OpenAI is building a full-stack performance advertising business inside ChatGPT
The Signal
OpenAI is building a full-stack performance advertising business inside ChatGPT. The company has hired ServiceNow CMO Colin Fleming to lead its business marketing push, is developing conversion-focused ad formats tied to cost-per-action pricing, and is constructing the measurement infrastructure — pixels, APIs, conversion tracking — required to compete directly with Google and Meta for performance budgets. Fifteen weeks into its ChatGPT ad pilot, OpenAI is targeting small and mid-sized businesses including local service providers such as dry cleaners, car washes, and appointment-based companies. New ad formats are designed to drive purchases, bookings, and contact form submissions, with advertisers paying only on completed actions. To support attribution, OpenAI is requiring advertisers to install a proprietary pixel and encouraging API integration so conversion data flows back into OpenAI's systems. The buildout comes as OpenAI eyes a public market debut as early as September 2026 at a reported valuation of approximately $1 trillion. Fleming, who called the opportunity one he would have "regretted not taking," joins as the company attempts to resolve early tensions between advertiser expectations and product reality that have accompanied every major ad platform at this stage.
Why It Matters
The hiring of Colin Fleming is the tell — not the ad formats, not the pixel, not the CPA pricing. OpenAI isn't building an ad business; it's building an IPO story, and it needs a credible revenue narrative before September. Here's what the coverage keeps glossing over: the SMB segment OpenAI is targeting — dry cleaners, car washes, appointment-based locals — is exactly the cohort that Google and Meta have spent fifteen years training to use self-serve performance tools. These aren't experimental advertisers looking for a new platform to test. They're operationally lean businesses that measure every dollar and have zero patience for attribution gaps. Winning them requires not just formats and pixels, but a support infrastructure and optimization loop that takes years to build. OpenAI is fifteen weeks into a pilot and already surfacing the tensions every major platform faced at this stage — except those platforms didn't have a $1 trillion valuation to defend while they figured it out. The deeper dynamic is this: performance advertising is a trust business. Meta earned SMB loyalty through granular targeting and a feedback loop that actually worked. Google earned it through intent signals no one else had. OpenAI's differentiator — conversational context — is genuinely powerful, but it's unproven as a conversion driver, and browser restrictions will erode pixel reliability faster than OpenAI can shore it up with API integrations. For marketing executives evaluating this channel, the question isn't whether ChatGPT ads will eventually work. They will. The question is whether OpenAI's IPO timeline gives the product enough runway to mature before advertiser expectations outrun product reality — and that gap is the real risk to watch.
The Exploit
Opportunity
Performance marketing leaders at mid-market B2C companies—particularly those spending $500K–$5M annually on Google and Meta—have a 90-day window to negotiate first-mover terms with OpenAI's ad team before IPO pressure standardizes pricing. Early pilot partners gain disproportionate access to product roadmap influence, preferred CPA rates, and attribution data that won't be available post-launch at scale.
Risk
Teams that wait for proof will pay full price into a crowded auction. The greater danger is misreading this as a 'test and learn' moment—allocating junior resources to run a low-budget experiment generates no leverage. CMOs who don't engage at the commercial level before September 2026 cede negotiating position permanently; IPO economics will demand margin expansion, not partner discounts.
The Move
Your CMO or VP of Growth should open a direct commercial conversation with OpenAI's advertising team before August 2026—not to run creative tests, but to negotiate pilot terms that include CPA rate locks, pixel data ownership provisions, and a seat in the product feedback loop. The checkpoint: a signed pilot agreement with contractual data rights before OpenAI files its S-1.
First-Mover Advantage
Gains
Risks
Teams that wait for proof will pay full price into a crowded auction. The greater danger is misreading this as a 'test and learn' moment—allocating junior resources to run a low-budget experiment generates no leverage. CMOs who don't engage at the commercial level before September 2026 cede negotiating position permanently; IPO economics will demand margin expansion, not partner discounts.
Window
Winners & Losers
Winners↑
API Integration and Ad Tech Middleware Vendors
OpenAI's explicit push toward API-based conversion tracking over pixel reliance creates immediate demand for middleware vendors who can bridge SMB back-end systems — booking software, CRMs, POS platforms — into OpenAI's data infrastructure. As browser restrictions erode pixel reliability, API-connected attribution becomes the premium tier, and vendors who can package that integration for operationally lean SMBs will own a critical chokepoint. The play here is to productise a 'ChatGPT-ready' integration layer now, before OpenAI builds it natively or locks in preferred partners ahead of the IPO.
