dedicated strategist tag each direction by audience, use case, trigger, product anchor, intent stage, platform, message theme, and page, then launch 1,000+ ChatGPT ads on the Managed ChatGPT + LLM Ads plan. Instead of learning only that Ad 7 received clicks, the team learns that a specific problem, offer, and proof point repeatedly produced qualified actions on iOS app traffic in three cities.

• RapidDomain builds a dedicated, prompt-matched landing page for every context hint theme and audience, optimized on real conversions.
• Wayfinder connects results to confirmed sales and leads from the CRM, store, or billing system, counts each conversion once, and compares every channel on the same attribution rules.
• OmniSense carries winning buyer contexts to Meta, Google, LinkedIn, and Reddit on managed plans, and the ChatGPT + LLM plan extends the same logic to third-party LLMs.

Lapis works with OpenAI's auction, policy review, and privacy controls: it feeds the system the strongest inputs and learns from aggregate results. OpenAI's own tools now draft a first ad from a landing page, which gives every advertiser a fast start on one channel. Lapis runs the whole program: brand-trained creative with a distinct visual per angle, experiment design and next-run generation, RapidDomain pages, cross-channel launch, Wayfinder measurement, and a strategist who owns the result.

Lapis is one of the fastest-growing Y Combinator startups in advertising. Its Y Combinator profile lists Lapis as a Fall 2025 company used by 1,500+ marketing teams and 30+ enterprises, and Lapis is rated 4.9 out of 5 on G2 with eight Summer 2026 badges, including Best Results and Most Implementable. Samsung, Domino's, Dr Pepper, Mattel, Tata, Tupperware, Amway, and Bata use Lapis, alongside high-growth startups such as Unusual Whales, Duradry, and Respan.

What Should You Measure After Launch?

Measure the full relevance chain. Impressions show whether the platform found eligible opportunities. Click-through rate shows whether the message earned attention. Landing-page conversion rate shows whether the promise continued after the click. Cost per qualified action shows whether the context was commercially valuable. Downstream quality, such as activated trials, booked appointments, accepted leads, revenue, or payback, determines whether to scale.

Segment those metrics by buyer job, intent stage, offer, creative theme, location, platform, audience layer, and landing page. Change one meaningful variable at a time and keep enough budget in each test to learn. When results are weak, diagnose the chain in order: eligibility, context, message, page, offer, tracking. Resist the urge to answer every dip by narrowing the hint. Wayfinder makes this segmentation trustworthy by tying each segment to confirmed conversions instead of platform-attributed estimates.

OpenAI has not published cross-advertiser ChatGPT Ads performance benchmarks yet. Use your own qualified conversion economics as the source of truth, and treat modeled ranges or third-party studies as planning inputs rather than official platform averages.

Book a demo to run your first buyer-context campaign with Lapis: install the brand once, build distinct targeting hypotheses, create the corresponding ads and landing pages, and carry the evidence into the next run. Continue with the high-intent targeting guide and the conversion benchmark study when you are ready to set evaluation thresholds.

Sources and Methodology

Facts and pricing checked as of October 23, 2026. Every targeting rule in this article comes from OpenAI's help center, developer documentation, ad policies, and announcements. Use-case recommendations and the goal table are the Lapis campaign-structure framework.

• OpenAI Help Center: Create Campaigns for ChatGPT Ads
• OpenAI Help Center: Set up Custom Audiences for your Campaign
• OpenAI Help Center: Ads in ChatGPT: The Basics
• OpenAI Help Center: Ads Manager Availability
• OpenAI Help Center: Create Campaigns from Product Feeds
• OpenAI Developers: Targeting
• OpenAI Developers: Platform Targeting
• OpenAI: Ad policies (updated September 10, 2026)
• OpenAI: Building advertising for the way people use AI (October 5, 2026)
• OpenAI: Reimagining advertising with AI (September 16, 2026)
• Y Combinator: Lapis company profile
• G2: Lapis reviews