When OpenAI opened up ChatGPT Ads as a self-serve platform earlier in 2026, our agency did what we always do with a new channel: we ran toward it with real client budgets instead of waiting for a case study someone else wrote.
Six months in, we have actual data. Not impressions. Not vibes. Not “early signs are promising.” Real CPCs, real conversion rates, real comparisons against Google and Meta across B2B and ecommerce accounts. This is what we found — including the parts that surprised us and the parts that should make you cautious.
- ChatGPT Ads are performing best for high-intent B2B audiences — CPLs are competitive with Google Search in several verticals, but volume is still limited.
- Ecommerce results are mixed. Conversion rates are lower than Google Shopping, but early ROAS is holding in the 2–3x range for direct-response creative.
- The audience is distinct from Google and Meta. Don’t assume your existing creative or bidding logic will transfer directly — it won’t.
- Attribution is the biggest unsolved problem right now. Without proper tracking infrastructure, you’ll misread the channel entirely.
- This is a complement to Google Ads, not a replacement. Treat it like Bing Ads circa 2018 — worth testing seriously, not worth abandoning your core channel over.
Why We Started Taking ChatGPT Ads Seriously Immediately
Most agencies waited. They wanted to see “the data” before committing client budget. We think that’s the wrong instinct — the agencies that were early on Google Ads in 2002, Facebook Ads in 2012, and Performance Max in 2021 captured disproportionate learnings when competition (and CPCs) were lowest.
ChatGPT has over 500 million weekly active users as of early 2026. A significant chunk of them are using it for product research, vendor comparisons, and purchase decisions. That’s not a social media browse audience. That’s an audience actively asking questions and looking for answers — which is about as close to search intent as you can get without being search.
The format is also different from anything else running right now. Ads surface contextually within conversations, tied to the query the user just typed. A user asking “what’s the best CRM for a 20-person sales team” can see a sponsored result from a SaaS vendor right there in the response flow. That’s not a banner ad. That’s intent capture at the moment of consideration.
We started testing across five B2B clients and four ecommerce clients in January 2026. Here’s what six months of real spend taught us.
ChatGPT Ads Performance Data: What the Numbers Actually Look Like
Let’s get specific, because vague impressions are useless to you.
For B2B accounts (SaaS, professional services, financial services), we’re seeing average CPCs ranging from $4.20 to $9.80 depending on vertical. That sits below Google Search for competitive B2B keywords, which routinely run $15–$40+ in categories like legal, finance, and enterprise software. The catch: volume is lower. You’re not replacing your Google Search budget here. You’re adding an incremental layer.
Cost per lead on our best-performing B2B account — a mid-market SaaS client targeting operations managers — came in at $61 over Q1 2026. Google Search CPL for the same client sits around $74. That’s a real, meaningful difference. But the ChatGPT Ads volume was about 30% of the Google Search volume in the same period. So the channel earns budget allocation, but not the majority.
For ecommerce accounts, it’s messier. Our best performer — a DTC home goods brand — hit a 2.4x ROAS on ChatGPT Ads over the first 90 days. Respectable, but below the 4.1x they’re pulling from Google Shopping. The conversion rate on ChatGPT was about 1.8% versus 3.2% on Google Shopping. Users clicking from a conversational AI interface are still in a slightly different mindset than users who just searched “buy [product] near me.” The purchase friction is marginally higher.
Our weakest performer was a competitive ecommerce account in apparel. The ROAS sat at 1.4x — not terrible for a brand-new channel in its first 60 days, but not enough to justify scaling without more optimization work. We’re still running it, but with a tightly controlled budget until the creative finds its footing.
The Creative Logic Is Different Here — and That Catches Most Advertisers Off Guard
This is where we see the most mistakes, even from experienced paid media teams. People pull their Google RSA copy, strip out the headlines, reformat it for ChatGPT Ads, and wonder why performance is flat.
The problem: Google Search users typed a keyword. ChatGPT users typed a question, and they’re inside a conversation. The psychological context is completely different.
