Retail data study
In four UK online shops running both, visits from ChatGPT ads produced a tiny fraction of the sales that Google ad visits did. The free links inside ChatGPT's answers told a very different story.
Somewhere in a conversation about sports kit or pendant lights, an advert now waits. The pitch is easy to believe. Shoppers already ask the chatbot what to buy, so a retailer who pays to appear at that moment should catch buyers at the point of decision. With Black Friday and Christmas budgets being set now, plenty of online shops are weighing whether to move money across.
The early numbers from UK shops argue for caution. In four online shops selling sports equipment and LED lighting, each tracked from the day its ChatGPT ads went live, a visit from a ChatGPT ad brought in about ten pence in sales, against more than three pounds for a visit from a Google ad over the same days.
The surprise sits elsewhere in the same accounts. Shoppers who clicked an ordinary, unpaid link in a ChatGPT answer spent freely, and across a wider group of 17 UK online shops those free visits were worth more on average than a Google ad visit. On this evidence, ChatGPT is already sending buyers; the paid ads are the part that has yet to prove it can.
Two sales from thousands of ChatGPT ad visits
Start with the plainest comparison. Per visit, Google ad traffic was worth roughly thirty-three times as much.
Two sales is a very thin base. The size of the gap is clear; the exact ChatGPT ad figure could move a lot with a handful more orders. These tests are also short, because ChatGPT ads are new: one shop ran its ads for four days, another for 41, and none of the windows includes Black Friday.
Shop by shop, the pattern repeats. No shop got close to its Google ad return from its ChatGPT ads. An LED lighting retailer (Shop F) sent more than 1,600 ChatGPT ad visitors to its site over a month and recorded no sales from them, while its Google ads earned close to five pounds a visit.
| Shop (sector) | Days of ChatGPT ads | ChatGPT ad visits | ChatGPT ad sales | ChatGPT ads, £ per visit | Google ads, £ per visit |
|---|---|---|---|---|---|
| B (sports equipment) | 6 | 292 | 1 | 0.54 | 4.66 |
| C (sports equipment) | 41 | 652 | 1 | 0.19 | 2.51 |
| D (sports equipment) | 4 | 122 | 0 | 0.00 | 1.86 |
| F (LED lighting) | 32 | 1,629 | 0 | 0.00 | 4.87 |
Clicks that returned pennies
The obvious defence of ChatGPT ads is price. Their clicks averaged 27p against 45p on Google. A low click price only matters if the clicks sell, though: every pound spent on ChatGPT ads brought back £0.34 in sales, against £7.82 per pound on Google Ads. Put another way, a ChatGPT ad click was worth about a penny at the rate Google ads were returning in the same shops.
Treat that return figure as a range around a clear direction. Sales are credited to the last visit before the purchase, so a shopper who first met a shop through a ChatGPT ad and came back days later through a Google ad counts as a Google sale. Some ChatGPT ad influence could be hiding in Google's column. It would need a great many hidden buyers to close a gap this wide.
The free links in ChatGPT answers did the selling
Here the received wisdom turns over. The same four shops, on the same days, also had visitors arriving from unpaid links in ChatGPT's answers, where the chatbot pointed to the shop as a source. Those visits brought in £3.74 each from 36 sales, against ten pence for a ChatGPT ad visit. That puts free ChatGPT traffic in the same territory as Google ads, and dozens of times more valuable than the paid version from the same chatbot.
| Traffic source (same four shops, same days) | Visits | Sales | Sales value per visit (£) |
|---|---|---|---|
| ChatGPT ads | 2,695 | 2 | 0.10 |
| Google ads | 72,720 | 1,362 | 3.31 |
| Free ChatGPT answer links | 1,914 | 36 | 3.74 |
So the audience is there and willing to spend. What separates the two is how the shopper arrived: recommended in an answer, or shown an advert.
Across 17 shops, free ChatGPT visits outsold Google ads per visit in most
Widen the view to 17 UK online shops, from streetwear and paint to water tanks and coffee, over September and October. Free ChatGPT links brought 11,085 visits and 271 sales, worth £3.76 a visit, against £2.53 for a Google ad visit; free ChatGPT visits earned more per visit than Google ads in 11 of the 17 shops.
| Measure | ChatGPT ads | Google ads | Free ChatGPT answer links |
|---|---|---|---|
| Four shops: visits | 2,695 | 72,720 | 1,914 |
| Four shops: sales | 2 | 1,362 | 36 |
| Four shops: sales value (£) | 281.55 | 240,832 | 7,165 |
| Four shops: £ per visit | 0.10 | 3.31 | 3.74 |
| Four shops: share of visits that bought | 0.07% | 1.87% | 1.88% |
| Four shops: spend (£) | 820.85 | 30,797 | not applicable |
| Four shops: cost per click (£) | 0.27 | 0.45 | not applicable |
| Four shops: sales per £1 spent (£) | 0.34 | 7.82 | not applicable |
| 17 shops: visits | not measured | 483,297 | 11,085 |
| 17 shops: sales | not measured | 8,446 | 271 |
| 17 shops: £ per visit | not measured | 2.53 | 3.76 |
| 17 shops: share of visits that bought | not measured | 1.75% | 2.44% |
Keep that in proportion. Free ChatGPT visits are a small slice of each shop's traffic, so a few large orders can swing a shop's figure. One office-supplies shop's result rests on a single sale. Google ads still deliver far more visitors and far more sales in total. The free links are valuable per visit; they are nowhere near a replacement for the volume.
What this means for UK businesses
Test ChatGPT ads small, and judge them on sales. Low-priced clicks look attractive on a dashboard. Set a capped test budget, tag every ad link, and measure purchases and revenue.
Look at the free ChatGPT visits you already get. Many shops are receiving buyers from ChatGPT answers without paying for them. Find out how many, what they buy, and which pages the chatbot links to.
Remember the limits. These are short, early tests outside the busiest weeks of the year. Re-check after the peak season, when there is more data.
When the next advert appears beside an answer about sports kit or pendant lights, the shops in this data suggest the better prize may be the link in the answer itself.
How we built it
The figures come from the GA4 analytics of UK online shops managed by the agency behind this analysis, with ad spend taken from each shop's OpenAI and Google ad accounts. Visits, sales and spend are measured counts taken directly from those accounts.
ChatGPT ad visits are GA4 sessions tagged chatgpt / cpc. Google ad visits are google / cpc, covering Search, Shopping and Performance Max. Free ChatGPT visits are sessions arriving from chatgpt.com links in answers. Sales are GA4 purchases and revenue credited to the last visit before the purchase (last click).
The ads comparison covers four shops (sports equipment and LED lighting) from 23 September to 2 November 2026, each measured from the day its ChatGPT ads started, so test lengths run from four to 41 days. One shop was left out for having fewer than 50 ChatGPT ad visits, and three because their GA4 does not record Google ad sales properly. Spend comparisons use the four shops where both ad accounts report spend; Google spend covers every campaign in each account. Only two ChatGPT ad sales were recorded, so the direction of the gap is clear while the exact rate is uncertain. No window includes Black Friday.
The free-links comparison covers 17 shops from 1 September to 2 November 2026. Free ChatGPT visits are a small share of each shop's traffic, and per-shop results swing with a few large orders. Shops are identified by letter and sector only.
