ChatGPT Ads Meet HubSpot: The European Attribution Guide

Your customers are already inside ChatGPT. Now the ads are too.

ChatGPT ads are live for Free and Go-tier users across 31 European markets — the 27 EU member states plus Iceland, Liechtenstein, Norway, and Switzerland.

At this stage, buying runs through OpenAI's Ads Solutions team and its agency and technology partners. A self-serve Ads Manager for Europe has been announced as coming.

That gap — ads live, self-serve pending — is your preparation window.

Because here's what most HubSpot-managed teams will discover too late: if you tag a ChatGPT ad the way you tag a Google ad, HubSpot files the click under Paid search and blends it with your Google Ads data. Your ChatGPT spend disappears into the wrong bucket. And you'll never be able to say what the channel did.

In this guide, you'll learn:

• Where ChatGPT advertising stands, and how it works in Europe
• How the traffic shows up in HubSpot — organic and paid
• The classification trap, and the UTM convention that avoids it
• A three-stage plan to prepare, launch, and evaluate

Let's start with the current state of the channel.

Where do ChatGPT ads stand today?

ChatGPT ads are live in the US, Canada, Australia, New Zealand, the UK, Mexico, Brazil, Japan, South Korea, and 31 European markets — the 27 EU member states plus Iceland, Liechtenstein, Norway, and Switzerland. Ads reach only Free and Go tiers; Plus, Pro, Business, Enterprise, and Education stay ad-free.

The channel has matured remarkably fast.

OpenAI's pilot ran US-only, managed exclusively through agency holding companies (Dentsu, Omnicom, Publicis, WPP), at around a €60/$60 CPM with a roughly €200k/$200k minimum commitment.

Within months, that model was gone. The US now has a self-serve Ads Manager (ads.openai.com) with no minimum spend.

Along the way, OpenAI added:

• Lower pricing — CPMs around €25/$25, CPC bidding at roughly €3–5/$3–5 per click (with bid floors by category)
• Conversion-optimized bidding, geo-targeting, and custom audiences
• A first-party measurement stack — the OpenAI Pixel plus a Conversions API

Tens of thousands of marketers have now advertised on ChatGPT.

The forecasts agree on the shape of the curve: small now, material soon. eMarketer expects ads in AI chatbots to generate under $1 billion in their first full year — about 3% of total AI ad spend. Its separate forecast projects US AI-search ad spend growing from 0.7% of all search ad spending to $25.9 billion, or 13.6% of search ad spend, within a few years. Third-party trackers already put monthly ChatGPT ad spend in the $100-million range.

One important difference in Europe at this stage: buying is not yet self-serve. Campaigns run through OpenAI's Ads Solutions team, agency partners, and technology partners.

That's exactly why the preparation work below matters now.

How do ChatGPT ads work?

ChatGPT ads appear below a response, clearly labeled as sponsored and visually separated from the answer. Each unit carries a business name, image, headline, and short description with a destination link. Targeting is contextual — matched to the topic and intent of the current conversation — not keyword-based.

Per OpenAI's help documentation, the system matches ads to:

• The conversation's topic and intent
• Your landing page and ad copy
• Advertiser-provided "context hints" — not exact-match keywords
• Where personalization is enabled: broader signals from the user's ChatGPT experience

OpenAI supports CPM buying (Reach objective), CPC buying (Clicks objective), and conversion optimization.

Three stated principles govern the format:

1. Answer independence — ads never influence answers
2. Conversation privacy — advertisers get only aggregate performance data
3. User control

For a search marketer, the mental shift is real: you're not bidding on keywords. You're describing context, and the system decides when your ad fits the conversation.

What's different about ChatGPT ads in Europe?

In the EEA and Switzerland, ChatGPT ads launched contextual-only — no personalized targeting — and OpenAI asks for explicit consent before any personalization, rather than relying on GDPR "legitimate interest." Ads reach only Free and Go (€7.99/month) users.

