For ecommerce brands that want to be first on a billion-dollar ad market
AI advertising is the future, and right now you can be one of the first advertisers there. Launch your first ecommerce campaign in 48 hours with the proven Hint Stacking Engine® and get the cheapest results in the market.
30-day money-back guarantee. Instant access, one-time payment, no subscription.
Small skincare brand, about $40k a month, and Meta has been eating us alive since spring. Had my first campaign live before lunch the next day. $9,412 in sales from $1,860 in ad spend over the first 21 days. CPC came in around $2.80, which I haven't seen on Meta since 2023.
We sell ayurvedic supplements, mostly to the US. I took the five questions our support inbox gets every day and turned them into hints. Two weeks in: 41 orders, $2,340 in revenue, $610 spent. The AOV is 30% higher than our Meta orders.
Outdoor gear brand doing about $180k a month. I thought $27 meant it would be fluff. It's not. August 4 to September 3: $31,180 in tracked sales on $4,920 spend. CPC averaged $3.40. Our Meta CPC is now above $2 and converts at a third of the rate.
Handmade ceramic mugs on Shopify, one person business. I had the video on one screen and ads.openai.com on the other and just copied every click. Went live at 4pm. First order came in at 11:40pm the same night. Ten days later I'm at 27 orders on $250 spent.
Men's grooming brand, UK, around £90k a month. Shifted 20% of our Meta budget across, nothing else changed. After four weeks: £19.60 CPA on ChatGPT vs £43 on Meta for the same SKUs. Conversion rate went from 1.1% to 2.9% once we built hints around the question instead of the product.
Pet supplements, Canada, tiny brand. Took the exact same $25 a day to ChatGPT Ads. Took me about 90 minutes including watching the video. Month one: $4,108 in sales from $775 spend. Same budget that was getting me $1,200 on Meta.
Full transparency, I bought this to prove it wouldn't work. Coffee subscription brand, nine years in ecom. We launched with 14 hints across the three layers and $40 a day. By day 18: $5,620 in first orders on $1,440 spend, 3.9x ROAS and that's before subscription LTV.
Modest jewelry brand, Australia. Set up the campaign the next morning during my daughter's swim lesson. Week three we did $11,300, our best week since we launched in 2021. About $6,800 of it came from ChatGPT Ads on $1,100 spend.
Home fitness equipment, US and Canada, roughly $220k a month. Google Shopping CPC hit $4.90 in July. First 30 days: $18,240 in sales, $3,150 spend, average CPC $2.95. That's 40% cheaper than Google Shopping for the same products, with a similar conversion rate.
Baby products, US, about $60k a month. Every non-Meta channel we've tried since 2023 has lost money. After 25 days: $7,860 in sales on $2,080 spend, first profitable channel outside Meta since 2023. Adding budget every week.
I thought so too. So before you read another line, here is the one number that matters.
In OpenAI's ads beta, product cards pulled between 4 and 7 percent click-through. Not on a hand-picked winner. Across roughly 3,850 placements measured by Criteo, OpenAI's first ad-tech partner.
Google Search, the most competitive auction on the planet, averages 6.4 percent. It took twenty years and every agency in the world fighting over the same keywords to get there.
ChatGPT's auction has been open to the public for four months.
That gap does not stay open. It never does. It closed on Facebook in 2014, on Instagram in 2017, on TikTok in 2022. Each time, the brands that were in during the first year bought customers for a fraction of what everyone else paid later.
You are reading this during the first year.
If you run a store, you already know the feeling. You raise the Meta budget by 30 percent and the cost per purchase climbs with it. You add a creative, it works for nine days, then the auction eats it. You try Google Shopping and discover that every competitor with a bigger margin is already there, outbidding you on your own product name.
Nothing is broken. You are simply fighting for attention in the two most crowded rooms on the internet, against people who have been in those rooms for a decade.
Meanwhile, roughly 800 million people a week ask ChatGPT what to buy. "Best running shoes for flat feet." "Gift for a dad who fishes." "Which air purifier for a 400 square foot apartment." Real purchase intent, typed out in full sentences, and until May of this year there was no way to put a product in front of it.
Now there is. And almost nobody is using it, because almost nobody knows how the targeting works.
Why do you actually want a new channel?
Not for the channel. You want to stop being a hostage. Hostage to one auction, one algorithm update, one account review that can switch your revenue off on a Tuesday. A second channel where you are early is not a growth hack. It is the difference between running your business and Meta running it for you.
ChatGPT Ads has no keywords. No interests. No lookalikes. No demographics. The only targeting lever you get, beyond geography, is a text box at the ad group level called context hints. You describe the conversations where your product belongs, and the system decides when your card shows up.
Most people write one sentence in that box, spend $300, see nothing and conclude "ChatGPT ads don't work". They never find out that the way you fill that box decides whether you pay $3 a click or $9, and whether you get shown at all.
