Flight log / Transmission 003

ChatGPT ads vs. Google Search ads: what changes when the answer is the ad

Google sells the question. ChatGPT sells the moment after it. One auction has 26 years of tooling behind it; the other opened self-serve in May 2026.

If you own a paid media budget, the question isn't whether ChatGPT advertising is interesting. It's whether it deserves dollars that would otherwise go to a channel you already know how to run. This post is that comparison, one dimension at a time.

New to the platform? Our ChatGPT ads guide covers the basics. Here we assume you know how Google Search works and want to know what's different.

ChatGPT advertising vs. Google Search ads at a glance

Platform details as of October 2026, from OpenAI's help center and Google Ads Help. Both change often.

Which has the stronger intent signal?

Google's signal is short and explicit. "SOC 2 compliance software pricing" tells you exactly what the person wants and roughly where they are in the buying process. Decades of keyword strategy exist to read that signal well.

ChatGPT's signal is long and implicit. Someone might spend five messages explaining that they're a 30-person startup, an enterprise prospect just sent a security questionnaire, and they have a quarter to get audited. They never typed a keyword. OpenAI selects ads primarily on relevance to the context and intent of the conversation.

So which is stronger? It depends on the stage. For "I know what I want," the query wins: it's precise and you can bid on it directly. For "help me figure out what I need," the conversation wins, because it holds context a query can't.

The line is blurring from Google's side too. Google now shows ads in AI Overviews and has been testing them in AI Mode, and in May 2026 Search Engine Journal spotted a help doc update saying search terms tied to AI experiences may reflect inferred intent rather than the literal query. More on that in our breakdown of ads in AI Overviews.

How much control do you give up on ChatGPT?

A lot, and that's the honest headline for budget owners.

On Google you choose keywords and match types, add negatives, read the search terms report, set bid strategies and layer audiences. On ChatGPT you write context hints, which OpenAI says are not exact-match keywords or targeting rules and do not guarantee delivery. Your levers are the hints, the ad copy, the landing page, location, platform (Android, iOS, desktop web), custom audiences, budget, objective and bid strategy. Even bidding leans automatic: since August 2026, eligible new ad groups default to Maximize results, and Manual: Max bid is the only hard ceiling.

The practical effect: on Google, you can usually diagnose a bad week down to a handful of search terms. On ChatGPT, you'll see that an ad group underdelivered or underconverted, and you'll have to reason about why from your hints and copy. That's slower learning, which argues for smaller, longer tests.

How does measurement compare?

Both platforms can tell you about conversions. The difference is what they can tell you about the click before it.

• ChatGPT: a measurement pixel, a Conversions API that accepts offline events with hashed user data, and UTM support. Reporting covers impressions, clicks, spend, CTR, average CPC, average CPM and conversions, in aggregate. View-through attribution is a fixed 1 day.
• Google: conversion tracking, offline conversion imports and the search terms report, plus years of third-party tooling built around it. AI Max adds reporting that labels which search terms and landing pages came from AI Max.

For B2B the gap that matters most is the search terms report. It's how you find wasted spend, new negatives and new angles. ChatGPT has no equivalent, so your CRM becomes the only place you can see whether the traffic was right.

Who will you actually reach?

Google Search reaches almost anyone with a question. ChatGPT ads reach logged-in adults on the Free and Go plans in supported markets. OpenAI doesn't show ads to Plus, Pro or Business subscribers, or to users it knows or predicts are under 18.

For B2B that cuts both ways. The senior buyer whose company pays for ChatGPT Business won't see your ad. The analyst or manager doing early research on a free account will. The pattern that shows up in most B2B accounts is that the person who builds the shortlist isn't the person who signs, and ChatGPT ads mostly reach the first one.

What changes in the creative?

A Google responsive search ad gives you up to 15 headlines of 30 characters and 4 descriptions of 90, which Google mixes and matches. A ChatGPT ad is one title, one description, an image and your logo. OpenAI doesn't publish character limits and third-party numbers disagree, so trust the counter in Ads Manager. What OpenAI does ask for is volume: several distinct title and description variations per offering. Instead of Google mixing your assets, you hand ChatGPT a stack of finished ads and let it pick.

Two writing shifts follow:

1. Name the need, not the category. The person already explained their situation to ChatGPT. Your title should sound like the next sentence in that conversation.
2. Offer the next step, not the final one. A checklist or benchmark fits someone mid-research better than "book a demo."

The landing page matters more on ChatGPT, not less. OpenAI uses landing page content as an input for relevance, and the visitor arrives with more context than a searcher. A generic page drops the thread on both counts.

Which costs more per click?

Headline CPCs look similar. OpenAI's help center recommends a $3 to $5 starting max CPC for Clicks campaigns. LocaliQ's search advertising benchmarks (page updated June 2026, based on thousands of its customers' Google and Microsoft Ads campaigns) put the average CPC at $5.42 across industries and $5.87 for business services.

Don't stop there. A cheaper click from someone who isn't ready to buy can cost more per qualified lead than an expensive click from someone who is. The only comparison worth making is cost per qualified lead, or pipeline per dollar, measured in your own CRM.

How should you split a test budget?

Our judgment, not a platform rule: fund ChatGPT as a capped test, not a reallocation. Take it from new money or from the weakest-performing slice of search, never from your best-converting brand or high-intent campaigns.

A worked hypothetical: say a 20-person SaaS team spends $30,000 a month on Google Search and is limited by search volume, not budget. Carving out $2,000 to $3,000 a month for ChatGPT is about $65 to $100 a day, comfortably above OpenAI's $25 minimum daily budget for USD accounts, across two or three ad groups for six weeks. That's tuition, not a forecast.

If the test looks promising, prove it's incremental before you scale it. A geo holdout will tell you more than any platform dashboard.

Which use cases fit each channel?

Nobody gets knocked out here. Google keeps the "I know what I want" moment. ChatGPT is contesting the research phase that used to sprawl across ten tabs. Most budgets will end up funding both, in very different amounts.

What we don't know yet

• B2B results at scale. Public ChatGPT ads wins so far are mostly ecommerce and self-reported. No credible B2B benchmark exists yet.
• Price stability. ChatGPT pricing has moved several times in 2026. Today's CPC comparison may not hold next quarter.
• How Google's AI surfaces change the comparison. As Google puts more ads inside AI answers, the gap in "conversation vs. query" narrows.
• Assisted impact. If ChatGPT ads shape the shortlist and the buyer later converts through a Google brand search, last-click reporting will credit Google.

Ready to run the test? Our step-by-step guide on how to advertise on ChatGPT walks through setup, tracking and the first 30 days.

FAQ