The first question marketers ask about every new advertising channel is almost always the same: Does it perform better than the last one we tested?
It happened with social. It happened with retail media. And now it’s happening with OpenAI.
The debate around ChatGPT advertising has quickly become a comparison with Google Search. Can it deliver lower CPAs? Does it convert as well? Is it ready to compete for performance budgets?
They’re all understandable questions. But they’re the wrong ones.
It’s a mistake to assume conversational AI should behave like search in the first place. The customer journey is fundamentally different. If marketers continue evaluating ChatGPT using search-era metrics, there’s a real risk they’ll dismiss one of the most significant new advertising channels before they’ve understood how it creates value.
Search captures intent. Conversational AI shapes it.
Google built the world’s most successful advertising business by harvesting intent.
If a user searches for “cheap hotels in Hawaii,” Google shares a list of relevant results. The commercial intent is already there. The advertiser’s job is simply to be the most relevant answer when the purchase decision is made. But conversational AI works differently.
Instead of typing in a few keywords, a user might go into detail and explain that they’re travelling to Hawaii for a friend’s wedding, that they’re on a tight budget but want their own room, oh, and they’d like to surf while they’re there. That’s no longer a search query; it’s a conversation. And that difference matters.
Rather than simply capturing existing demand, conversational AI helps users explore options and refine priorities before they buy. When the traveler eventually visits the website to book, they’ve often already used AI to compare all the alternatives and understand exactly what they’re looking for.
The advertising opportunity is participating in how intent develops.
Higher acquisition costs don’t mean lower value
One of the first things marketers notice when they begin testing conversational AI is that the acquisition metrics can look uncomfortable.
CPAs can appear higher than those achieved through search. For anyone used to judging performance through traditional attribution, that’s an easy reason to conclude the channel isn’t competitive.
That was my first reaction, too.
If it looks like search and behaves like search, surely it should produce similar economics. But once you step back and look at the customer journey rather than the click, that conclusion becomes much harder to support.
A user spending several minutes discussing products, asking follow-up questions and refining their requirements is behaving very differently from someone typing a short keyword into Google. They’re effectively qualifying themselves before they ever reach an advertiser’s website.
That means acquisition might cost more, but the customer arriving at the other end can also be significantly better informed and a much stronger commercial fit.
Judging conversational AI purely on CPA risks confusing efficiency with effectiveness.
We’ve spent 25 years optimizing for clicks
The measurement frameworks we’ve inherited aren’t suited for where the consumer journey is going.
For more than two decades, performance marketing has been built around clicks because clicks were observable. Attribution models rewarded whichever interaction happened closest to the conversion, giving marketers a practical way to optimize campaigns at scale.
That made perfect sense in an internet dominated by search engines, display advertising and relatively linear customer journeys.
Conversational AI changes that, because the interaction influencing a purchase often happens long before a click ever occurs. And when the eventual conversion happens later through another channel, traditional attribution naturally gives the credit elsewhere.
That doesn’t necessarily mean the initial conversation had little value; it simply means the measurement framework wasn’t designed to capture it.
Early evidence suggests that when marketers move beyond attribution and begin measuring incremental business outcomes instead, conversational AI looks very different from the picture painted by last-click metrics.
AI advertising can work very effectively. But unless marketers are measuring it in a way that reflects how the channel actually influences buying decisions, it can create a colossal education gap.
The industry is asking the wrong question
The conversation today largely centers on whether OpenAI can compete with Google.
A far better question is whether conversational AI should be evaluated like search at all.
Every evolution in advertising has forced marketers to rethink measurement. Social wasn’t simply another display channel. Retail media wasn’t just paid search inside a retailer’s website. CTV couldn’t be judged solely through linear TV metrics.
Conversational AI deserves the same mindset shift.
Google monetized search intent. Conversational AI has the potential to monetize the conversation that creates it.
Those are fundamentally different commercial models. If marketers continue measuring one using the rules of the other, they risk underestimating not only the performance of OpenAI ads, but the role conversational AI is likely to play in the future of digital advertising.
“Data-Driven Thinking” is written by members of the media community and contains fresh ideas on the digital revolution in media.
Follow INCRMNTAL, a Smartly company and AdExchanger on LinkedIn.
For more articles featuring Maor Sadra, click here.