SMB Performance Marketing Consultants and Boutique Agencies
The SMB segment OpenAI is targeting — dry cleaners, car washes, appointment-based locals — has almost no internal capability to evaluate a new ad channel, install pixels, configure API connections, or interpret CPA performance data without outside help. Every new platform that enters this space creates a consulting wedge, and the early-mover advisors who build ChatGPT campaign fluency in Q2-Q3 2026 will be positioned as the trusted guides when OpenAI's self-serve tools mature post-IPO. Agencies should begin documenting pilot learnings now to build proprietary benchmarks that justify retainer relationships.
Conversion Rate Optimisation and Landing Page Specialists
CPA pricing models only pay out on completed actions, which means advertiser success on ChatGPT's platform is directly downstream of post-click conversion performance — landing pages, booking flows, and form completion rates become the rate-limiting factor for ROI. CRO specialists who can credibly connect their work to ChatGPT CPA outcomes have a strong new value proposition to sell into the SMB market that OpenAI is actively recruiting. The smart move is to build case studies now using early pilot advertisers and position conversion optimisation as the unlock for making ChatGPT ads actually work.
Independent Measurement and Attribution Platforms
OpenAI's proprietary pixel and closed attribution ecosystem creates the same walled-garden tension that has driven demand for independent measurement tools alongside Meta and Google for years. Advertisers running multi-channel budgets will need a neutral third party to reconcile ChatGPT's self-reported CPA data against their broader performance stack, especially given how early and unproven OpenAI's measurement infrastructure is. Platforms like Northbeam, Triple Whale, or Rockerbox-style independents should be actively scoping ChatGPT integrations now to be ready when advertiser demand for cross-channel validation materialises.
Losers↓
Google and Meta SMB Advertising Business Units
The SMB performance segment is the most profitable and defensible cohort both platforms have built over fifteen years, and OpenAI is directly targeting it with a pricing model — cost-per-action with zero waste spend — specifically designed to appeal to the ROI-sensitive operators Google and Meta rely on for long-tail revenue. While OpenAI's product is immature, the competitive framing alone will force both platforms into defensive moves — accelerating AI-native ad formats, lowering CPA floors, and investing more in SMB support infrastructure to reduce churn risk. The real threat isn't Q3 2026 budget shifts; it's that OpenAI's conversational intent signal, once proven, undercuts the search and social intent models that justify Google and Meta's pricing power.
Self-Serve Performance Ad Platforms Targeting SMB Segments
Mid-tier self-serve platforms — including vertical-specific ad networks, local search advertising products, and newer performance channels — are structurally most exposed because they compete for the same lean SMB budgets without Google's or Meta's scale advantages or OpenAI's conversational differentiation. These platforms have typically won on the basis of simplicity, category focus, or lower CPAs, and OpenAI's CPA-only pricing model directly attacks the cost-efficiency argument while the ChatGPT brand name attracts experimenters away from smaller incumbent platforms. Their defensive move is to double down on category depth and human support — the two dimensions where a platform fifteen weeks into a pilot and racing toward a $1 trillion IPO is least equipped to compete.
Strategic Outlook
The timing is the context. OpenAI is roughly four months from a reported September IPO window, which means what looks like a product buildout is also a revenue narrative under construction — and those two things are in tension in ways the coverage undersells. Meta ran a version of this playbook in 2012, rushing its mobile ad product to market ahead of its IPO before the measurement infrastructure was ready, and it took nearly two years of post-IPO turbulence before advertisers trusted the performance loop. Google's path was more deliberate — AdWords scaled slowly through 2002 and 2003 before SMBs adopted it at volume, and that adoption was earned through a self-serve interface and a feedback loop that actually optimized. OpenAI is attempting to compress that maturation cycle into a pre-IPO window. The Colin Fleming hire signals that the company understands it needs institutional credibility, not just product features — ServiceNow's marketing under Fleming was a B2B positioning story built around enterprise trust, which is precisely the register OpenAI needs to address CMO skepticism. Meanwhile, this arrives as Google faces antitrust remedies that could structurally open search advertising to competition for the first time in a decade. OpenAI isn't just building an ad business into a stable market. It's building one into a market that may be about to shift.