ChatGPT Ads Manager Is Open: Should Ecommerce Brands Test It?

Most ecommerce brands should test ChatGPT ads only after their product feed is connected and their conversion tracking works, and then with a small, fixed budget they can afford to learn from. There is no minimum spend any more and you can bid per click, but there are no reliable public benchmarks yet, so treat any first test as research, not as a channel you scale into Q4.

Here is what the platform actually offers as of October 2026, who we think should try it, and how to measure it without fooling yourself.

What is ChatGPT Ads Manager, and what changed in 2026?

OpenAI’s ad business moved fast this year.

• February 2026: a pilot for large advertisers. One agency that tracks pricing, Top Growth Marketing, reports it was sold at a $60 CPM with minimum commitments of $200,000 to $250,000.
• May 5 to 6, 2026: a self-serve beta Ads Manager opened to US advertisers. Axios reported that cost-per-click bidding was added (the pilot had been CPM only) and the former $50,000 minimum spend was removed. Search Engine Journal confirmed CPM bidding stays available alongside CPC, and that measurement now includes a pixel and a Conversions API that can track purchases, sign-ups and leads.
• Late August 2026: product carousel ads launched.
• September 4, 2026: OpenAI added an Ads Manager plugin that lets you build and edit campaigns in plain language inside ChatGPT, editable custom audiences, more pixel fields (hashed phone, name, region, postal code), total-budget pacing, and a view-through conversion model that was still rolling out. ChatGPT shopping results also became exclusive to brands with connected product feeds, and ads expanded to Europe, India, the Middle East and North Africa (Search Engine Roundtable).

OpenAI says ads are available in over 40 countries, are “always clearly labeled and separate from ChatGPT’s answers,” and that “advertising does not influence the answers ChatGPT provides” (OpenAI, August 31, 2026).

That last point matters for planning. Ads and organic product recommendations are two separate systems. Buying ads does not get you into the organic shopping results, and the organic results need a feed regardless. We cover the organic side in how to get your products recommended by ChatGPT.

How much do ChatGPT ads cost?

Nobody outside OpenAI has solid numbers yet. Here is what is public, and how much weight to put on each.

The same agency says ads show to users on the Free and Go tiers, not to paid Plus or Pro subscribers, and that some categories (it lists health, financial services, political, dating and adult content) are restricted. We have not seen OpenAI’s own policy page confirm that list, so check the current ad policies inside Ads Manager before you plan creative for a regulated product.

For context, a $3 to $5 click is not cheap for most Amazon categories we manage. Run the maths before you start. For a $30 product with a 35% margin ($10.50 of gross profit per order), a $4 click needs a conversion rate above 38% just to break even on the first order. No cold traffic converts like that. ChatGPT ads only make sense if you value something beyond the first order: repeat purchases, a high average order value, or new-customer acquisition you can’t get elsewhere.

Which ecommerce brands should test ChatGPT ads before Q4?

We would put brands into three groups.

Test now: brands with a DTC site that converts well, an average order value comfortably above $75, a connected product feed and working pixel or Conversions API tracking. Higher-consideration products (things people research and compare) are a better fit than impulse buys, because that is what people use ChatGPT for.

Wait: brands whose sales are almost all on Amazon. Your ad click has to land somewhere you can measure. If the landing page is your Amazon listing, you lose most of the conversion data, and as we explain in why ChatGPT rarely sends traffic to Amazon, Amazon is walled off from ChatGPT in other ways too.

Skip for now: low-price impulse products, brands without clean tracking, and brands that haven’t fixed their own listings and site. Spend there buys clicks on a page that won’t convert them. That is the pattern we wrote about in PPC is an amplifier, not a fix: a client increased ad spend 25% after a listing rebuild and net profit rose 81.5%, because the rebuild came first.

How should you measure a ChatGPT ads test honestly?

This is where most early tests will go wrong.

Research from Similarweb found that shopping journeys which start in ChatGPT and end with a purchase somewhere else happen about 5 times more often than journeys with a direct click to a retailer. Last-click reporting will undercount ChatGPT. The view-through model OpenAI is introducing will probably overcount it. Neither is the truth.

A simple setup that gets closer:

1. Install the pixel and the Conversions API before you spend anything. Test that a purchase fires correctly.
2. Tag every ad URL with UTM parameters so the clicks show up separately in your own analytics, not just in Ads Manager.
3. Record a baseline for four weeks before launch: branded search volume in Google Search Console, direct traffic, and new-customer orders.
4. Run the test for a fixed period (we suggest six weeks) at a fixed budget. Do not change the budget mid-test.
5. Compare three numbers at the end: in-platform conversions, conversions in your own analytics, and the change in branded search and new-customer orders against the baseline.
6. Hold a short pause of two weeks afterwards. If branded search and new-customer orders drop when the ads stop, the ads were doing something. If nothing moves, they probably weren’t.

If you want a fuller view of what AI traffic is worth once it lands, we pulled together the 2026 data in does AI traffic convert.

What starter budget makes sense?

Top Growth Marketing suggests $1,500 to $5,000 a month for an initial test. That is a reasonable range for a mid-sized DTC brand, with one caveat: if the agency’s $3 to $5 CPC estimate holds, $1,500 buys roughly 300 to 500 clicks a month. That is enough to see whether people engage, and not enough to judge return on ad spend with any confidence.

Set the budget as an amount you have decided to spend on learning. If you need the test to be profitable inside six weeks, it is the wrong test for you this quarter, and the money is better spent on Amazon, Google Shopping or the channels where your measurement already works. Our Google Ads management clients are mostly testing Shopping ads in Google’s AI Mode first, where the feed and the tracking are already in place. We cover that in selling in Google AI Mode.

What we'd do this week

1. Connect your product feed. You need it for organic shopping results anyway, and it is required for carousel ads. If you are on Shopify, check Shopify Agentic Storefronts first.
2. Install the ChatGPT pixel and Conversions API on your DTC site and test a purchase event.
3. Write down your baseline numbers for branded search and new-customer orders.
4. Register for Ads Manager and check what it shows for your category: allowed formats, policies, and suggested bids. Those in-platform numbers beat any agency estimate, including the ones above.
5. Decide, in writing, the budget, duration and the result that would make you continue. Then do not touch it until the test ends.

All figures are as of October 2026. The platform is still in beta and changing monthly, so confirm pricing, minimums and policies inside ChatGPT Ads Manager before you launch.

When you judge ChatGPT ads against Meta, use the same attribution rules for both. Our breakdown of Meta’s attribution changes and ROAS explains why the comparison is often skewed.