Three regulatory frameworks explain why Europe came later and in a restricted form:

• GDPR / ePrivacy. OpenAI runs a two-tier model: contextual ads (current conversation topic, approximate location, device type, time of day, language) as the default, with personalized ads requiring a separate explicit opt-in. Personalization is not active in the EEA and Switzerland at this stage.
• Digital Services Act. OpenAI's DSA transparency disclosure reported approximately 159.1 million average monthly active recipients for ChatGPT search in the EU — far above the 45-million threshold. The European Commission is weighing whether ChatGPT search should be designated a Very Large Online Search Engine, which would trigger ad-transparency, ad-repository, and systemic-risk obligations.
• EU AI Act. Enforcement powers are now active. The Act's prohibition on manipulative or subliminal AI techniques raises unresolved questions about ads embedded in a trusted conversational flow.

For European marketers, this matters twice over.

First, the regulators are paying attention. National data-protection authorities across the EEA actively enforce consent rules — Norway's Datatilsynet recently fined the retailer Elkjøp NOK 20 million over customer-club consent practices, one of a steady stream of consent-related penalties across Europe. Consent discipline on your own landing pages and forms is not optional.

Second, the adoption backdrop is stronger than most teams assume. Eurostat's latest ICT survey (roughly 330,000 individuals aged 16–74 across 33 European countries) found roughly one in three EU individuals (32.7%) had used a generative AI tool in the previous three months.

The leading markets sit far above that: Norway at 56% (the highest surveyed), Denmark at 48.4%, Estonia at 46.6%, and Malta at 46.5%.

Your customers are already living inside ChatGPT — even if your competitors' marketing teams aren't yet.

How does ChatGPT traffic show up in HubSpot?

HubSpot classifies organic ChatGPT referrals under its native AI Referrals traffic source, which recognizes nine platforms — ChatGPT, Claude, Perplexity, Copilot, Gemini, Meta AI, Mistral, Poe, and Grok — with the platform in Drill-Down 1 and campaign parameters in Drill-Down 2. Paid ChatGPT ad clicks are a different story, covered below.

Organic ChatGPT referrals are the cleanest signal of the AI engines. OpenAI appends utm_source=chatgpt.com to ChatGPT search referral URLs, and the referrer is commonly chatgpt.com.

But treat that number as a floor, not a true count. Here's why:

Many AI clicks arrive with stripped referrers — mobile, in-app browsers, copied links, privacy features — and land in Direct traffic.

Most AI exposure never produces a click at all. Ahrefs' Brand Radar study of 31,000+ brand mentions (drawn from a 150-million-prompt database) found AI assistants link to the brands they mention only 10.7% of the time on the low end (Google AI Overviews) and 51.6% on the high end (Perplexity).

For every measurable click, there are often two or three brand exposures your analytics can't see.

Note: HubSpot and GA4 will disagree, systematically. HubSpot uses first-touch source assignment and strict bot filtering; GA4 uses last-non-direct click. GA4 typically shows higher ChatGPT numbers. Neither is wrong — they answer different questions. Set that expectation with stakeholders before the first report goes out.

Paid ChatGPT ad clicks behave differently again.

OpenAI's Ads Manager auto-appends its own click identifier, oppref=, to every ad landing-page URL (used for OpenAI's own measurement). It also exposes dynamic macros — {campaign_id}, {ad_group_id}, {ad_id} — you can map into your own UTM string.

The crucial detail: ChatGPT ad clicks do not automatically carry your UTM parameters. You must add them to the landing-page URL yourself, or every paid click lands in your analytics as generic referral or direct traffic. (If UTM discipline is new territory, start with our UTM tracking guide.)

How does HubSpot classify ChatGPT ad traffic? (The trap)

A ChatGPT ad click tagged with utm_medium=cpc will be classified as "Paid search" in HubSpot and blended with your Google Ads data — because HubSpot's Paid-search rule is evaluated before its AI Referrals rule, and the cpc token alone is enough to trip it. To keep ChatGPT ads visible as their own line, tag them with a distinct medium instead.

This is the finding every HubSpot-managed team needs to internalize before spending.

Here's the mechanism. HubSpot's traffic-source rules are applied in a fixed order — first match wins:

1. The Paid-search rule fires when any of utm_source, utm_medium, or utm_campaign contains "adword," "ppc," or "cpc."
2. The AI Referrals rule fires later, on recognized AI domains.

So a ChatGPT ad tagged utm_medium=cpc gets bucketed as Paid search — even though the referring domain is chatgpt.com, and even if utm_source=chatgpt. The rule doesn't require the source to be a recognized search engine. The token is enough.

