# ChatGPT Ads vs Google Ads: Search, Shopping and PMax URL: https://adsonomy.com/guides/chatgpt-ads-vs-google-ads Published: 17 September 2026 Last updated: 17 September 2026 Plain-text version of the page above. Adsonomy is an independent product and is not affiliated with OpenAI. ## Summary Compares ChatGPT ads with Google Ads (Search, Shopping and Performance Max): the intent model (keywords versus context hints), ad formats, Smart Bidding versus ChatGPT's fixed and maximise strategies, measurement and attribution windows, the controls Google has that ChatGPT ads lack, and when each platform, or both, is the better choice. ## Key facts - Google Search matches on keywords with three match types, broad, phrase and exact; ChatGPT ads match by meaning against up to 2,000 free-text context hints per ad group, with no query-level report. - Google's ad auction is recalculated for every search from six factors, including bid and Quality Score; ChatGPT ads run on a bid and a billing event set at the ad group, with no auction insight report. - Performance Max runs one campaign across YouTube, Display, Search, Discover, Gmail and Maps; ChatGPT ads run only inside ChatGPT, through two creative formats, a chat card or a feed-based product ad. - Google's Smart Bidding (Target CPA, Target ROAS, Maximize Conversions, Maximize Conversion Value) sets bids per auction from historical and real-time signals; the ChatGPT Ads API offers only fixed bid, maximise clicks or maximise conversions. - Google's Enhanced Conversions send hashed first-party data server-side through the Google Ads API; OpenAI's Conversions API does the same with SHA-256 hashed identifiers, deduplicated against its Pixel. - Google Shopping and Performance Max pull structured data from a Google Merchant Center feed; ChatGPT product ads pull from OpenAI's own product feed schema, uploaded over SFTP. ## Full text Google Ads and ChatGPT ads solve a similar problem, getting an offer in front of someone close to acting, with almost nothing else in common in how they do it. Google runs an auction on words people type, or hands a whole account to Performance Max to work across its ad inventory with AI-combined creative and audiences. ChatGPT ads run on written descriptions of a situation, matched to a conversation by meaning, inside two lean formats. This guide sets out what is actually documented on each side: the intent model, the formats, how bidding and measurement work, and where one platform is simply the better choice. **Key takeaways** - Bid on keyword intent in Google Search; describe situations with context hints in ChatGPT ad groups, since there is no keyword auction there. - Treat Performance Max and ChatGPT ads as different scales of the same idea: PMax spans Google's whole inventory from one campaign, ChatGPT ads only reach conversations inside ChatGPT. - Rebuild Smart Bidding's target CPA or ROAS yourself for ChatGPT, since the API has no equivalent field; a controller has to step fixed bids toward the target. - Run two separate feeds if you sell products: a Google Merchant Center feed for Shopping and PMax, and OpenAI's own product feed schema for ChatGPT product ads. - Use Google Search and Shopping to capture demand people already typed; use ChatGPT ads to reach the conversation before or instead of that search happens. ### How does ChatGPT's context matching differ from Google's keyword auction? Google Search bids on keywords a person types, using three match types, broad, phrase and exact, that trade reach for precision. ChatGPT ads carry no keywords at all: an ad group holds up to 2,000 free-text context hints, and the platform matches them to the meaning of a conversation, not to specific words a person typed. Google's auction runs on every search and is decided by six factors: bid, ad and landing-page quality, the expected impact of ad assets and formats, Ad Rank thresholds, search context and auction competitiveness, with Ad Rank recalculated for each individual search. ChatGPT ads have nothing equivalent published: an ad group carries a bid and a billing event, hints describe a situation, and there is no per-hint bid or negative hint list in the documented object model. The [guide on how ChatGPT ads differ from search and social ads](https://adsonomy.com/guides/how-chatgpt-ads-differ-from-search-and-social-ads) covers the ad-group model in more depth, and the [guide on context hints](https://adsonomy.com/guides/contextual-targeting-and-context-hints-in-chatgpt-ads) covers how to write them. ### How does ChatGPT compare with Google Shopping and Performance Max? Google Shopping and Performance Max are feed and AI-driven rather than keyword-driven: Shopping ads use Merchant Center product data instead of keywords