How to Run ChatGPT Ads: A Step-by-Step Guide for Marketers

ChatGPT’s self-serve advertising platform opened up to everyone in May 2026, and it’s quickly becoming a channel marketers can’t afford to ignore. Unlike a typical social platform, ads on ChatGPT appear directly inside a user’s conversation at the exact moment they’re asking about something relevant — which puts it much closer to Google Search ads in spirit than to Meta or Instagram.

The catch is that ChatGPT Ads work differently to almost every platform marketers are used to. There are no keywords in the traditional sense, targeting relies on a very different mechanism, and the reporting available to advertisers is still limited. This guide walks through exactly how to get a campaign live, based on how the platform currently works.

Who you can actually reach?

ChatGPT Ads only serve to users OpenAI believes are adults, and only on the Free and Go tiers — including people using ChatGPT while logged out. If your audience is mostly paying for a higher subscription tier, this platform won’t reach them. But if your customers are the kind of people using the free version of ChatGPT regularly, it’s a genuinely new pool of intent-driven attention to tap into.

What it costs to get started?

Average cost-per-click currently sits somewhere in the $2 to $5 range depending on industry, and the platform enforces a minimum bid of around $3 — bidding lower than that risks your ad simply not being shown at all. On the plus side, the barrier to entry is low: campaigns can run on a daily budget as small as $25, a dramatic drop from the six-figure minimum commitments required during the platform’s earlier pilot phase.

Step 1: Confirm your business is eligible

OpenAI restricts advertising in a handful of categories — adult content, gambling, tobacco, alcohol, and financial services among them — so it’s worth checking OpenAI’s ad policies before you invest any time in setup. Targeting is currently limited to a handful of countries, including the US, Australia, Canada, the UK, Japan, New Zealand, and South Korea, with more expected to be added over time.

Step 2: Set up your ad account

You’ll need your legal business name, a business tax ID, a payment method, and a favicon before creating an account directly at ads.openai.com. One quirk worth knowing: OpenAI doesn’t want agencies creating accounts on a client’s behalf — instead, the business creates the account and adds the agency as a user. If you’re working with an SEO and digital marketing agency to manage this alongside your other channels, this is the kind of platform-specific detail worth getting right from day one.

Step 3: Structure your campaign

If you’ve used Google Ads or Meta Ads before, the structure will feel familiar: Campaign > Ad group > Ad.

At the campaign level, you’ll choose an objective (Reach, Clicks, or Conversions), set your budget, target or exclude locations, and optionally upload first-party audience data. At the ad group level, you set your bid and a field called “context hints,” which is where ChatGPT Ads really departs from anything you’ve used before.

Bidding works similarly to familiar platforms too — you can bid for clicks (paying per click, similar to manual CPC bidding), for reach (paying per thousand impressions), or for conversions if you’ve set up tracking. Ad placement is decided through a relevance-weighted auction that factors in landing page quality and ad relevance alongside your bid, in a similar vein to Google’s Ad Rank.

Step 4: Get your head around targeting (there are no keywords)

This is the biggest mental shift for marketers coming from traditional PPC. There’s no keyword list, no demographic targeting, no in-market audience segments. Instead, the primary lever you have is a freeform “context hints” field, where you describe the conversations or topics your product might be relevant to.

Early testing suggests a more literal, keyword-style approach to this field currently tends to outperform long, conversational prompts — though this is a fast-moving platform, and what works today may shift as OpenAI refines its matching. Since there’s genuinely no established playbook yet, this is an area worth testing directly with your own product and audience rather than assuming best practices from other platforms will transfer over.

Step 5: Create your ads

Ads appear at the bottom of a ChatGPT conversation, clearly labelled as sponsored content. There are two formats: static ads (a headline, description, and image) and product ads, which pull from a submitted product feed and are gradually gaining more detail like pricing and ratings.

Character limits are tight — headlines cap out at 50 characters and descriptions at 100 — so the copy has to work hard in a very small space. This is exactly the kind of constraint where experienced copywriting makes a measurable difference, since there’s very little room for a weak headline to hide.

Step 6: Set up measurement before you launch, not after

OpenAI provides a tracking pixel similar to what you’d install for Google or Meta Ads, letting you fire conversion events for specific actions like button clicks or completed purchases. It’s worth adding UTM parameters to every destination URL as well, so you can cross-reference ChatGPT Ads traffic against your own analytics rather than relying solely on in-platform reporting. If your Google Analytics setup isn’t already capturing campaign-level detail cleanly, it’s worth tightening that up before you start spending, since it’s much harder to reconstruct performance data after the fact.

Step 7: Optimise with limited visibility

This is the part that will frustrate marketers used to granular reporting. ChatGPT Ads currently only report performance by country, even though you can target more specific locations, and you get no visibility into the actual conversations or themes that triggered your ad to show. In practice, this means treating early campaigns the way many marketers treated Google’s early Performance Max rollout — largely trusting the system’s matching while testing variations in your ads and context hints to see what moves the needle.

Where ChatGPT Ads fits into your channel mix?

Because ads show up while someone is actively researching a topic in conversation, ChatGPT Ads behave more like search intent advertising than social advertising — closer in spirit to your Google Ads or PPC management strategy than to a paid social campaign. If you’re already seeing meaningful organic traffic referred from ChatGPT, that’s a reasonable signal the paid side of the platform is worth testing too. Worth noting: running ads doesn’t influence whether your brand shows up organically in ChatGPT’s responses, so the two remain separate levers.

Early data comparing traffic quality is mixed — engagement time and conversion rates from ChatGPT Ads traffic have generally trailed well behind equivalent Google Ads traffic in early tests. That doesn’t necessarily make it a poor channel; it means the audience and intent are different, and the real question is what it costs you to acquire a customer through it compared to your other channels.

Should you test it?

For most businesses, yes — cautiously. The minimum budget is low enough that testing doesn’t require a major commitment, and that cost of entry will likely only increase as more advertisers join the platform and competition for placements grows. Go in expecting to invest real time into experimentation rather than expecting an established playbook, since the platform is still young and OpenAI continues to roll out changes to targeting, reporting, and ad formats.

Conclusion

Between the unfamiliar targeting model, tight ad copy constraints, and limited reporting, ChatGPT Ads reward marketers who set things up carefully the first time rather than iterating blindly. If you’d rather have this managed alongside your existing paid and organic strategy instead of running it as a standalone experiment, the Rank My Business PPC management team can fold ChatGPT Ads testing into a broader approach that’s already tracking what’s actually driving results for your business.