And there's no official escape hatch:

• HubSpot's AI Referrals source is organic-only
• There is no "Paid LLM" category (a gap users have raised in HubSpot's Ideas forum)
• HubSpot has published no dedicated guidance on tracking paid AI ads

That forces a genuine trade-off, and practitioners split on it:

1. Tag with utm_medium=cpc — mirrors Google and GA4 conventions, but in HubSpot it hides ChatGPT ad spend inside Paid search.
2. Tag with a distinct medium — such as utm_medium=conversational-ai, plus utm_source=chatgpt and a clear utm_campaign, built via HubSpot's Tracking URL Builder. The traffic lands in "Other campaigns," where it can be isolated cleanly — at the cost of the "paid" label.

Our recommendation for a HubSpot-first stack: take option 2. Keep ChatGPT ads visible as their own line. Reconstruct the "paid" view through your utm_campaign naming convention plus OpenAI's own Ads Manager reporting.

There is currently no way to make ChatGPT ads sit in both a "paid" bucket and an "AI" bucket in HubSpot at once. Pick isolation over a misleading merge.

Setting up ChatGPT-ready tracking — UTM governance, source rules, and the reports that prove lift — is exactly the kind of foundation work that pays for itself before the first euro of spend. Talk to Superwork and we'll wire it into your portal before you launch — it's the same setup behind our ChatGPT Ads integration service.

How do you track ChatGPT ads in HubSpot? (Reporting that holds up)

Use HubSpot's Original Traffic Source and Latest Traffic Source properties with Drill-Downs 1 and 2 — they propagate from contacts to associated companies and deals, so a contact whose first touch was a ChatGPT ad or AI referral can be reported at the deal level.

For revenue attribution, HubSpot offers pre-built multi-touch models in the custom Attribution report builder: first touch, last touch, linear, U-shaped (40/20/40), W-shaped, and time decay. (How to choose between them is the subject of our marketing attribution guide.)

Revenue attribution requires Marketing Hub Enterprise; contact- and deal-create attribution is more broadly available.

Note: you cannot define fully custom probabilistic attribution logic. You're choosing among the pre-built models — and the model you choose changes the answer dramatically. Agree on the model before the channel debate starts.

Three saved reports to build at the start of any ChatGPT-era engagement:

1. AI-referred sessions by platform (traffic analytics)
2. Contacts created with Original Source = AI Referrals — or your ChatGPT-ad campaign, once live
3. Deals with any AI/ChatGPT touch in the attribution path

What about HubSpot's own AI-visibility tools?

HubSpot now ships tooling for the organic side of this picture.

HubSpot AEO (in beta) tracks how your brand appears across ChatGPT, Gemini, and Perplexity — with a Brand Visibility Score, competitor share-of-voice, citation tracking, and recommendations. It's included in Marketing Hub Pro and Enterprise, and available standalone at around €50/$50 per month (roughly €45/$45 billed annually), with a limited number of tracked prompts.

There's also a free, one-time AI Search Grader that scores a brand out of 100 across five weighted dimensions: Sentiment (40), Presence Quality (20), Brand Recognition (20), Share of Voice (10), and Market Competition (10).

Early adopters suggest the stakes: HubSpot cites Docebo attributing nearly 15% of its leads to AI traffic.

Is ChatGPT traffic actually worth it?

AI referrals are roughly 1% of web traffic today — dominated overwhelmingly by ChatGPT — but they convert at parity-to-higher rates than non-branded organic search, with far longer sessions. Small now, strategic later.

The volume picture, from two independent datasets:

• Conductor's AEO/GEO benchmark puts AI referrals at about 1.08% of total web traffic, with ChatGPT at ~87% of it, growth around 340% year-over-year, and AI-referred sessions lasting roughly 2.3x longer than traditional search sessions.
• Previsible's AI Traffic Report (6.77 million LLM-driven sessions across 166 sites) found ChatGPT sends 92.4% of all trackable standalone AI referral traffic.

The conversion picture is where it gets interesting. Visibility Labs' GA4 analysis (reported by Search Engine Land) compared 9.46 million non-branded organic sessions to 135,000 ChatGPT sessions across 94 ecommerce brands.

