MEIKAIInterparfums on ChatGPT Ads: 45.5% Visibility Lift
Criteo has published a case study of Interparfums' first ChatGPT Ads campaign. Interparfums led the campaign. Criteo, OpenAI's first advertising technology partner, bought and served the inventory. Meikai measured how the ads affected brand visibility within the conversations.
The measurement design makes this case worth examining. A before-and-after visibility chart cannot, on its own, show whether a ChatGPT Ads campaign caused a change. Interparfums used a stronger comparison: the same prompt set with ads and without them.
In summary: across May and June 2026, combined brand visibility was 45.5% higher when ads appeared. Interparfums accounted for 16.6% of the observed ad space and achieved a strong advertising presence. Targeting alignment was 2.8 times higher on category prompts. According to Criteo, Interparfums' brands had not previously reached this level of total visibility.
Disclosure. Meikai was Interparfums' measurement partner for this campaign. The figures in this article were published by Criteo and measured by Meikai. Meikai does not sell ChatGPT inventory or receive a commission on media spend; Criteo participates in the media buying. This article is therefore an interested vendor's analysis of a public case, not an independent audit of Criteo's buying operation.
Campaign design
Interparfums, Inc. develops, manufactures and distributes prestige fragrances through an international network covering more than 120 countries. It joined Criteo's inaugural ChatGPT Alpha without predefined KPIs. The aim was to understand how a labelled placement in ChatGPT performed for a fragrance portfolio, then decide whether the channel was worth developing further.
This was a disciplined way to approach a format that had barely existed at the start of the year. The costly alternative would have been to treat ChatGPT like paid search: bid on the most obvious category, then attribute every change in a visibility dashboard to media spend. Criteo and Interparfums designed the campaign to measure that effect instead of assuming it.
What we measured
Interparfums asked Meikai to compare brand visibility with and without ad delivery. Targeting alignment on category prompts formed part of the same brief. The design was deliberately simple: keep the prompt set fixed and observe the conversations according to whether a labelled ad appeared. The difference measures the incremental effect of the placement. It avoids drawing conclusions from a before-and-after chart that may have moved because a cited page changed, a model evolved or a competitor published new material.
This is the method we advocate in our writing on generative engine advertising. Paid delivery figures should come from the ad platform. Visibility should be compared across the same prompts with and without ads. Combining an ad impression and an organic citation in one score removes the distinction that matters most. OpenAI states that ads in ChatGPT are labelled, displayed separately from the answer and do not influence organic responses. GEA spend cannot buy a GEO citation. It can, however, increase the brand's presence in the experience a person actually sees.
What the numbers say
Criteo published three campaign figures, all from the activity in May and June.
| Result | What it means |
|---|---|
| +45.5% combined visibility when ads appeared | On the same prompt set, total brand presence in ChatGPT was substantially higher with ads than without them. |
| 16.6% share of ad space, with a strong advertising presence | In the measured conversations where ads could appear, Interparfums occupied a meaningful share of the labelled placements. |
| 2.8x higher targeting alignment on category prompts | The campaign matched particularly well with unbranded category questions, where the assistant builds its shortlist. This indicates more precise targeting than a broadly defined media buy. |
The campaign also surfaced more than ten themes associated with the brands, ranging from luxury positioning and heritage to gifting appeal and iconic status. This indicates how the brands were portrayed, not simply whether they appeared. For a fragrance house, a passing mention is not equivalent to being presented as a gift, a heritage brand or an icon. These themes help creative and media teams understand the buying situations in which the ads appeared.
What this does not prove
A public case study has important limits.
- It does not prove that ads changed organic answers. Combined visibility includes the labelled ad placement itself. OpenAI's policy states that advertising does not influence ChatGPT's organic responses. The 45.5% figure therefore describes visibility in the experience containing the ad, not an increase in GEO citations.
- It covers one Alpha campaign. The activity ran for two months, involved one advertiser and formed part of Criteo's inaugural ChatGPT programme. The result is real, but it is neither a benchmark nor a rate card.
- The public account does not disclose the sample, markets or brand mix. Interparfums is a multi-brand group. Criteo does not identify the participating fragrance houses or the prompt volume behind the percentages. We will not speculate about those details.
- Share of ad space is not the same as share of voice. 16.6% describes presence in the labelled inventory, not Interparfums' share of organic mentions against competitors.
What other brands can reuse
The headline figure is 45.5%, but the most useful part of the case is its measurement design. It suggests four practices that Ad Planner can build into the next campaign rather than apply only during the final review.
- Compare the same prompts with and without ads. A before-and-after visibility chart does not measure incrementality. We made this point when we launched Ad Planner as ChatGPT Ads opened in Europe. This case provides a practical example.
