ChatGPT Ads Complete Guide: Campaigns, Tracking, and Measurement
What ChatGPT Ads Is
ChatGPT Ads shows sponsored placements inside ChatGPT conversations: ads render below the end of a response, clearly labeled, visible to users on the Free and Go plans. OpenAI began testing ads in the US in February 2026 and opened self-serve buying through the ChatGPT Ads Manager at ads.openai.com in July. The platform is younger than any incumbent - accounts, features, and documentation are still evolving - but the core loop is already complete: campaigns, targeting, creative, measurement, and reporting all work end to end. This guide is the hub for the whole cluster: each section covers one layer and links the deep-dive guide for it.
Campaigns, Objectives, and Budgets
A campaign carries the objective, budget, dates, and targeting; ad groups hold context hints and bid strategy; ads carry the creative. The campaign setup guide walks the full flow, and the objectives deserve a line each here: Views buy impressions at CPM, Clicks buy valid clicks at CPC, and Conversions - oCPC - keep billing clicks while optimizing toward one tracked conversion event, as the oCPC guide details. Ad groups additionally carry the automated Maximize results bid strategy, which adjusts bids toward the campaign goal without a manual bid cap. Budgets run as campaign-total or daily, with the daily variant averaged over a rolling seven-day window capped at twice the daily amount per day and seven times per week (US minimum $25 daily). The one-way doors to remember: objective and budget type lock at creation, and Meta-style always-on daily optimizations do not map one-to-one onto this platform yet.
Targeting: Context Hints Instead of Keywords
The targeting guide covers the model in depth: OpenAI selects ads from conversation context and intent, your creative, advertiser-provided context hints, and campaign targeting selections, ranked by relevance and bid. Geography runs at country level with US state, DMA, and ZIP granularity; platforms split into the iOS app, Android app, and web. Custom audiences - uploaded customer and prospect lists - can be included or excluded per campaign, with the custom audiences setup covering the 25,000 matched-user threshold, file formats, and the EEA restriction. Context hints deserve the most care of the three: they describe who the product helps and when, and OpenAI is explicit that they are not exact-match controls.
Measurement: Pixel, Conversions API, and Events
The measurement layer follows the same two-source architecture Meta made standard. The setup installs the OpenAI pixel and wires the server-side conversions API - a POST to bzr.openai.com with your Pixel ID and key - and the two copies deduplicate through a shared event_id. The event vocabulary is OpenAI's own: twelve standard events plus custom, with exact data shapes and minor-unit money rules in the standard events guide. Attribution adds platform-specific behavior every practitioner should know from the attribution guide: click-attributed conversions only in the main metric, a separate fixed one-day view-through metric, a 24-48 hour reporting delay, and modeled conversions where available.
Click IDs and the Funnel
OpenAI appends its click reference, oppref, to every landing URL, and the pixel stores it in a first-party cookie for later matching. The oppref guide covers the full mechanics; the operational summary is that every funnel hop must carry the parameter forward, and server events should return it to OpenAI whenever available. Funnel builders and trackers are where oppref dies - which is why the tracker and server integrations below matter as much as the campaign itself.
Integrations: Trackers, MMPs, sGTM, and the API
Four integration paths cover the professional stacks. Trackers: RedTrack ships a preset, and the Keitaro integration guide builds the manual equivalent - capture oppref in a custom traffic source, then bridge postbacks to the CAPI. Mobile: AppsFlyer and Adjust are the officially supported measurement partners with click-based attribution. Infrastructure: a server-side GTM container can deliver CAPI events through a custom template on Google's sendHttpRequest API. Programmatic: the OpenAI Ads API manages campaigns, ad groups, ads, and insights - one key per ad account - while feed management and the product feed workflow stay in Ads Manager with a delta API for price updates. Multi-brand sites get a dedicated multiple pixels guide for separating advertisers on shared pages.
Validation and Troubleshooting
The platform surfaces its own diagnostics: ad statuses with hover reasons, plus a per-data-source event quality score with named warnings - missing customer identifiers, lost click information, delayed server events, matches needing review. The troubleshooting guide turns both into a checking order: tooltip, setup, delivery delay, event quality, support last. Fire one conversion, confirm it lands attributed and deduplicated in Ads Manager, then spend - and when something does not land, the troubleshooting guide and the Pixel Activator validation flow cover the diagnosis.
Where ChatGPT Ads Fits Against Google and Meta
The full comparison guide keeps the comparison on documented facts: intent capture models, targeting primitives, billing, and measurement maturity. The short version for budget holders: existing demand belongs to Google, audience scale belongs to Meta, and ChatGPT Ads earns a test allocation sized for learning - which only pays off when the measurement pipeline underneath it works. Beyond paid, the search visibility guide covers the organic side: which OpenAI crawlers to allow and how to track the referrals ChatGPT search sends.
Automating Delivery: Pixel Activator and MOST
The free Pixel Activator validates one OpenAI Ads event per landing via pixel.way2.us, and MOST delivers tracked conversions server-side to ChatGPT Ads alongside Meta, TikTok, Reddit, Snapchat, and Pinterest from one postback - covering the tracker and CAPI legs this guide described for teams that would rather not build them.
