Best of LinkedIn: MarTech Insights CW 18/ 19

00:00:00: This episode is provided by Thomas Allgaier and Frennis, based on the most relevant LinkedIn posts about MarTech from Calendar Weeks Eighteen and Nineteen.

00:00:07: Frennis a B-to-B market research company that supports enterprise marketing teams in unlocking full potential of their customer data with the help of AI.

00:00:15: You can find more info in description.

00:00:18: So welcome to The Deep Dive!

00:00:20: Today we are unpacking top MarTech trends everyone's talking about across LinkedIn right now.

00:00:25: Yeah it has been really wild couple of weeks for B to be marketing professionals, I mean we are seeing this massive shift away from just you know buying endless software and moving straight into the era of

00:00:37: AI execution.

00:00:38: Exactly!

00:00:39: We're going cover everything from shocking new numbers on the Martek landscape to why buying more tools is completely failing...the real world deployment of a genetic AI....and finally these massive security risks hiding in all the new AI integrations.

00:00:52: it's a lot

00:00:54: is a lot, but it's crucial because if you open your inbox today.

00:00:57: You're just flooded with vendors promising that their new AI features going to like magically fix your pipeline.

00:01:04: But the conversations happening among actual operators tell a totally different story

00:01:08: Right?

00:01:08: so let's ground this in The Numbers right off the bat Because the New State of Mar-Tec twenty twenty six report which Scott Brinker and Frans Riemersmo have been heavily discussing It has This headline number.

00:01:18: That feels honestly kind of counterintuitive.

00:01:20: Yeah the fifteen thousand Number.

00:01:22: Fifteen thousand five hundred and five individual products currently on the market.

00:01:27: But, The growth rate of landscape has totally flatlined.

00:01:30: we're talking about a microscopic point seven nine percent year-over-year growth.

00:01:34: So I mean if the market is basically perfectly still lake right now does that means Martek?

00:01:38: Is dead or are We just addicted to buying the same shiny objects in different packaging?

00:01:43: Well, it's definitely not dead.

00:01:44: On the surface yeah a point seven nine percent growth rate makes it look like this still stagnant lake.

00:01:49: you'd think innovation just stopped.

00:01:52: but um Edith of Barmintloo and Justin Thaibo pointed out that underneath the flat percentage, there is this incredibly violent churn happening.

00:01:59: Violent

00:01:59: churn?

00:02:00: Like.

00:02:00: what does it look like in data?

00:02:01: It means over last year one thousand four hundred eighty eight tools were added to landscape but at exact same time one thousand three hundred sixty seven tools are completely removed or absorbed.

00:02:13: Oh wow!

00:02:14: So the lake isn't still all these vicious undertow dragging platforms down while new ones pop up.

00:02:20: What exactly getting dragged under?

00:02:22: Mostly content marketing tools.

00:02:24: That first generative AI bubble from a few years ago, it burst completely!

00:02:28: The standalone tools built just to write text became obsolete.

00:02:30: overnight the big AI labs absorbed those features and massive enterprise platforms embedded text generation right into their workflows.

00:02:38: Right because why buy separate tool when your CRM does that for you now?

00:02:42: Exactly But at same time we saw massive growth in other areas like CMSs and e-commerce tools grew by roughly twenty percent

00:02:49: Which is fascinating.

00:02:50: because a content management system, like one of the oldest categories in digital marketing.

00:02:56: Why this sudden twenty percent surge there?

00:02:58: Because the whole architecture is shifting.

00:03:01: SAS has becoming the infrastructure layer.

00:03:03: you know...the plumbing and AI stepping-in as an actual value layer.

00:03:08: Companies are realizing their websites have the wrong architecture entirely.

00:03:12: They're rebuilding from the ground up so that machines and AI agents can read them, prioritizing structured data over just pretty visual

00:03:20: interfaces.".

00:03:20: I love it!

00:03:21: It actually makes me think of this incredible analogy that Dia Costa shared on LinkedIn building.

00:03:26: some thoughts for Stephanie Jenin.

00:03:28: Scott Brinker originally called a chrysalis phase.

00:03:30: like the old software caterpillar dissolves into soup becomes beautiful butterfly

00:03:34: Right...the butterfly metaphor.

00:03:36: Yeah, but deep pushback on that.

00:03:37: He said Martek is in the butterfly.

00:03:39: it's a sci-fi xenomorph.

00:03:40: The new AI driven architecture is literally just bursting violently out of the chest Of these old legacy systems.

00:03:48: That I mean...that Is A perfectly terrifying and accurate way to describe what It feels like To manage a marketing stack right now?

