There’s a new study making the rounds that should get the attention of anyone responsible for a digital marketing budget. And not just because of what it says about ChatGPT advertising.
Researchers studying nearly 1,900 online retailers, representing $16 billion in annual sales, found that visitors arriving through paid ChatGPT advertisements performed significantly worse than visitors from Google ads, organic search, direct traffic, and even ChatGPT’s own unpaid recommendations.
How much worse? For every 100 visitors arriving through paid ChatGPT ads, the retailers generated an average of just $89 in sales, compared with $245 from 100 visitors arriving through paid Google search. That’s revenue, not profit, and it doesn’t even account for what advertisers paid for those clicks. Compared with direct visitors, ChatGPT ad traffic also converted at an 84% lower rate.
That’s a pretty sobering finding for a platform being hailed as the next big thing in digital advertising.
But before anyone declares ChatGPT advertising dead, there’s a much more important lesson here. And it has less to do with artificial intelligence than with how businesses make marketing decisions.
The Technology Isn’t Necessarily the Problem. The Assumptions Might Be.
Let’s be clear. This study doesn’t mean ChatGPT is failing as a marketing channel. It doesn’t mean AI-driven search is a fad. And it certainly doesn’t mean businesses should abandon efforts to make themselves more visible in AI-generated recommendations.
We’ve been advocating exactly the opposite.
The research actually reveals something interesting. People who arrive at a website through an unpaid ChatGPT recommendation behave quite differently from those who click on a paid advertisement within the same platform.
That distinction matters.
For years, marketers have understood that organic search visibility and paid search advertising serve different purposes. Now we’re discovering that the same may be true, perhaps to an even greater degree, in AI-driven discovery.
And there’s another important qualification. This research focused on ecommerce. What works, or doesn’t work, for an online retailer may look very different for a manufacturer selling complex equipment, a professional services firm, or a company with a lengthy B2B sales cycle.
So the takeaway shouldn’t be that ChatGPT advertising doesn’t work.
It should be that nobody should assume it works simply because it’s ChatGPT.
We’ve Seen This Movie Before
When I started Trivera in 1996, businesses were scrambling to get websites. Some had a clear understanding of what they wanted those websites to accomplish. Others just knew their competitors were getting them, and they didn’t want to be left behind.
Then came search engine marketing. Suddenly, buying keywords was the answer to everything. Never mind whether those clicks turned into customers.
Social media followed, bringing an obsession with followers, likes, and engagement. For a while, businesses were measuring their marketing success by how many people clicked a thumbs-up button.
And, remember when everybody thought their business needed a mobile app?
Every one of those technologies created legitimate opportunities. Some fundamentally changed how businesses attract and serve customers. And we embraced those changes, helping our clients take advantage of them.
But every one also created opportunities to waste enormous amounts of money.
Thirty years later, the technology has changed, but the temptation hasn’t.
Today, it’s AI. Companies are being told they need GEO, AIO, AI-powered content, AI agents, and now advertising inside AI platforms. Some of those opportunities are absolutely worth pursuing. Others need more testing. And some may never make financial sense for a particular business.
The challenge isn’t knowing what’s new. It’s knowing what’s worth doing.
From Predicting the Opportunity to Measuring the Reality
Back in 2024, we wrote about the possibility of ChatGPT becoming a serious competitor to Google as a search engine. We recognized that changing search behavior would eventually create new opportunities for marketers, including advertising.
In our subsequent articles about SEO, GEO, and AIO, we’ve consistently made the case that businesses need to prepare for AI-driven discovery without abandoning the proven foundations of digital marketing.
And earlier this year, we explained where Trivera stands on GEO and AIO: These emerging disciplines deserve attention, but they should complement sound SEO practices, credible content, and a broader digital strategy rather than replace them.
Now we’re getting actual performance data about one of those emerging opportunities.
Does that mean our position has changed? Not at all.
It means we have more information to work with.
That’s how digital marketing strategy is supposed to evolve. You recognize a change in the landscape, evaluate the opportunities it creates, test what makes sense, measure the results, and adjust your recommendations as new evidence becomes available.
There’s nothing inconsistent about recognizing ChatGPT’s enormous potential while questioning whether its current advertising model deserves a significant share of your marketing budget.
In fact, I’d be more concerned about an agency that wasn’t asking that question.
Your Marketing Partner Should Know When to Say No
Here’s where this gets important for business leaders.
The digital marketing industry has never been short on people eager to sell the next big thing. Every new platform, technology, and acronym seems to generate a new collection of experts promising extraordinary results.
And there’s nothing wrong with exploring new opportunities. At Trivera, we’ve spent three decades doing exactly that.
But there’s a difference between being an early adopter and being an early investor in something that hasn’t demonstrated its value.
A good marketing partner should be able to tell you not only what’s possible, but what’s practical for your business.
That means understanding your customers, your sales cycle, your competitive environment, and what you’re actually trying to accomplish. It means knowing which opportunities deserve a serious investment, which deserve a limited experiment, and which simply aren’t worth pursuing right now.
Sometimes the smartest recommendation is to try something new.
Sometimes it’s to double down on something that’s already working.
And sometimes it’s to tell a client that the exciting new technology everybody’s talking about probably isn’t where their next marketing dollar should go.
That last recommendation can be particularly difficult for an agency whose revenue depends on selling services.
But your marketing budget shouldn’t be funding your agency’s enthusiasm for the latest trend.
It should be funding activities that have a reasonable path toward producing meaningful business results.
And when the evidence changes, the recommendations should change with it.
What This Means for You
The ChatGPT advertising study isn’t a reason to stop exploring AI marketing opportunities. It’s a reminder that every new opportunity deserves the same scrutiny you’d apply to any other business investment.
Before committing significant resources to a new platform, technology, or marketing strategy, make sure whoever is guiding your digital marketing efforts can answer five questions.
1. Why this opportunity? How does it support your specific business objectives? Being visible on a popular platform isn’t necessarily the same as reaching the people most likely to become your customers.
2. Why now? What evidence suggests the opportunity is worth pursuing today? Is there a real business case, or are you responding to the fear of being left behind?
3. What are we measuring? Clicks, impressions, and visibility can be useful indicators, but they aren’t the finish line. What matters is whether your investment is moving prospects toward becoming customers.
4. How much are we willing to learn? Experimentation is essential, especially with emerging technology. But a smart experiment has a defined budget, measurable objectives, and a point at which you evaluate whether to continue.
5. What’s our next move? What results would justify increasing the investment? What would tell you it’s time to change direction? And what would convince you to walk away?
ChatGPT advertising may eventually become an enormously valuable marketing channel. Or it may never deliver the returns advertisers expect. The research we’re seeing today is an important early signal, not a final verdict.
Either way, your business shouldn’t have to gamble its marketing budget on finding out.
Being ahead of the curve doesn’t mean jumping on every bandwagon. It means knowing which ones are actually going somewhere.
That’s the kind of strategic judgment we’ve spent 30 years developing at Trivera. And if your marketing efforts are generating plenty of activity, but you’re not sure whether they’re generating meaningful business results, that’s a good place to start a conversation with us.
