Focal
Focal
Book a Call
Back to blog
Published July 5, 2026

Are ChatGPT Ads Worth It? Real ROAS & Performance Data

Henry Purchase

Henry Purchase

Co-Founder

Are ChatGPT Ads Worth It? Real ROAS & Performance Data

Yes, but only for certain business types. The channel produces genuine returns for B2B SaaS, professional services, and considered consumer purchases where buyers research extensively before deciding. For impulse retail, hyper-local same-day services, or low-value transactional categories, it usually does not clear the hurdle.

The distinction matters. ChatGPT Ads reached $100 million in annualized revenue in its first six weeks with over 600 advertisers, according to Reuters. The demand is real. The question is whether your specific business is one of the ones the maths works for. This guide covers what the actual performance data looks like in 2026 and when to move.

TL;DR: ChatGPT Ads Performance in 2026

  • Average CTR. 0.91 per cent (Adweek data from the early beta).
  • Top-quartile CTR. One per cent or above (Similarweb Q1 2026).
  • Conversion lift. 1.5x other digital channels for LLM-referred traffic (Criteo Q1 2026).
  • Time on site. 60 to 80 per cent longer than social traffic for the same user profile.
  • Best for. B2B SaaS, professional services, financial services, education, and high-consideration consumer purchases.
  • Worst for. Impulse purchases, hyper-local same-day services, low-AOV transactional retail.
  • First-mover window. Currently open, expected to close in 18 to 24 months as advertiser density catches up to Google and Meta.

The Early Performance Data

Six weeks after launch on February 9, 2026, OpenAI's ChatGPT Ads pilot had crossed $100 million in annualized revenue and signed over 600 advertisers per Reuters reporting. At that stage, 85 per cent of Free and Go tier users were eligible to see ads, but only 20 per cent saw them daily. Advertiser spend outpaced inventory availability by a wide margin.

That gap explains the pricing behaviour. Early CPMs held at $60 for two months before compressing to $25 as OpenAI expanded inventory and opened self-serve. If ads had underperformed, prices would have collapsed sooner. They did not.

Watch

I Spent $10,000 on ChatGPT Ads and Learned This

Henry breaks down what actually happened after spending $10,000 on ChatGPT Ads: what worked, what did not, and what the CTR, conversion, and ROAS numbers looked like across different campaign types. If the "worth it" question is what brought you to this post, this is the real-world case study that answers it directly.

CTR Benchmarks (2026)

The widely cited "1.3 per cent industry average CTR" figure from early 2026 commentary is skewed by survivor bias from a handful of high-performing early advertisers. The real benchmark distribution, from Similarweb's Q1 2026 dataset:

  • Display ads (average): 0.35 per cent
  • ChatGPT Ads (average): 0.68 per cent
  • ChatGPT Ads (top quartile): one per cent
  • ChatGPT Ads (top ten per cent): 1.57 per cent

ChatGPT Ads sit roughly twice the display average. Above one per cent, creative is meaningfully working. Below the 0.35 per cent floor, something is broken. Google Search remains higher at around 6.4 per cent, but the comparison is misleading. Search users have declared explicit intent by typing a query. ChatGPT users are still forming intent inside a conversation.

The Conversion Quality Argument

CTR is not the important metric. Conversion quality is.

Criteo data from Q1 2026, based on 500 US retailers, showed that users referred from LLM platforms like ChatGPT convert at approximately 1.5x the rate of other referral channels. Time-on-site data is even more striking: ChatGPT referral visitors spend 60 to 80 per cent longer on advertiser sites than social traffic.

The mechanism is intent depth. When someone asks ChatGPT for advice on a purchase, they arrive on the advertiser's landing page with more context than a Google searcher and much more than a social scroller. They have already declared what they need. The ad answered their question. The click is not exploratory.

For a business selling high-consideration products, this lift matters more than the higher CPM.

Who ChatGPT Ads Actually Work For in 2026

Six categories that consistently see returns:

B2B SaaS with mid-market or enterprise pricing. Buyers research extensively before booking demos. ChatGPT is now a primary research surface for this cohort.

Professional services (agencies, consultants, accountants, coaches). Discovery calls converted from ChatGPT-led research traffic tend to be higher quality than those from paid social.

Financial services and fintech. Research-heavy category with detailed comparison behaviour. High CPM justified by high customer LTV.

Education and coaching programmes. Buyers evaluate options extensively before enrolling.

Home services with national reach. Contractors, moving services, home improvement specialists all see qualified enquiries.