The result: ChatGPT traffic converted at 1.81% vs. 1.39% for non-branded organic — 31% higher, beating organic in 10 of 12 months.

The caveats: average order value ran lower ($204 vs. $238), and organic volume stayed 70x larger. The proposed mechanism is "intent compression" — users refine their needs inside ChatGPT before they ever click.

The honest read? Other studies range from ChatGPT converting worse than Google (a 973-site study) to 4–5x better (Semrush and Seer Interactive datasets). Methodologies differ enormously, the evidence skews ecommerce, and volume remains 47–200x smaller than Google organic.

But the direction — higher-intent, better-qualified visitors — is consistent across nearly everything published.

The B2B reach ceiling: why AEO still beats the ad slot

ChatGPT ads never reach paid-tier subscribers — Plus, Pro, Business, Enterprise, and Education users see no ads at all. Since that's where senior, budget-holding B2B buyers concentrate, organic citation inside the answer is the only way to reach them in ChatGPT.

This is the most important strategic caveat for European B2B SaaS.

One practitioner poll of 100 SaaS and tech marketers (Omni Lab) found 74% were on paid ChatGPT tiers. No bid, at any price, reaches them inside ChatGPT.

The organic answer, however, has no tier gate. Every user on every plan reads it.

That's the entire strategic case for treating the paid ad slot as the smaller prize and organic AEO/GEO citation as the bigger one.

And it's structural, not a temporary beta limitation. Plan the paid channel as incremental reach alongside LinkedIn, Google, and organic AEO — never as a replacement.

How should you prepare? (The 3-stage plan)

Stage 1 — Prepare now, before you spend a euro

Step #1: Baseline your organic AI-referral traffic in HubSpot today. Open Traffic Analytics → AI Referrals. Save a report of AI-referred sessions by platform and contacts created with Original Source = AI Referrals. You cannot prove lift later without a baseline now.

Step #2: Fix the classification before it happens. Adopt a written UTM convention for ChatGPT ads that avoids the cpc/ppc token — e.g., utm_source=chatgpt, utm_medium=conversational-ai, utm_campaign=chatgpt-<offer> — and build every ad URL through HubSpot's Tracking URL Builder. Document it so no one defaults to cpc.

Step #3: Grade your AI visibility. Run the free AI Search Grader (or our own instant AEO checker). If you're on Marketing Hub Pro or Enterprise, turn on HubSpot AEO to baseline brand visibility and competitor share-of-voice across ChatGPT, Gemini, and Perplexity.

Step #4: Invest in AEO/GEO content groundwork. Comparison pages, clear pricing, structured and authoritative content that answer engines cite. This is the only route to the ad-free paid-tier buyers.

Stage 2 — When you launch

Step #5: Install the OpenAI Pixel and Conversions API together, using a shared event ID for deduplication. The server-side Conversions API is more durable than the browser pixel against ad blockers and Safari ITP — OpenAI itself recommends it.

Step #6: Preserve oppref through redirects, and map OpenAI's {campaign_id}/{ad_group_id}/{ad_id} macros into your UTM string. That's what lets you reconcile OpenAI's dashboard with HubSpot.

Step #7: Start with a contained test budget — and measure on assisted conversions and pipeline influence, not last-click. Given contextual (not behavioral) targeting and the conversation gap, a strict last-click lens will make an early awareness channel look like a failure.

Stage 3 — Evaluate and scale

Step #8: Build a cohort view. Compare ChatGPT-sourced leads and trials against Google-sourced ones on activation, time-to-paid, and average contract value.

Step #9: Use benchmarks that change the decision.

• Scale if blended CPA lands within ~1.5x your Google Search CPA after 90 days and pipeline-influenced revenue is materializing.
• Pause or reduce if CTR falls below ~0.3% or CPC exceeds ~€8/$8 in your vertical without downstream pipeline.
• Re-check whether your ICP actually sits on Free/Go tiers before committing serious budget.

Want the tracking foundation, the reporting suite, and the launch plan built in your portal before the European self-serve Ads Manager opens? Get in touch with Superwork — we set up ChatGPT-ready measurement and AEO programs for European B2B teams on HubSpot.

Frequently asked questions

Are ChatGPT ads available in Europe?