- Start with category prompts. A branded question already names the company. A category question leaves the assistant to build its own shortlist. The 2.8x alignment figure shows that the campaign reached this more demanding set of conversations.
- Analyse the themes, not only the overall increase. Heritage, gifting, luxury and iconic status are useful inputs for the next campaign brief, not decorative labels for a presentation.
- Keep media buying metrics separate. Impressions, clicks, CPC and conversions should still come from the ad platform. The visibility layer answers a different question: when the ad appeared, was the brand more present in the conversations that mattered?
Why planning comes next
Interparfums validated the measurement approach. The next opportunity lies earlier, when the media plan is being prepared. Without that planning, brands may bid on the most obvious category without knowing whether the spend addresses a genuine gap. Ad Planner analyses the same GEO data used in this case and turns it into a plan before the first impression. It identifies where ads already appear, where the brand's organic presence is weak and where paid media would duplicate conversations the brand already wins.
Planning the buy and serving the ad are separate functions. Criteo buys the inventory. Ad Planner helps decide whether the placement is worth buying and what it should contain. It produces four outputs in sequence.
- Opportunity selection. It ranks the buying intentions for which ChatGPT already serves ads, using ad penetration and the brand's organic gap. Category prompts matter because the assistant builds its shortlist there. They are also where Interparfums recorded 2.8x higher targeting alignment.
- A decision for each opportunity: Buy, Defend, Organic first or Monitor. Some recommendations advise against spending. That matters when inventory is limited and priced at a premium. Ad Planner creates value through the budget it preserves as well as the budget it deploys.
- An ad group ready for launch. It includes context hints in the language of the monitored prompts, a suggested maximum bid informed by organic strength and chat cards based on real products. The themes surfaced by this campaign, including heritage, gifting, luxury and iconic status, are precisely the situations those hints should describe.
- Measurement using the same design. Paid metrics come from the ad platform, while visibility is compared with and without ads on the same prompt model. The Alpha campaign required a manually designed protocol. Ad Planner makes that protocol the default for each campaign.
The channel is still emerging, which creates room to test before buying patterns become established. Brands that enter the auction with a plan can address contexts their competitors have not yet mapped. Meikai Ad Planner for ChatGPT is available as a standalone product or alongside a monitoring plan. To see which contexts we would recommend buying, defending or leaving alone for your brand, talk with our team.
Method and sources
The campaign details and three published figures come from Criteo's Interparfums case study, consulted at the time of publication. The company background on Interparfums comes from the same page. Meikai's role as measurement partner is stated there and confirmed here. Statements about the labelling of ChatGPT Ads and their lack of influence on organic responses follow OpenAI's published advertising documentation. Descriptions of Meikai's capabilities reflect what was documented on this website at publication.
Frequently asked questions
What results did Interparfums see from its ChatGPT Ads campaign with Criteo?
Criteo reports that combined brand visibility was 45.5% higher when ads appeared. Interparfums accounted for 16.6% of the observed ad space, with a strong advertising presence, and targeting alignment was 2.8 times higher on category prompts. The campaign ran in May and June 2026 as part of Criteo's ChatGPT Alpha. According to Criteo, Interparfums' brands had not previously reached this level of total visibility.
How did Meikai measure the Interparfums ChatGPT Ads campaign?
Interparfums asked Meikai to compare brand visibility with and without ads on the same prompt set. Targeting alignment on category prompts formed part of the same brief. This design isolates the incremental effect of the labelled placement, unlike a before-and-after visibility chart that may move for unrelated reasons.
Does a ChatGPT ad increase organic AI visibility?
No. OpenAI states that ads in ChatGPT are labelled, displayed separately from the answer and do not influence organic responses. The increase in combined visibility includes the ad placement in the experience the person saw. It should not be interpreted as a gain in organic citations or generative engine optimisation.
Who bought the media? Who measured it?
Criteo bought and served the ChatGPT Ads inventory as OpenAI's first advertising technology partner. Meikai was Interparfums' measurement partner and compared visibility with and without ads. Meikai does not sell the inventory or receive a commission on media spend.
What is Meikai Ad Planner? How does it relate to this case?
Ad Planner is Meikai's planning product for ChatGPT Ads. It ranks the buying intentions for which ads already appear, then classifies each opportunity as Buy, Defend, Organic first or Monitor. It prepares an ad group and measures the increase in visibility by comparing the same prompts with and without ads, as in the Interparfums case. Criteo buys the inventory; Ad Planner helps decide where that inventory is worth buying. It is available as a standalone product or alongside a monitoring plan.