Here's your ungated copy of The State of Martech 2026 report

Want to skip straight to the goods? Here you go, completely ungated:

Download the full State of Martech 2026 report
Download the 2026 Marketing Technology Landscape

With those in your hands, let me guide you through some highlights.

In short, the state of martech is strange, stressful, and — somehow — strong.

For the first time in 15 years, the growth of the martech landscape effectively flattened. The number of commercial martech products grew from 15,384 last year to just 15,505 this year — a net increase of only 121 products, or 0.79%.

After a decade and a half of relentless expansion, from 150 products in 2011 to more than 15,000 today, have we finally hit peak martech?

Maybe.

But “flat” is also a misleading word to describe what’s happening. Because underneath that near-zero net growth, the market is churning furiously:

1,488 products were added.
1,367 products were removed.

At long last, the removal rate has nearly caught up with the addition rate. The inflow of new products dropped 40% from last year, from 2,489 additions to 1,488. Meanwhile, the outflow of existing products climbed 13%, from 1,211 removals to 1,367.

So who dropped out in this round of martech musical chairs?

Perhaps not surprisingly, the exits are concentrated among smaller SaaS companies, especially those from the 2010–2019 startup wave. That cohort accounts for 51.7% of this year’s removals. Nearly 80% of removed products had 50 employees or fewer. By revenue, the $1M–$10M band accounted for 45.5% of removals.

Yet these weren’t all tiny experiments that never found a customer. Many were companies that found some traction — but not enough durability to survive the next phase of the market.

But there's a lot of variance across different categories of martech.

Take Content Marketing. (No, please…)

When generative AI went mainstream, Content Marketing was one of the first categories to explode. AI writing tools, content repurposing tools, social content generators, blog assistants, creative production platforms — suddenly everyone had a “content AI” product.

From 2023 to 2025, the category nearly doubled. But in 2026, Content Marketing had the largest outflow with 176 products removed. What happened?

Three forces converged:

First, the major AI labs absorbed a lot of the functionality. “Generate a blog post,” “write ad copy,” “turn this into a LinkedIn post” — these are now table stakes in ChatGPT, Claude, Gemini, and the major productivity suites.

Second, incumbent SaaS platforms moved fast. Adobe, HubSpot, Salesforce, and many others embedded AI content capabilities directly into the workflows customers already use.

Third, product-market fit proved harder than expected. Generating content fast and generating content that actually works are very different things.

The second most-contracted category? Sales Automation, Enablement & Intelligence — which not coincidentally was the second fastest-growing category in the gen AI boom.

Yet still, it’s worth noting that each of these categories added new products too.

Meanwhile, other categories are legitimately in a “Growth” phase, with high inflow to relatively low outflow:

Notice the through-line: every one of these is benefiting from the disruption that’s hitting other parts of the stack. They’re growing not in spite of AI, but because of AI and how it’s rewiring everything around them.

This brings us to one of the more important — and still underappreciated — shifts happening underneath the surface: martech platforms are starting to metamorphose from apps humans operate into infrastructure agents can use.

For most of the past 20 years, martech was organized around applications. Marketers logged into platforms, clicked through dashboards, configured workflows, launched campaigns, and monitored reports.

The UI was the product.

But in an agentic world, a growing share of martech’s value moves below the UI. The question becomes less, “Can a marketer use this app?” and more, “Can an AI agent safely access the right data, content, workflow, permission, or action at the right moment?”

The momentum of MCP is one of the clearest signals of this shift.

There are now over 29,000 MCP servers listed in the different registries out there, in just 18 months. (That escalated fast!) Many major martech platforms have released or announced their own. In ChatGPT’s and Claude’s “app store” equivalents, you can find official integrations with Adobe, Airtable, Amplitude, Asana, Braze, Canva, Clay, Figma, HubSpot, Intercom, Klaviyo, Mailchimp, Notion, Semrush, Slack, Snowflake, Tableau, Webflow, WordPress, Workato, Zapier, ZoomInfo, and many more.

In the last era, martech platforms competed to be where marketers worked. In the next era, they’ll compete to be what agents can work with.

The cover of this year’s report features a chrysalis — the structure a caterpillar builds around itself before becoming something else entirely. Inside a chrysalis, the caterpillar dissolves. Its old structure breaks down. From that strange biological soup, something entirely different assembles itself.

Same DNA. Unrecognizable form.

That feels much closer to what’s happening in marketing today. A structural reorganization of the technology, the practice, the roles, and the relationship between brands and customers.

We frame this transformation across three stages:

Caterpillar: marketing as we’ve known it.
Chrysalis: where most of us are today.
Butterfly: where we believe this is heading.