Considered consumer purchases. Furniture, appliances, health and wellness products, high-AOV apparel. Anything a buyer researches before purchasing.

The common thread: buyers who compare options before spending, in categories where the customer lifetime value is high enough to absorb higher CPCs.

Who They Don't Work For

Four categories where the maths rarely clears:

Impulse consumer retail. Low-AOV fashion, quick-serve food, novelty gifts. Users are not in ChatGPT for these decisions.

Hyper-local same-day services. Emergency plumbing, roadside assistance, "near me" queries. Google Local still wins these.

Broad brand awareness for undifferentiated products. ChatGPT users are researching, not passively browsing. Awareness campaigns for commodity products underperform.

Categories with restricted eligibility. Pharmaceuticals, adult content, political advertising, and other regulated categories are limited or excluded on the platform in 2026.

If your product falls into one of these categories, wait. Test the channel when the format expands or the auction gets cheap enough that the maths flips.

The First-Mover Window

The strongest ROAS argument for 2026 is not about creative quality or targeting. It is about timing.

CPMs dropped from $60 to $25 in three months as OpenAI opened inventory. Advertisers who moved during the pilot paid the pilot premium and built the account history. Advertisers who move now pay the current CPMs and still have far less competition than they will in 12 or 24 months.

Most paid media leaders expect the channel to reach Google-comparable advertiser density within 18 to 24 months. The brands testing now will have compounded account history, creative learnings, and audience data by the time competition catches up. Brands waiting until 2028 will start from scratch, at 2x to 3x current CPCs.

For businesses in the fit categories above, the ROI on testing right now is not just the direct revenue. It is the learning advantage that closes the gap on brands who are already testing.

Watch

ChatGPT Ads: The $300B Opportunity in 2026

Henry lays out the full first-mover argument: why ChatGPT Ads represent a $300 billion opportunity by the end of the decade, what the timing looks like across the next 18 to 24 months, and which categories are best positioned to compound the current arbitrage. Worth watching if the ROAS numbers on this post look interesting and you want the wider context on where the channel is heading.

Realistic ROAS Expectations

By category, based on what Focal sees across the campaigns we run:

  • B2B SaaS (mid-market). 3x to 5x return on ad spend on qualified pipeline within 90 days.
  • Professional services. 4x to 8x on qualified discovery calls within 60 days.
  • Considered consumer. 2x to 4x return on ad spend within 45 days.
  • DTC with product feed ads. 2x to 5x on catalogue campaigns within 60 days.

These ranges assume tracking is set up correctly (see our OAIQ pixel setup guide), the campaign structure matches the business type, and creative is written for the format rather than repurposed from Google or Meta.

Underperforming these ranges is usually a signal that the tracking is broken, the creative is off, or the category is a poor fit. Overperforming them means you found a strong audience-hint match early. Both are worth investigating.

How Focal Helps Businesses Test ChatGPT Ads Profitably

Focal is a done-for-you ChatGPT Ads agency that runs campaigns for DTC, B2B SaaS, and service businesses. Our team ships first campaigns within seven days of kickoff and reports monthly on ROAS with creative-level breakdowns.

If you want a specialist team running the channel and measuring outcomes against pipeline instead of impressions, book a free 30-minute discovery call. We will tell you on the call whether ChatGPT Ads are worth it for your specific business.

Frequently Asked Questions

Are ChatGPT Ads worth it in 2026?

For businesses selling considered purchases (B2B SaaS, professional services, financial products, high-consideration consumer goods), yes. Early data shows 1.5x conversion lift vs other digital channels and 60 to 80 per cent longer time on site. For impulse retail, hyper-local same-day services, and low-AOV categories, the maths rarely clears.

What ROAS should I expect on ChatGPT Ads?

3x to 8x ROAS on qualified pipeline within 60 to 90 days for B2B SaaS and professional services. 2x to 5x for considered consumer categories within 45 to 60 days. Below 2x usually means the tracking is broken or the category is a poor fit.

Are ChatGPT Ads more profitable than Meta or Google?

On CPM basis, ChatGPT is more expensive than Meta and comparable to premium Google Search. On conversion quality, ChatGPT delivers about 1.5x the conversion rate. Whether that offsets the CPM premium depends on your average customer value and funnel.

When will ChatGPT Ads stop being cost-effective?

Most operators expect advertiser density to catch up to Google levels within 18 to 24 months, at which point current CPCs will roughly double. The current pricing reflects the early phase of a new channel. Brands moving now lock in the arbitrage.

Watch Henry Purchase's ChatGPT Ads walkthroughs on YouTube for weekly platform updates.