to decide how and where to show, and Performance Max combines bidding, budget, audiences and creative through Google's AI across YouTube, Display, Search, Discover, Gmail and Maps from a single campaign. ChatGPT ads have no equivalent cross-surface campaign: every campaign only reaches conversations inside ChatGPT, using a chat card or a product ad built from a feed. Performance Max treats the audience signals you provide as a guide, not a restriction, and Google states it may show ads to relevant audiences outside those signals if they have a strong likelihood of converting; it can also auto-generate extra headlines and descriptions from your landing page and existing assets, with an opt-out. ChatGPT's product ads are built from OpenAI's own product feed schema, a different format from a Merchant Center feed even though both carry similar item data such as price and availability; the [guide comparing the two feed formats](https://adsonomy.com/guides/openai-product-feed-vs-google-merchant-center) covers the required fields on each side. ### How do the ad formats compare? Google's formats scale with the campaign type: Responsive Search Ads take 3 to 15 headlines of up to 30 characters and 2 to 4 descriptions of up to 90 characters, Shopping ads are generated from feed data, and a Performance Max asset group can hold 3 to 15 headlines, 1 to 5 long headlines, 2 to 5 descriptions, images, logos and video. ChatGPT ads have exactly two formats: a chat card, with a title of 3 to 50 characters, a body of up to 100 characters and a square image of at least 640 by 640 pixels, and a product ad template built from a feed item, with at most one non-archived product ad per feed ad group. There is no video format and no rotation across many creative variants inside one ChatGPT placement, unlike a Performance Max asset group that can carry up to five videos and up to twenty images each for its horizontal and square ratios. Every ChatGPT ad is reviewed before it can serve, and editing a creative sends it back to review; every new ChatGPT campaign, ad group and ad also starts paused. ### How does Smart Bidding compare with ChatGPT Ads bidding? Google's Smart Bidding strategies, Target CPA, Target ROAS, Maximize Conversions and Maximize Conversion Value, are all machine-learning bidding that sets a bid for every individual auction from historical data plus auction-time signals such as device, location and time of day. The ChatGPT Ads API offers three strategies only: fixed_bid, maximize_clicks and maximize_conversions, with no target CPA or ROAS field and no bid cap on the maximise strategies. Billing differs too. A ChatGPT campaign chooses a bidding type of impressions, clicks or conversions, and a conversion-optimised campaign billed per impression, generally available since 16 September 2026, is the option OpenAI recommends; amounts are set in micros. Google keeps Manual CPC as a direct bid alongside Smart Bidding; enhanced CPC, the older hybrid, stopped being available for Search and Display campaigns from the week of 31 March 2025. Because ChatGPT ads have no target-based strategy, matching a cost-per-order or ROAS goal there means stepping fixed bids and budgets from the outside, which the [guide on budgets, pacing and bidding strategies](https://adsonomy.com/guides/budgets-pacing-and-bidding-strategies-for-chatgpt-ads) sets out. ### How does measurement and conversion tracking compare? Both platforms now push conversion tracking server-side. Google's Enhanced Conversions send hashed first-party data, SHA-256, normalised before hashing, either via the on-site tag or server-side through the Google Ads API up to 24 hours after the conversion. OpenAI's Conversions API works the same way: PII hashed SHA-256, deduplicated against Pixel events by a shared event ID, with a timestamp required to be within the last seven days. Attribution windows are documented differently. Google Ads offers last-click and data-driven attribution, with data-driven the default for most conversion actions; older models such as first-click and linear are no longer supported. ChatGPT's click attribution window is set per conversion event with an attribution_window_days field that has no stated default, and insights requests carry no window parameter of their own. View-through conversions are reported as a separate figure rather than folded into clicks; since 16 September 2026 Ads Manager lets you choose a 7, 14 or 30-day click window and a 0 or 1-day view window per column set. The [guide on measuring ChatGPT ads](https://adsonomy.com/guides/measuring-chatgpt-ads-ga4-utm-incrementality) covers GA4, UTMs and how to reconcile the two systems' numbers. ### What controls does Google Ads have that ChatGPT Ads does not? Google