# ChatGPT ads pull $15M in test budgets but 4x higher cost-per-acquisition keeps holiday spend on Meta and Google *Brands are testing OpenAI's ad platform, but conversion economics don't justify shifting proven Q4 budgets—yet.* By **Jenny Huang Goodman MPA MSc MHSA, Principal** — The Stash Edge, Hako Shikin LLC. Published 2026-10-10. Canonical: https://www.pops4.com/stash/articles/meta-and-openai-2026-10-10t00-6 Subject: Meta and OpenAI Tags: ad platform testing, cost per acquisition, holiday budgets, chatgpt ads, performance marketing, conversion economics --- OpenAI launched ChatGPT ads in September, and brands moved quickly to test. According to Modern Retail, early adopters allocated roughly **$15 million** in exploratory spend across Q3 and early Q4. But the platform's cost-per-acquisition is running **three to four times higher** than comparable Google Search or Meta feed placements, and conversion rates have not justified pulling budget from proven channels during the critical holiday window. The mechanics are straightforward. ChatGPT serves sponsored messages within conversational responses—users ask a question, the AI answers, and a brand-sponsored follow-up appears inline. OpenAI pitches this as high-intent placement: the user is already engaged, context is rich, and the brand message arrives at the moment of consideration. Brands in home goods, gifting, and apparel tested the format with standard direct-response creative, tracking clicks through to product pages and checkout. The problem is unit economics. Modern Retail reports that while click-through rates on ChatGPT ads are comparable to search, the cost-per-click is materially higher—roughly **$8 to $12** versus **$2 to $4** on Google Shopping for similar queries. Conversion rates from click to purchase have been lower, likely because ChatGPT users are in research mode rather than buying mode. The result: a blended cost-per-acquisition that makes the channel uncompetitive against Meta retargeting or Google branded search, especially when holiday budgets are locked and performance thresholds are non-negotiable. The underlying mechanism is adoption risk in performance marketing. Brands with physical products operate on known return-on-ad-spend targets. A new platform must deliver comparable or better unit economics within a short test window, or the budget reallocates. ChatGPT ads are not failing—they are simply unproven at scale, and Q4 is not the quarter to experiment with unproven channels. The brands testing ChatGPT are treating it as a 2025 line item, not a 2024 performance driver. A small physical-product brand can replicate the test discipline without the waste. Allocate **$500** to ChatGPT ads in January or February, when you can afford to learn. Structure the campaign as a single product, single audience, single creative variant. Track cost-per-click, conversion rate, and blended CAC against your Meta or Google baseline. If ChatGPT delivers a lower CAC on a statistically valid sample—say, **50 conversions**—you have a new channel to scale. If not, you have data to revisit in six months when OpenAI iterates the format or lowers pricing. The discipline is the same as any new platform: test small, measure hard, scale only on proof. The broader pattern is that new ad platforms earn budget when they solve a problem existing platforms do not. Google solved search intent. Meta solved social proof and retargeting. ChatGPT's promise is conversational context, but that context must convert at a competitive cost. Until it does, holiday budgets stay where the math works, and new platforms wait for February. ## The takeaway Test new ad platforms in off-peak months with tight budget caps and conversion tracking; scale only when unit economics beat your baseline. --- ## Publisher **Hako Shikin LLC** — Virginia Beach, Virginia. Founded 1997. ASI 217876 · DUNS 18-204-6339. Principal and author: **Jenny Huang Goodman MPA MSc MHSA**. - Author: https://www.huanggoodman.com/about - LLM context: https://www.pops4.com/stash/llms.txt - MCP endpoint, for AI agents: https://mcp.pops4.com/mcp - Client dashboard: https://dashboard.pops4.com/ - Catalogue: 70,000+ products, 200+ brands