Vibe coding in marketing? We surveyed 300+ marketing leaders on its adoption

tl;dr — Download the free vibe coding in marketing research I did with UserEvidence.

“When the person who feels the pain is also empowered to build the solution, the result is magical and utterly unique.”

That quote by David Peterson, from a post he wrote in 2020, always struck me as one of the most eloquent arguments for letting ops teams and savvy business users build their own apps, automations, and now agents. Back then, that was enabled by no-code tools, from (A)irtable to (Z)apier.

Today, we call it “vibe coding.”

(To preempt pedantic debate, I’m using the term loosely to describe people — particularly people who aren’t professional developers — building agents, apps, and automations primarily through conversational interfaces powered by AI.)

Vibe coding is, admittedly, a somewhat polarizing proposition. As was no-code before it. The idea of letting non-engineers build software stuff still strikes some people as a scourge rather than a superpower. (“1995 just called. They want their IT policy back.”)

I kid my IT friends. To be fair, there are legitimate concerns with security, data access, compliance requirements, reliability, continuity, etc., if you let anybody build anything, any time, any way they want. Brand, finance, and legal would all raise flags on that too.

But it’s a straw man argument. Nobody’s advocating for anarchy, blithely letting anyone build anything. Vibe coding can and should be done within guardrails and governance. Don’t be thrown by the “vibe” language. It’s not the business operations equivalent of the free love movement. It’s people solving problems and innovating their work, using the latest tools of their trade. That’s exactly what we want people to be doing.

“But non-engineers can’t build good software!” I call bull****. Sorry, but that blanket dismissal drives me nuts. First, most vibe-coded projects aren’t architecting the Eiffel Tower. They’re usually small, simple programs. Second, those builders know more about their domain — the problems to solve, the opportunities to seize — than any non-domain developer ever could, no matter how great of a software engineer they are.

It’s like IQ vs. EQ. Two very different kinds of knowledge. A vibe-coding business user has a structural advantage in their tacit know-how that AI can’t substitute. Judgment and taste rule there. Meanwhile, knowing how to write good code to implement a good idea… well, AI’s getting pretty good at that.

But enough of my strongly-held opinions. I brought data!

The wonderful team at UserEvidence and I collaborated on a joint research project to understand how vibe coding was being adopted in marketing orgs. We surveyed 302 marketing leaders — 55% from mid-market companies, 26% from larger enterprises. We had two qualifying filters:

Our instinct on that second filter proved right, as respondents reported pretty strong institutional adoption:

It’s an impressive level of organizational adoption given how new this whole vibe coding phenomenon is. Most respondents (57%) have only been using AI code generation or builder tools for less than a year:

(I know, the term “vibe coding” was only coined last February, if you’re wondering about the 12% who claim to have been doing this for more than 2 years. But the use of AI-powered no-code tools and GTM engineering had been on the rise before that.)

One interesting finding was how confidence in a team’s vibe coding capacity developed over time. It started high — 44% very confident. (Marketers tend to be a confident lot by nature.) But instead of steadily increasing, it hovered there for a while. One year in, the number who reported hitting walls on complexity then peaked at 46%. It was only after 2+ years that participants reported a significant uptick in their confidence (64%).

Kinda maps to the Gartner Hype Cycle, right?

So what are marketing teams vibe coding? The answer is straight from David Peterson’s quote: they’re building solutions to the pains they feel. And in the complex, messy world of martech, there’s no lack of aches and afflictions to attack.

The top three use cases participants said they were vibe coding:

These are almost all internal applications, the backstage and boiler room work by which modern marketing runs. But for years, marketers and marketing ops teams have been constrained by the way individual martech products think they should do workflows and automations — coincidentally tilted toward the interests of each individual vendor.

Because martech stacks are a mix of many different products, it was hard for marketers to implement processes that spanned across all those differently opinionated platforms. But with vibe coding, marketers can rise above that. They can design workflows the way they want work to flow.

Of course, to do that, you need integration across those products. And in every survey of martech challenges I’ve seen for the past decade, integration has always been at the top of the list. Again, most vendors offer their own pre-built integrations, but with a subset of partners and a subset of functionality. Vibe coding lets marketers route around those obstacles and work directly with each product’s APIs — or MCP servers (75%).

Less common for now is vibe coding external projects, such as creating landing pages or microsites (28%) or building customer-facing apps (17%).

One explanation for that is caution. Marketers want to hone their vibe coding chops before they take them front-of-house. But I hypothesize it’s because that’s less about relieving pain and more about innovating new kinds of marketing. It’s a new dimension of creativity, and it will take a little time to discover the possibilities.

These are the capabilities that participants in our survey said vibe coding enabled for them that weren’t possible or practical before:

I’d sum it up as more speed with fewer dependencies.

As the pain relief theme suggests, a lot of vibe coding is being applied toward making existing tasks better, faster, easier. So we asked marketers what they were doing before:

In most cases, it was manual elbow grease (54%). Spreadsheet jujitsu and copy-paste karate. However, there were substitutions that speak to the economic disruption of AI on labor and the SaaSpocalypse:

This is solid evidence of a shift in the build vs. buy calculus. Marketers aren’t building their own CRMs or customer engagement platforms. (And I wouldn’t recommend it. That’s more Eiffel Tower software architecture.) But they are building bespoke and niche solutions that previously they would have purchased a specialized tool for or outsourced to a service provider.

Let’s hug the elephant in the room. Vibe coding can be governed, should be governed, and for most of our study participants is being governed. 83% said their organization has a formal policy or governance framework specific to the marketing team’s use of vibe coding.

The majority have an approved tool list, data access or privacy restrictions, and a QA process before deployment:

Even more encouraging is that there was perfect correlation between governance and maturity. 92% of the marketing orgs who reported that vibe coding is fully embedded as a part of how their team routinely operates have formal governance in place.

And one of the most impactful ways of developing that maturity is through formal training and onboarding. Marketing teams who do that are more likely to become very confident in their capabilities (52% vs. 44% overall) and to see ROI within days or weeks (41% vs. 30% overall).

And to dispel the shadow IT myth, the people marketers primarily turn to for guidance? IT or engineering colleagues (53%).

Vibe coding in marketing isn’t about aimless tinkering or novel navel-gazing. The early adopters are sharply focused on driving value:

And it doesn’t take long to see returns from their investment in time and effort. 43% reported that it only took 1-3 months before the practice delivered real value — and 32% reported achieving that within days or weeks.

It’s therefore not surprising that essentially all (89%) expect their usage of vibe coding to expand in the next 12 months, with 43% expecting to expand usage significantly.

I’ll close with one more telling data point. The marketing orgs most likely to see fast ROI were those where people within marketing led the charge to embrace vibe coding. 47% of those achieved fast ROI in contrast to only 30% when vibe coding was introduced by IT. The lowest success rate (14%) was when it was pushed by a vendor.

When the person who feels the pain is also empowered to build the solution, the result is magical and utterly unique. It delivers real value. And it can be well-governed.

There is a wealth of more data and insights in our full research report, which you can download for free. I’ve only covered a fraction of it here. It was a real joy to collaborate with UserEvidence on this project, and I'm so grateful to everyone who participated. This is the data I was looking for.

Vibe check?

Scott