One product went from 27 to 90 conversions a week after its brand turned on ChatGPT ads, according to Jackson Blackledge (@blvckledge), who has been testing the ads for ecommerce clients and shared early findings on X in September 2026.
been testing chatgpt ads for some clients still early but the results have been much better than i expected some of our findings: 1. chatgpt ads are looking strong for top of funnel. ROAS tends to be much higher on first-click attribution than on last-click. 2. there’s a strong correlation between chatgpt ads and higher sales on other channels. one product jumped from 27 to 90 conversions a week after we turned them on. 3. product feed campaigns often deliver better ROAS than standard campaigns (2.28x vs. 1.6x in one account). scaling product feed campaigns ads is quite difficult though.
— Jackson Blackledge (@blvckledge) · View on X
The clients are not named, and Blackledge calls these early insights that might change over time.
Before
The post gives two starting points:
- One product: 27 conversions a week before ChatGPT ads were switched on.
- Standard ChatGPT ad campaigns in one account: 1.6x ROAS.
It does not give spend, CPC or revenue figures.
What they changed
- Used ChatGPT ads for top of funnel. Blackledge says ROAS looks much higher on first-click attribution than on last-click.
- Ran product feed campaigns next to standard ones. Feed campaigns often delivered better ROAS, though they were hard to scale.
- Bid on clicks instead of reach. Reach campaigns produced very high CPMs and CPCs.
- Started CPCs well below the suggested bids. The team used its Google Shopping CPCs as the benchmark.
- Picked products that already sell on Google Shopping. Those products tended to do better on ChatGPT ads as well.
- Wrote calm, specific copy. The ad appears after the organic answer, so the user has already read a detailed reply, and aggressive or clickbait copy did not work.
After
| Metric | Before or standard | After or product feed |
|---|---|---|
| Weekly conversions, one product | 27 | 90 |
| ROAS, one account | 1.6x (standard campaigns) | 2.28x (product feed campaigns) |
Blackledge reports a strong correlation between ChatGPT ads and higher sales on other channels, which is where part of the 27 to 90 jump showed up. That makes it a correlation, and the post does not isolate the ads' own effect.
How can you test this on your store?
- Start with products that already sell well on Google Shopping.
- Bid on clicks and set CPCs from your Google Shopping numbers, below the suggested bids.
- Run one product feed campaign next to one standard campaign and compare ROAS.
- Track first-click and last-click ROAS together, plus total weekly sales of the advertised products across all channels.
- Write copy that adds something to the answer the user just read, and skip hype.
Frequently asked questions
Where does a ChatGPT ad appear?
After the organic answer, according to Blackledge. By then the user has read a detailed reply, which is why aggressive copy did not work in these tests.
Why do the results depend on attribution?
Blackledge says ROAS looks much higher on first-click than on last-click, and that the ads correlated with higher sales on other channels. Judging the ads on last-click alone would likely undercount them.
What is the biggest limitation right now?
Reporting. Blackledge says there is little data on search terms and audience demographics.
Sources
- Jackson Blackledge (@blvckledge) on X: https://x.com/blvckledge/status/2102020208915529997