What’s working for us is answer-first creative. Instead of leading with a brand claim or a feature benefit, the best-performing ads frame themselves as a direct continuation of what the user already asked. Think: “Looking for [solution to their problem]? Here’s how [brand] approaches it differently.” It matches the conversational register of the interface.
Calls to action also behave differently. Hard CTAs like “Buy Now” or “Get a Free Quote” are underperforming softer engagement CTAs like “See how it works” or “Compare your options.” Users in a research mindset want to keep learning, not commit. Push them too hard and they scroll past.
If you want a framework for testing which message angles are actually working — on any channel — our Google Ads A/B testing framework translates well to ChatGPT Ads creative iteration. The principles are the same: isolate variables, run clean tests, don’t declare a winner on 40 clicks.
Attribution Is Broken Right Now — Here’s How We’re Working Around It
This is the part nobody talks about in the breathless “ChatGPT Ads are amazing” posts, and it’s the most operationally important issue you’ll face.
OpenAI’s native reporting is still maturing. The conversion tracking setup is workable but not as battle-tested as Google’s infrastructure, which has had 20 years of iteration. Right now, if you’re not tagging your ChatGPT Ad destination URLs with proper UTM parameters and passing them into a reliable analytics setup, you’re flying completely blind on attribution.
We’ve seen ChatGPT-sourced traffic show up as “direct” or “referral” in GA4 for clients who hadn’t implemented URL tracking before we took over. They were getting real conversions from the channel and couldn’t see them. Months of spend with zero visibility.
The same discipline you need for Google Ads URL parameters and tracking templates applies here — arguably even more so, because the channel is new and the default reporting gaps are larger. Tag everything. Every ad, every variation, every audience segment.
We’re also being careful about the attribution model used to evaluate ChatGPT Ads results. The channel skews toward upper-funnel and mid-funnel behavior. If you’re running last-click attribution, you’ll systematically undervalue it — users often come through ChatGPT, don’t convert immediately, and close later through a branded Google search. Data-driven attribution handles this better, but you need enough conversion volume for it to be reliable.
How ChatGPT Ads Fit Into a Multi-Channel Paid Media Strategy
The honest answer to “should I run ChatGPT Ads?” is: it depends on your current channel mix and how efficiently you’re running it.
If you’re not maxing out on Google Search — if there’s impression share available on your core terms, if your Quality Scores are dragging your CPCs up, if your landing pages are leaking conversions — fix that first. ChatGPT Ads won’t save a broken core. If you’re not thinking carefully about the relative strengths of Google versus Meta for your lead generation goals, adding a third channel before you’ve mastered two is how budgets get diluted and results get muddy.
But if your Google and Meta accounts are well-structured and you have budget to expand, ChatGPT Ads is the most interesting new channel we’ve seen since Performance Max launched. The audience intent is high. The CPCs are still early-adopter efficient. And the lack of sophisticated competition right now means your ad doesn’t need to outbid dozens of well-optimized competitors — yet.
Think of the channel allocation this way for most B2B advertisers in 2026: Google Search gets 55–65% of your paid search budget. Meta or LinkedIn gets 20–30% depending on your audience. ChatGPT Ads earns 10–15% of budget as a test channel, with a 90-day review to decide whether to scale or hold.
For ecommerce, the weighting shifts. Google Shopping and Search remain dominant, and you should be deeply invested there before touching ChatGPT Ads. If your ecommerce campaign structure is solid and you’re hitting target ROAS consistently, then a 10% budget allocation to ChatGPT Ads as a test is reasonable.
The Honest Verdict: What OpenAI Ads Results Tell Us About Where This Is Going
Six months in, we’re believers — with conditions.
ChatGPT Ads is a real channel with real intent and real conversion potential. It’s not hype, it’s not a toy, and agencies that dismiss it as “too new to bother with” are making the same mistake agencies made about mobile search in 2013. The behavior is already there. The ad infrastructure will catch up.