Yes. ChatGPT ads are live for Free and Go-tier users across 31 European markets — the 27 EU member states plus Iceland, Liechtenstein, Norway, and Switzerland. At this stage, buying in Europe runs through OpenAI's Ads Solutions team and its agency and technology partners; a European self-serve Ads Manager has been announced as coming.

Who sees ChatGPT ads?

Only users on the Free and Go (€7.99/month) tiers. Plus, Pro, Business, Enterprise, and Education subscribers never see ads — which matters for B2B, since senior buyers concentrate on paid tiers.

How are ChatGPT ads targeted in Europe?

Contextually only at this stage: the system matches ads to the current conversation's topic and intent, approximate location, device type, time of day, and language. Personalized targeting requires separate explicit opt-in consent and is not active in the EEA and Switzerland at launch.

How much do ChatGPT ads cost?

In the US, where the market is most developed, CPMs run around €25/$25 (down from roughly €60/$60 at pilot) and CPCs roughly €3–5/$3–5 with category-based bid floors. There is no minimum spend in the self-serve US Ads Manager. European pricing is set through OpenAI's sales and partner channels at this stage.

How does HubSpot classify ChatGPT ad traffic?

By default, badly — if you tag ad URLs with utm_medium=cpc, HubSpot's Paid-search rule fires before its AI Referrals rule and files the clicks under Paid search, blended with Google Ads. Use a distinct medium (like utm_medium=conversational-ai) with utm_source=chatgpt and a clear utm_campaign so the traffic stays isolated and reportable.

Does HubSpot have an AI referrals traffic source?

Yes. HubSpot's native AI Referrals source classifies organic referrals from nine platforms — ChatGPT, Claude, Perplexity, Copilot, Gemini, Meta AI, Mistral, Poe, and Grok — with the platform in Drill-Down 1. It is organic-only; there is no "Paid LLM" category.

Does ChatGPT traffic convert?

The best-documented study (94 ecommerce brands, GA4) found ChatGPT sessions converting at 1.81% vs. 1.39% for non-branded organic search — 31% higher — attributed to users refining their intent inside ChatGPT before clicking. Studies conflict on the exact multiple, and B2B evidence is thinner, but the direction — smaller volume, higher intent — is consistent.

Should European B2B companies buy ChatGPT ads or invest in AEO?

Both, in that order of preparation but the reverse order of ceiling: the ad slot only reaches Free and Go users, while organic citation in the answer reaches every tier — including the paid tiers where senior B2B buyers sit. Treat ads as incremental reach and AEO/GEO as the structural investment.

The bottom line

ChatGPT ads in Europe are at the rare stage where preparation beats budget.

The channel is real: live in 31 European markets, maturing fast, with conversion data pointing at higher-intent visitors.

And the constraint is equally real: the clicks won't attribute themselves, HubSpot will misfile them under Paid search if you let it, and no ad budget reaches the paid-tier buyers who matter most in B2B.

So do the unglamorous work now:

• Baseline your AI referrals
• Write the UTM convention that keeps ChatGPT ads out of the Paid-search bucket
• Wire Original Source into your deal reporting
• Grade your AI visibility and start earning citations — because the answer itself is the one placement money can't buy

Then, when European self-serve opens, you launch with measurement that holds up instead of a blended mess.

If you want your HubSpot portal ready before the spend starts — tracking, attribution, AEO baseline and all — reach out to Superwork and we'll build the ChatGPT ads foundation with you.

This is a fast-moving topic, and we update this guide as the rollout evolves. A few honest sourcing notes: OpenAI has not fully documented its attribution windows or view-through methodology, and ad blockers, stripped referrers, and the dark funnel (users who see a ChatGPT mention, then arrive via branded search or direct) all cause undercounting — treat both HubSpot and OpenAI numbers as floors, and consider a post-conversion "how did you hear about us?" field. Conversion-quality studies conflict and skew toward ecommerce; B2B SaaS evidence is thinner, so don't over-index on any single "AI converts Nx better" figure. The 56% Norwegian consumer adoption figure is solid Eurostat data; enterprise-adoption figures circulating in industry commentary are less well sourced. Pricing, formats, and availability shift month to month — verify against OpenAI's current documentation before committing budget.