The Hint Stacking Engine® is the method I built for that box. Three layers, stacked in a fixed order:
Short buying-intent fragments the way a shopper types them, not the way a marketer writes them. In published tests, keyword-style hints beat conversational descriptions on both delivery and cost per click.
The exact questions ChatGPT gets asked right before a purchase in your category. This is where product cards get their 4 to 7 percent.
The "is X worth it", "X vs Y", "does X work for" conversations. Cheapest impressions in the system, because nobody stacks them.
One stack per product category, one ad group per stack, product feed as the creative, and a bid rule that keeps you above the delivery floor without overpaying. That is the whole engine. It takes 60 minutes the first time and 15 minutes every time after.
No slides. No theory module. I sit down with a store, open a blank ads.openai.com account and go from zero to a live campaign while you watch. Every chapter is time-stamped so you can pause, copy, and continue.
The setup screens OpenAI does not explain.
Product feed ads are the strongest format on the platform. They are also the clunkiest to submit.
The heart of the engine. Watch me write all three layers for a real category, then reuse the pattern for yours.
Where most first campaigns die quietly.
You will not see which conversations triggered your ad. Here is how to optimise anyway.
Fifty ready-to-paste context hints across ten ecommerce categories: apparel, supplements-adjacent wellness products that are allowed, home, pets, beauty, outdoor, kitchen, baby, hobby and gifts. Each one written in all three layers. Swap the product name, paste, done.
Twelve checks to run on your product feed before you submit it, in the order the platform rejects them. Fixes the "blank card" problem before it costs you a week.
One page. What to bid in September, what to do when CPMs start climbing in October, and when to stop and hold. Built from the public benchmarks and the 20/3 rule I use on Meta accounts spending $2,000 a day.
My name is Alexander Wozniak. I build and scale paid acquisition for ecommerce brands and online experts. I am not a ChatGPT guru. Until May there was no such thing. I am a media buyer who has spent the last ten years inside Meta and Google auctions, scaling accounts past $2,000 a day, and who has watched every "new channel" go from cheap to expensive on the same predictable curve.
ChatGPT Ads is a billion-dollar industry from day one. OpenAI is targeting $2.5 billion in ad revenue this year, and the brands that learn the auction now will be the ones still paying cheap clicks when everyone else arrives.
When OpenAI opened self-serve ads, I did what I always do with a new auction: opened an account, put real money in, and worked out where the lever was. The lever is the hint box. The Hint Stacking Engine® is what I found works in it. I am selling it for the price of a lunch because I want the people who buy it to be the ones who are actually in the auction in Q4, sending me screenshots.
Take the full 30 days. Follow the walkthrough, build your stack, switch the campaign on. If at any point you decide it is not for you, or you simply could not get a campaign live, send one email and I refund the whole $17.
You decide whether it was worth it. Not me.
Roughly 500 new advertisers enter the ChatGPT auction every month right now. OpenAI has said out loud that it wants $2.5 billion from that auction this year, and Q4 is where ad budgets go to die. The CPMs that clear at $25 in September will not clear at $25 in November. They never have, on any platform, in any year.
You can learn this channel now for $17 and thirty minutes a week, while the clicks are cheap and the competition is thin. Or you can learn it in February, from an agency deck, at whatever the auction costs by then.
Hint Stacking Engine®
30-day money-back guarantee. Instant access by email. Launch price, goes to $47 soon.
You can open an advertiser account and run campaigns in the United States, Canada, the United Kingdom, Australia, New Zealand, Japan, South Korea, Mexico and Brazil. Your store can be anywhere as long as you ship to and target one of those markets.
The platform minimum is $25 a day per campaign. The walkthrough is built around exactly that number for the first 30 days.
An exportable product feed and the ability to add a pixel. Both are covered on screen. WooCommerce and BigCommerce work the same way.
Not any more. OpenAI switched Instant Checkout off in March 2026 to focus on discovery. The ad sends the shopper to your store, where your own checkout does the work. That is a feature: you keep the customer data.
Honest answer: nobody has reliable conversion benchmarks yet. What is public is click-through of 4 to 7 percent on product cards, CPCs of $3 to $5 in retail, and one early case where a ChatGPT click converted worse than a Google click on the same site. Cheap attention, unproven conversion. You buy this to run the test while it is cheap, not because someone promised you a ROAS.
No. It is one recording of one campaign being built, plus the three files you need to build yours. If you want a 12-module academy, this is not it.
30-day money-back guarantee. Instant access, one-time payment.
Results depend on your product, margin, feed quality and the auction on the day you run. The numbers quoted on this page are published third-party benchmarks (Criteo, eMarketer, OpenAI documentation, Search Engine Land) from 2026 and are not a promise of your results. ChatGPT is a trademark of OpenAI. This product is not affiliated with or endorsed by OpenAI.
Where should I send the three things behind a $50,000 month?Drop your details below
Check your inbox in a minute.
3
2
4
2
3
2
3
2
4
3
