ChatGPT: 7 Inspiring Ways to Unlock Success

On October 5, OpenAI announced that visual ads will appear inside ChatGPT while users generate images. A first group of U.S. advertisers will be testing later this month. In the same announcement came the measurement stack: ten attribution partners (AppsFlyer, Triple Whale, Northbeam and Kochava among them). There are also conversion-data integrations with Hightouch, Tealium and LiveRamp, and geo-incrementality testing through Haus, Measured and WorkMagic. All of it pointed to a platform with 1.2 billion weekly users.

The conventional executive read is that ChatGPT has become another performance channel. The CFO will ask for ROAS, and the media team will fund it out of search. A test will run for a quarter against cost per acquisition. That read is wrong. It will quietly cost some companies a great deal of money. OpenAI is selling a brand-building moment with activation instruments. The ChatGPT visual ads strategy will be decided by whether leadership notices the mismatch before the first dollar is spent.

Binet and Field Would Call This a Category Error in the Making

Les Binet and Peter Field spent years mining the IPA Databank for what actually makes marketing effective. Their central finding, the heart of The Multiplier Effect, is that brand building and short-term activation are complements, not substitutes. Brand work builds mental availability, the odds that you come to mind before anyone is in-market. Activation harvests demand that already exists. The evidence-backed default for consumer categories is roughly 60 percent brand and 40 percent activation, and closer to 55/45 or even 50/50 in B2B.

Think of activation as cardio: you see it on the scale this week. Brand is strength training: nothing shows up this week, but it raises the metabolic rate for every workout afterward. Binet and Field’s uncomfortable corollary is that anything measured inside a short attribution window will flatter the cardio and starve the strength work. Only one of them can be seen in the dashboard. Companies do not choose to underinvest in brand; their measurement chooses for them.

Now look at where OpenAI is placing the ad. A person generating an image is imagining something: a kitchen, a logo, a trip, a gift. The company says its ads will show how products and services could fit into daily life. That is not a coupon at checkout. It is closer to a billboard inside a daydream, which is exactly the emotional, creative, broad-reach territory that builds memory structures. Yet, the infrastructure announced alongside it is built to answer one question: did a conversion happen?

The incrementality partners are a genuine step forward because they test whether ads caused sales rather than merely preceded them. But they still count conversions inside a window. Nothing in the announced lineup is designed to see what a brand impression does six months later.

What I’ve Seen When the Dashboard Decides the Strategy

After twenty years and several nine-figure businesses, the pattern is consistent. The budget line that cannot defend itself on the dashboard is the first to go, and it goes in the quarter the board is most anxious. At Bolt Goodly, I watched sophisticated clients with beautiful performance reporting cut brand spend not because they believed brand did not work, but because it had no row on the scoreboard. Within a few quarters, their cost per acquisition crept up, and the very channels they trusted were harvesting a pool of demand that nobody was refilling.

At Pinnacle, the discipline behind the 60 percent cost reduction was deciding in advance which spend we would defend on principle rather than on in-quarter attribution, and what we would accept as evidence for it. That was a strategy decision made before the measurement existed, and it had to be, because the measurement would never have volunteered the answer.

Fund It From Brand Budget, or Don’t Fund It

Here is the position I would defend in front of any board. The budget line that pays for ChatGPT image ads decides how it will be judged. That decision is more important than the creative, the targeting or the bid. If it comes out of search, it will be judged against the most high-intent, bottom-funnel inventory ever built. It will look expensive and will be cut after two quarters, which is exactly how brand spend gets killed. If it comes out of the brand budget, then leadership has to agree up front on a different scoreboard: reach, recall, and the lift in branded search and direct demand over time.

This cuts against the standard playbook, which says to put one or two percent of budget into a test and read it on the channel’s native metrics. That is precisely the setup that guarantees the wrong answer. The ten attribution vendors in the announcement are not villains, but they are an ecosystem that earns its living on the activation reading. A platform that arrives with plumbing for one half of the Multiplier Effect will always make the other half look optional.

I should be clear about what I do not know. The format has not launched, there is no outcome data, and image-generation ads may turn out to behave nothing like a brand format. That is the point. A CMO who has not decided in advance what she is buying will be unable to interpret whatever the data says.

The Question for Your Next Board Meeting

The channel is new, but the failure mode is thirty years old. Efficient measurement makes activation look like the whole of marketing. The companies that confuse the two end up harvesting demand they have stopped creating. So before a dollar goes into ChatGPT’s image-generation ads, can you tell your board whether you are buying brand or buying conversions? Which scoreboard will you accept a year from now if the answer is brand?