Ads has a deep control layer built up over two decades: negative keywords at campaign or account level, ad scheduling by hour and day, device bid adjustments, frequency capping on Display and Video, and campaign drafts and experiments for testing a change before it goes live. None of these are documented in the ChatGPT Ads API. What ChatGPT ads do have at account level is a set of spend controls: up to 60 non-overlapping spend limit windows, and a daily spend limit for postpaid invoice accounts, narrower than Google's per-keyword and per-schedule controls but real. Shaping delivery hour by hour or device by device has to be built on top of the ChatGPT Ads API by a management layer instead. ### When should I use Google Ads, when ChatGPT ads, and when both? Google Ads is the better choice when there is already a meaningful volume of people typing relevant searches, when a large product catalogue benefits from Performance Max's cross-surface automation, or when a mature conversion history exists for Smart Bidding to learn from. ChatGPT ads are a smaller, single-surface, newer channel, without dayparting or a target-based bid strategy, better suited to reaching a conversation that a keyword search would never capture. Most advertisers who run both are not choosing one over the other: Google Search and Shopping capture demand people already typed, and ChatGPT ads reach people earlier, or elsewhere, in the same decision. Because the two systems share almost no mechanics, running both means keeping two separate playbooks rather than porting one campaign structure across. Comparison: Google Ads versus ChatGPT Ads. - Intent signal: keywords typed into search, broad, phrase or exact match (Google) versus free-text context hints, up to 2,000 per ad group, matched by meaning (ChatGPT). - AI-managed campaign type: Performance Max, one campaign across YouTube, Display, Search, Discover, Gmail, Maps (Google) versus none, every campaign reaches only ChatGPT conversations (ChatGPT). - Ad formats: Responsive Search Ads, Shopping ads, Performance Max asset groups, text/image/video (Google) versus a chat card (title, body, image) or a product ad built from a feed (ChatGPT). - Bidding strategies: Smart Bidding (Target CPA, Target ROAS, Maximize Conversions, Maximize Conversion Value), Manual CPC (Google) versus fixed bid, maximise clicks, maximise conversions (ChatGPT). - Billing event: click, or impression depending on format (Google) versus impression or click, conversion campaigns billed per impression (ChatGPT). - Conversion tracking: Google tag plus Enhanced Conversions, hashed, server-side via the Ads API (Google) versus Pixel plus Conversions API, hashed, deduplicated by event ID (ChatGPT). - Attribution windows: data-driven attribution (default), last-click alternative (Google) versus attribution_window_days per conversion event, no default stated; Ads Manager click window 7, 14 or 30 days and view window 0 or 1 day; view-through reported separately (ChatGPT). ## Frequently asked questions Q: Can I run something like Performance Max on ChatGPT ads? A: No. Performance Max is a Google-specific campaign type that spans YouTube, Display, Search, Discover, Gmail and Maps from one campaign. The ChatGPT Ads API has no cross-surface campaign type: every campaign only reaches conversations inside ChatGPT, through a chat card or a feed-based product ad. Q: Does ChatGPT Ads support Target CPA or Target ROAS bidding? A: No. The Ads API only offers fixed_bid, maximize_clicks and maximize_conversions as bidding strategies, with billing per impression or click. There is no target CPA, target ROAS or bid cap field on the maximise strategies, unlike Google's Smart Bidding. Q: Can I use my Google Merchant Center feed for ChatGPT product ads? A: No, not directly. ChatGPT product ads are built from OpenAI's own product feed schema, uploaded over SFTP or a hosted URL, which is a separate format from a Google Merchant Center feed even though both carry similar item data such as price and availability. Q: Does ChatGPT Ads have negative keywords like Google Ads? A: No. Google Ads supports negative keywords at the campaign and account level. The ChatGPT Ads object model has no documented negative hint list, so an ad group that matches the wrong conversations is corrected by rewriting its context hints, not by excluding terms. Q: Should I move budget from Google Ads to ChatGPT ads? A: Usually not as a swap. Google Search and Shopping capture demand people already typed, at a scale and with controls, such as dayparting and Smart Bidding, that ChatGPT ads do not have. ChatGPT ads reach a conversation that may happen before or instead of that search. Most advertisers run both rather than moving budget between them.