In the market, we’re moving from marketer-controlled “journeys” to customer-controlled conversations inside ChatGPT, Claude, Gemini, Perplexity, and other AI interfaces — headed toward a world where AI agents increasingly operate on the customer’s behalf.

In martech, we’re moving from deterministic SaaS platforms and rules-based workflows to a messy rebundling struggle of incumbent platforms, AI-native products, and homegrown agents — headed toward context-as-a-service infrastructure that orchestrates a potentially unlimited array of dynamic AI apps and agents.

In marketing roles, we’re moving from campaign manager to agent operator to value engineer. In marketing ops roles, we’re moving from system admin to stack wrangler to context engineer.

It’s messy. But the caterpillar doesn’t skip — can’t skip — the chrysalis. Neither can we.

The report also includes one of the nerdier-but-useful pieces of analysis we’ve done. 70 AI use cases in marketing, mapped across the six major categories of the martech landscape. For each use case, we asked whether companies were using AI inside existing SaaS platforms, new AI-native tools, custom-built solutions, or none of the above.

The overarching pattern that emerged was not build-vs-buy, SaaS-vs-AI. It’s build and buy and borrow-from-the-SaaS-platform-you-already-have.

The Data category is a great example.

Across 13 use cases, AI-native tools are doing well in areas like coding and competitive intelligence, places where fast iteration and model quality matter. SaaS and service providers are holding ground in audience and identity resolution, places where trust, pipes, permissions, and operational history matter.

Then there’s the awkward part: governance.

Data lineage, compliance, privacy, and consent are not moving as fast as the more glamorous use cases. Which is… not ideal, given that every AI initiative eventually backs into the same uncomfortable question: do we actually trust the data, context, and permissions this thing is acting on?

That’s the tension running through all 70 use cases.

Marketers are not waiting for the perfect architecture. They’re experimenting everywhere — inside their suites, with AI-native tools, and through homegrown agents and apps. But the more AI spreads, the more the bottleneck shifts from “can we do this?” to “can we make this reliable, governed, and context-aware?”

If there’s one word that keeps surfacing across all of this, it’s context.

Context is the difference between AI that generates plausible output and AI that creates meaningful value.

It’s the difference between “send an email” and “send the right email, to this customer, at this moment, with awareness of what they’ve already done, what they’re trying to accomplish, what we promised them, what we’re allowed to say, what inventory we actually have, what sales already discussed, what support already resolved, and what the brand should sound like while doing it.”

In the report, we describe three kinds of context that need to come together:

The magic is in the overlaps.

The overlap between company and customer is value engineering. Value engineering asks: where can we create meaningful value for the customer and the company?

This is not really a technology question. It’s strategy, empathy, economics, taste, and judgment. It’s understanding what moments matter, what outcomes are worth improving, and what kind of experience would actually be better — not merely more automated.

The overlap that includes systems is context engineering. This is the technical and operational work of making the right data, content, permissions, instructions, models, workflows, and guardrails available to the right agent at the right time.

Where all three converge — your goals, your customer’s needs, and your systems’ ability to deliver in real time — is what we’ve dubbed Golden Context.

Value engineering identifies the value. Context engineering makes it actionable. Together, they become the new center of gravity for marketing operations.

It’s important to recognize that context isn’t one thing. It operates in pace layers.

Some context changes by the second: the customer’s current query, real-time state, the conversation happening right now.

Some changes over minutes or hours: the session, the actions they’ve taken, the content they’ve consumed.

Other context unfolds over days, weeks, months, or years: journey stage, relationship history, preferences, lifetime value, brand, strategy, governance, and the broader market environment.

AI accelerates the oscillation across all of these layers. It lets us respond faster in the moment — but only if the slower layers are stable, accessible, and aligned underneath.

A real-time agent that knows the current query but not the customer relationship, the company’s promises, or the governance boundaries isn’t truly intelligent. It’s just fast. Possibly making customers furious.

Golden Context is when the fast layers and slow layers move together. Immediate enough to be relevant, grounded enough to be trusted.

For marketing ops, martech, RevOps, GTM engineering, and tech-savvy marketing leaders, this is an extraordinary moment. The app-centric era of martech made marketers operators of tools. The agentic era will make them designers of context.

That, I think, is the butterfly starting to form.

I’ve only covered a slice of what’s in the State of Martech 2026. You can download the full report and the clickable landscape PDF, all ungated. Feel free to share with colleagues, in articles, or on social media. We’d love to hear your thoughts. Tag us on LinkedIn, and we’ll join the conversation.

A huge thank you to our sponsors — GrowthLoop, Hightouch, Knak, MoEngage, Pega, Progress, and SAS — whose support made it possible for us to produce this research and share it freely with the community.

Here’s to what’s taking shape,

Scott

P.S. If you missed our live #MartechDay keynote, the on-demand replay is available for the rest of this month. Catch it here.