State of MarTech: Marketing Automation

 

For a long time, marketing automation has focused primarily on automation itself. Increasingly, marketers are asking whether that automation is worth its cost and complexity.

While automation vendors have dramatically expanded what their platforms claim to do, marketing teams are now being forced to determine which of those capabilities actually deserve to be automated. While faster email production, summaries, campaign setup, and workflow creation can deliver productivity gains, but they do not automatically produce better commercial outcomes.

The future of marketing automation may instead be tied to better automation economics, and better decisions about what deserves automation in the first place.

 

Starting at the journey’s beginning

Greater scrutiny of automation costs is raising a very simple question – just because a process can be automated doesn’t mean it should be.

That thinking is evident at online travel agency TravelOnline, which has spent years developing much of its marketing automation capability in-house, from campaign creation and behavioural tracking through to automated customer journeys.

According to marketing manager Sherri Adamson, after years of adding capability, its experience has led to a relatively simple conclusion – that more automation does not necessarily mean better marketing.

“The biggest thing is to start with your customer journey – not the automation platform,” said Adamson.

“It is really tempting to jump in there and build lots of sophisticated workflows, but if they aren’t solving a real customer need, then I think they quickly become an unnecessary complexity.”

The lesson was to “start small and focus on a handful of high-impact journeys and get those performing really well before you start expanding”.

That approach turned the traditional measure of automation maturity on its head, with sophistication judged by whether individual journeys solved a genuine customer problem and performed well enough to justify their continued existence, rather than simply by the number of workflows a team could create.

 

Capability versus value

Investment in marketing automation also continues to grow rapidly. Grand View Research valued the global market at US$6.7 billion ($9.4 billion) in 2024 and expected it to reach US$15.6 billion ($21.8 billion) by 2030. In Australia, it forecast even faster growth, from US$190 million ($266 million) in 2024 to US$545 million ($763 million) by the end of the decade.

Money is also beginning to flow into the next generation of automation. Grand View Research estimated the AI orchestration segment of marketing automation generated US$1.8 billion ($2.5 billion) in 2025 and expected it to approach US$8 billion ($11.2 billion) by 2033.

Yet greater investment and expanding capability do not necessarily translate into better commercial outcomes, even with the addition of AI.

Gartner vice president analyst Ben Bloom described the current market as “peak hype”, with the gap between technological promise and demonstrated value at its widest.

“Most marketing automation tools are still stuck in the world of lead management and manual, deterministic journey orchestration, while marketing and sales are still unsure about how to transform themselves to both collaborate around qualifying and converting demand into customers,” Bloom said.

“Gen AI assistants are helping somewhat in this market, but marketers we speak with still struggle to create testable AI native use cases and properly screen for value instead of things that sound advanced or exotic.

“Early AI use case successes have been typically the use cases that don’t necessarily directly intervene in the customer journey, like creating scaled content and assets and variations, rather than fully automated Agentic marketing operations.”

IBRS analyst and advisor Dr Joseph Sweeney said the challenge was distinguishing automation that looked sophisticated from that which could actually demonstrate a commercial benefit.

“I recommend customers ask this question: what does the system decide that the team could not already decide, and where is the control group?” Sweeney said.

He pointed to service deflection, predictive offer allocation, and self-updating CRM as examples that could pass that test, but said most of the remaining agentic use cases lacked published control-group evidence.

“In short, look at the evidence,” he said.

 

The new economics of automation

The need for stronger evidence is becoming more important as the economics of marketing automation also change.

According to Bloom, many marketing automation platforms have historically been priced around relatively predictable measures such as contacts or database size. But Bloom said AI was introducing consumption-based pricing based on credits, tokens, and activity, often layered on top of existing SaaS costs.

“Our research so far has found that marketers have immature guardrails around consumption-based pricing which puts total cost of ownership at risk, especially when legacy SaaS pricing models are mixed with Gen AI or Agentic AI features that are priced based on credits or tokens,” Bloom said.

That creates a different management challenge. As the cost of automation becomes increasingly tied to usage, marketers would need greater visibility into not only whether an automation works, but what it costs to run and whether the resulting benefit justified that expenditure.

Bloom expected this would push marketing teams towards disciplines already familiar to technology leaders.

“We envision leading marketing teams as succeeding when they adopt approaches similar to CIO FinOps,” he said.

Gartner predicted that by 2029, marketing FinOps would reduce waste by 30 percent and enable organisations to fund three times as many journeys from the same budget.

The introduction of consumption-based pricing could also change how marketers thought about automation itself.

While adding another workflow or journey to a traditional system generally carried little additional marginal cost once the platform had been purchased, the use of automated actions that consumed credits or tokens could give marketers a stronger incentive to understand what each automation cost and what it contributed.

Sweeney said organisations needed “bounded objectives with named owners”, “hard limits on budget, audience, and action scope”, and “cost observability per agent run”.

They also needed “a measurement design agreed before launch, with a kill criterion”.

“None of these can be bought as a product or feature, but all must be factored into new martech product buying decisions,” Sweeney said.

 

Value over volume

That suggests maturity in marketing automation may increasingly be defined not by the volume of processes automated, but by an organisation’s ability to identify, measure, and remove low-value automation.

The cofounder of the B2B marketing community Generate, Lara Vandersluis, said that was also reflected in what marketers were achieving with AI today.

“ROI is real, but it’s concentrated in specific well-scoped use cases rather than a broad transformative uplift,” she said.

In that environment, the next phase of marketing automation may be less about finding more processes to automate, and more about becoming disciplined enough to know which ones are worth keeping.