A campaign is the top-level unit in ChatGPT Ads Manager that defines one advertising objective and one budget for the ad groups and ads grouped beneath it. It sits above ad groups and individual ads in the hierarchy, so it controls spend and goals, while targeting themes and creative live one level down.
If you advertise inside ChatGPT, how you split work into campaigns decides whether your reporting shows which objective and region actually drove results. Pool mismatched goals into one campaign and you lose the ability to read performance or shift budget toward what converts.
Every ad you run in ChatGPT belongs to a campaign, the level where ChatGPT Ads Manager records your objective, budget, schedule, and the countries where delivery is eligible. It groups ad groups and ads that share one goal, and its settings flow down to everything inside it.
A campaign in ChatGPT Ads Manager carries a title, an objective, a budget, start and end dates, target countries, optional platform targeting, and optional custom audiences to include or exclude. The objective is the pivotal choice: Views priced per thousand impressions, Clicks priced per valid click, or Conversions optimized toward an event you select. OpenAI does not let you edit a campaign into a different objective after creation, so the choice sets pricing and optimization for the life of that campaign.
The campaign sits one level above ad groups, which hold your bids, context hints, and targeting themes, and two levels above individual ads. Paid campaigns run alongside the organic citations your brand earns inside ChatGPT answers, and AirOps tracks that earned visibility so you can see where paid and organic reinforce each other.
Setting up a campaign in ChatGPT Ads Manager follows a fixed order, and each choice narrows the ones after it.
Pick the objective. Choose Views, Clicks, or Conversions. This sets how your ads are priced and what delivery optimizes toward, and you cannot change it later without building a new campaign.
Set the budget. Choose a daily budget or a campaign-total budget. OpenAI sets a minimum daily budget of twenty-five dollars for accounts billed in US dollars (OpenAI Help Center), and daily delivery can run above or below that target within weekly limits.
Define the schedule. Set start and end dates so delivery lines up with launches and promotions.
Choose targeting. Select the countries where ads are eligible, optionally narrow to platforms like iOS app or desktop web, and include or exclude custom audiences.
Build underneath. Add ad groups with context hints and bids, then add the ads themselves, which go live after they pass OpenAI's review.
The campaign tells you how one objective and budget performed across its ad groups. It will not tell you which conversation triggered an ad, because ChatGPT matches on context you never see directly.
Campaign structure is a buying decision, because it determines what you can learn from your spend and how fast you can move budget toward the ads that actually pay off.
Clean attribution starts here. OpenAI's Ads reporting breaks out performance, including conversions, at the campaign, ad group, and ad level, so a campaign built around one clear goal lets you read results without untangling mixed objectives.
Budget control lives at this level. Your campaign budget caps spend across every ad inside it, so a sloppy structure can quietly drain spend on a weak objective before you notice, which is the single most common failure mode that new advertisers hit.
Mismatched goals break optimization. Pack Views and Conversions into one campaign and delivery optimizes toward a blur, which starves your real goal and burns through the week's budget before you can diagnose the problem.
SEO managers use a campaign to run a Conversions objective against product pages while they track the organic citations the same pages earn in ChatGPT answers.
Growth marketers use a campaign to test a new region with a capped daily budget before committing total spend to the market.
Demand gen leads use a campaign to separate top-of-funnel Views from bottom-of-funnel Conversions so pipeline reporting stays clean.
A campaign sits inside your ad account and above its ad groups and ads, so every setting you choose at the campaign level, from objective to budget to country targeting, cascades down to the creative and targeting nested beneath it.
The objective you pick, whether Views, Clicks, or Conversions, fixes both how ChatGPT prices your ads and what its delivery system optimizes toward, and OpenAI locks that choice once the campaign exists.
Instead of keyword lists, ChatGPT decides when a campaign's ads are eligible by matching the live conversation against advertiser signals like context hints, so your targeting themes shape delivery without guaranteeing any single chat.
Groups related ad groups and ads under one objective so reporting stays readable.
Controls total spend with a daily or campaign-total budget cap.
Runs on ChatGPT, reaching users on the Free and Go plans as they research purchases.
Targets by country and optionally by platforms like iOS app or desktop web.
Separates objectives so Views, Clicks, and Conversions each optimize cleanly.
Scales across regions by cloning into a new objective or billing model.
Build one campaign per distinct goal, so delivery and reporting both stay focused on a single objective.
Pick the objective deliberately, because OpenAI will not let you edit it later without rebuilding the campaign.
Start with a daily budget, so you can watch delivery before you commit a campaign-total amount.
Name campaigns clearly and consistently, so reporting and handoffs stay legible as your account grows.
Explicitly include your target countries, because excluding others does not restrict delivery to the ones you want.
Create several ad groups inside each campaign, so you can test themes and messaging without multiplying campaigns.
Avoid the habit of folding every product line and region into one catch-all campaign. It feels tidy, but it hides which objective and market actually drove results, and it leaves ChatGPT optimizing toward an averaged goal that serves none of them well in the end.