But the AI ads benchmarks for 2026 are early-stage data. Our numbers are real, but they’re based on a limited sample size and a platform that is actively changing month to month. What’s working for creative in Q1 may need to be rethought by Q3. The negative keyword equivalent — whatever content-exclusion tools OpenAI builds into the platform — will matter enormously for efficiency and we’re watching that closely.
The advertisers who will win on this channel over the next 18 months are the ones who treat it with the same rigor they apply to Google: disciplined tracking, structured testing, honest performance reviews, and a willingness to pause what isn’t working instead of rationalizing it.
The ones who will lose are the ones who dump budget in, look at surface-level metrics, and either declare it a miracle or write it off after 30 days because the ROAS wasn’t perfect immediately.
Sound familiar? It should. It’s the same mistake people have been making on Google Ads for two decades.
Frequently Asked Questions
What are ChatGPT Ads and how do they work?
ChatGPT Ads is OpenAI’s self-serve advertising platform that launched in 2026. Ads appear contextually within ChatGPT conversations, surfacing as sponsored placements when users ask questions relevant to the advertiser’s targeting. Advertisers can set budgets, define audience parameters, upload creative, and track results through OpenAI’s ad manager — similar in structure to Google Ads or Meta Ads, but optimized for a conversational AI context.
How do ChatGPT Ads CPCs compare to Google Ads in 2026?
Based on our agency data across B2B and ecommerce accounts, ChatGPT Ads CPCs are running 30–60% lower than Google Search CPCs in competitive verticals. B2B CPCs are ranging from $4–$10 on ChatGPT versus $15–$40 on Google Search for comparable keywords. That gap will likely narrow as more advertisers enter the platform, which is one reason early adoption matters.
Are ChatGPT Ads better for B2B or ecommerce?
Our data points to stronger early results for B2B. The high-intent, research-oriented behavior of ChatGPT users aligns naturally with B2B buying cycles, where prospects are asking complex questions and evaluating multiple options. Ecommerce results are more mixed — conversion rates are lower than Google Shopping, but ROAS can be positive with the right creative approach. We’d prioritize B2B testing first.
How do I track ChatGPT Ads conversions properly?
You need UTM parameters on every destination URL — at minimum source, medium, campaign, and content variables. Verify that your GA4 or analytics platform is correctly attributing ChatGPT traffic and not bucketing it as direct or organic. OpenAI’s native conversion tracking is improving but not yet at parity with Google’s infrastructure, so don’t rely on it exclusively. Cross-reference with your CRM data wherever possible.
Should I move budget from Google Ads to ChatGPT Ads?
No — at least not yet. Treat ChatGPT Ads as an incremental budget expansion, not a reallocation away from a proven channel. Google Search still delivers higher volume, more mature optimization tooling, and better attribution transparency. The right approach in 2026 is to run both, with Google as your primary channel and ChatGPT Ads as a 10–15% test budget that earns its way into a larger allocation based on actual results.
What kind of creative works best on ChatGPT Ads?
Answer-first messaging that matches the conversational context of the platform. Users are in research mode, not shopping mode. Creative that continues the user’s line of thinking — rather than interrupting it with a hard sell — consistently outperforms in our tests. Softer CTAs (“see how it works,” “compare your options”) are outperforming direct-response CTAs (“buy now,” “get a free quote”) at this stage of the platform’s development.
Thinking About Adding ChatGPT Ads to Your Channel Mix?
Here’s the honest checklist before you do: Is your Google Ads account running efficiently, with clean conversion tracking, a tightly managed negative keyword list, and a campaign structure that isn’t cannibalizing itself? If the answer to any of those is “we’re not sure,” that’s where your attention and budget belong first.
If your core paid search is dialed in and you want a second opinion on how ChatGPT Ads might fit your specific business — or you want someone to manage the test so you get clean data instead of noise — that’s exactly the kind of work we do. We’re a Google Ads management agency that has now expanded into multi-channel paid media precisely because our clients need honest answers about where their next dollar should go, not cheerleading for whatever channel is getting press this quarter.