AirOps: tracks your brand's organic visibility and citations inside ChatGPT answers, so you can see where paid campaigns complement the coverage you already earn.
OpenAI Ads Manager: the native console where you create campaigns, set objectives and budgets, manage targeting, and read conversion reporting.
Google Analytics 4: measures the downstream traffic and conversions from your campaign's landing pages when you tag ad URLs with UTM parameters.
Map your goals. List the distinct objectives and markets you want to advertise toward, and sketch one campaign per goal. You can do this in a spreadsheet this week with no budget approval.
Open an ad account. Create an account in OpenAI Ads Manager and complete identity verification and billing, since ads cannot deliver until both are approved.
Create your first campaign. Choose the objective, set a daily budget, pick start and end dates, and explicitly include your target countries so delivery reaches only the markets you chose.
Build ad groups and ads. Add ad groups with context hints and bids, then write clear creative variations with genuinely different angles, and skip near-duplicate copy.
Launch and monitor. Publish the campaign, wait for review to clear, then watch delivery and conversions in Insights and shift budget toward what performs best before you scale spend.
A campaign is the top-level unit in ChatGPT Ads Manager that pairs one objective with one budget for the ads beneath it.
You build it by choosing an objective, setting a budget and schedule, and targeting countries before adding ad groups and ads.
The objective is locked once the campaign exists, so a wrong choice forces a full rebuild instead of a quick edit.
Mixing unrelated goals into one campaign muddies attribution and pushes delivery toward an averaged target.
The leverage is in structure: one clear goal per campaign keeps reporting clean and budget flowing to what converts.
A campaign defines your objective and budget, and an ad group lives one level inside it to handle targeting, bids, and context hints. The campaign answers what you are trying to achieve and how much you will spend; the ad group answers who should see the ads and under what themes. One campaign can hold several ad groups, which is how you test different value propositions or audiences against the same objective without splitting your budget reporting. Because the objective is fixed at the campaign level, every ad group inside inherits that pricing and optimization goal. If you want a genuinely different objective, build a separate campaign instead of adding an ad group. Keeping the two levels straight is what lets your reporting show both the goal that worked and the message that carried it.
Start small with a daily budget so you can watch delivery before you commit to larger amounts. OpenAI enforces a minimum daily budget that varies with your account's billing currency, so the floor you work from depends on where you are billed. There is no universal right number, because what you need also depends on your objective and how competitive your category is in ChatGPT's relevance-weighted auction. A sensible opening move is to fund one campaign at or just above the minimum, let it run long enough to gather meaningful performance data, and then raise the budget on the campaigns that convert. Remember that a daily budget is a delivery target, so actual spend flexes day to day and your firmer ceiling is the weekly limit. Treat the first week as a learning budget, read the Insights reporting, and scale deliberately once you can see what each campaign returns.
Your campaign spend varies day to day because a daily budget behaves as an average delivery target, so ChatGPT can spend more on high-demand days and less on quiet ones. Across the weekly budget period the system smooths delivery toward your target, which means an underspent day can be made up later in the same week. Demand itself fluctuates: the volume of relevant conversations, the competitiveness of the relevance-weighted auction, and your targeting all shift from day to day, and your spend follows that eligible inventory. OpenAI also applies guardrails so a single day will not run away from your intended pace, and the weekly total stays within a fixed multiple of your daily amount. If the swings feel too large, lengthen the window you judge performance over, because one day is rarely enough signal. Judge each campaign on its weekly trend, and change the budget only when that trend tells you to, ignoring the noise of any single day.
No, you cannot pick the exact conversations your campaign shows up in, and that is by design. ChatGPT does not use keyword lists or let you buy a specific chat; instead it matches ads to the live conversation using signals like the topic, your landing page, your ad copy, and the context hints you provide. What you can influence is strong: write context hints that read like the real questions your buyers ask, keep your creative specific and clearly relevant, and target the countries, platforms, and audiences that fit your offer. Those inputs raise the odds that your campaign is eligible when a relevant conversation happens, without guaranteeing any single one. You also never receive private conversation transcripts, so you optimize toward patterns in the aggregate and leave the individual chats to the matching system. Treat delivery as something you steer through inputs, never something you hand-pick.
A well-structured campaign carries one clear objective, one budget sized to that goal, and a set of ad groups that test genuinely different angles toward it. The simplest benchmark is readability: if you can open your reporting and immediately say which objective, region, and message drove results, your structure is working. Good campaigns map to real business initiatives, so a product launch, a seasonal push, and an always-on acquisition effort each live in their own campaign instead of sharing one. Naming is consistent enough that a teammate can understand the account without a guided tour. Budgets are set deliberately, with the daily or total choice matching how tightly you need to control spend. Inside each campaign, several ad groups let you learn what messaging resonates without inflating your campaign count. If your account has one sprawling campaign trying to do everything, that is the clearest sign the structure needs to be split before you scale spend.