00:03:56: IT IS NOT A GRACEFUL EVOLUTION!

00:03:58: No

00:03:58: its messy Which brings up a glaring contradiction.

00:04:02: If the market is turning this violently, why are companies still buying software the old way?

00:04:08: Nick Friedman Tavaakhorst had this brutally honest post about companies routinely wasting like a hundred grand on tools they already own.

00:04:16: Oh, the attribution platform story?

00:04:18: Yes!

00:04:19: He shared his example of an enterprise client about to sign a six-figure contract for new attribution tool.

00:04:24: Nick's team comes in maps architecture and finds that company already owned three different tools which could do exactly same thing.

00:04:31: They just weren't wired together correctly

00:04:33: And That is The Norm.

00:04:34: It's not an outlier at all.

00:04:35: Premadam actually highlighted this stat showing that fifty percent of enterprise sauce licenses are totally unused, half the multi-million dollar stack is basically doing nothing.

00:04:44: That

00:04:44: is insane!

00:04:46: So when we talk about Stack Consolidation in twenty-twenty six... We're not talking about a retreat it as sign of operational maturity.

00:04:53: Sergi Scripnik has this amazing test for what he calls new RevOps.

00:04:58: Oh I saw this The Consultant Test.

00:05:00: He asks, if your external consultant or systems integrator disappeared tomorrow would your system still produce revenue?

00:05:07: If the answer is no.

00:05:08: you don't have a revenue engine.

00:05:09: You have a

00:05:10: dependency.".

00:05:11: You build something so fragile that your own team can't even drive it.

00:05:14: but okay let me play devil's advocate for a second.

00:05:17: How does an enterprise accidentally buy three of the same tool?

00:05:21: like finance approves budgets IT runs security checks.

00:05:25: Where Is The Disconnect Happening?

00:05:26: It?

00:05:27: siloed procurement Different regional teams buy things in isolation or a new VP comes in and just buys whatever tool they used at their last job.

00:05:34: Right, And then the CMO gets blamed when things don't work.

00:05:38: Anna Maron brought up a brilliant point about delegated authority here.

00:05:42: CMOs are held accountable for tech choices made by IT or finance.

00:05:48: it's like being a Michelin star chef but The Finance Department buys all your ingredients from a discount bin.

00:05:56: That's

00:05:56: a great way to put it.

00:05:57: How can marketing be held accountable when they don't even hold the keys?

00:06:00: They can't, and that is why marketing leaders have to take back architectural control because once The next step isn't buying more software, it's adding a digital workforce.

00:06:10: Yes let's get into agentic AI because this is moving so fast.

00:06:14: Mohammed binish pointed out that AI is no longer just a tool for writing blogs.

00:06:19: It has full marketing work force.

00:06:20: Absolutely Martin Keen argues the future stack needs distinct third layer.

00:06:26: You have data ops you have your martech software and now you've got the agentec layers sitting on top.

00:06:30: I want to make sure we explain this mechanically for the listener, because traditional Martek is like a power tool.

00:06:35: It makes the carpenter faster but you still have to hold it and pull the trigger.

00:06:40: Agenic AI is hiring an independent contractor who brings their own tools just builds your house while you sleep.

00:06:46: Exactly!

00:06:47: Agents are goal-oriented.

00:06:48: You give them an objective And they dynamically navigate systems overcome errors execute.

00:06:54: Look at what Katie UN shared about Databricks, they deployed three AI agents in production for their marketers.

00:07:00: even named them right Marge Tagatha and Atlas.

00:07:02: Right?

00:07:03: Marge helps marketers query complex data sets instantly tag of the structures and tags years of historical content.

00:07:11: Atlas handles dynamic audience segmentation.

00:07:14: They serve three hundred marketers with trusted data

00:07:16: access.

00:07:17: Aqua Burraman at mail moto shared another one that blew my mind frog, Google Sheets and Slack to just fully automate on-page SEO across hundreds of pages.

00:07:29: The agent reviews the crawl data finds broken links proposes fixes in a sheet and slacks.

00:07:39: Henry soon proved that you can take a traditional content marketing function, which costs about three hundred and sixty six thousand dollars.

00:07:46: A year to run an replace it with one hybrid human operator equipped with about twelve thousand dollars of AI tools.

00:07:53: wait Three hundred and Sixty-six K down the what?

00:07:56: Down to one hundred sixty two thousand dollars.

00:07:58: It drastically cuts the cost.

00:08:00: okay But if we are turning marketing into this automated assembly line run by agents like Margin Atlas Where does human strategy fit in?

00:08:08: Like, are we just automating ourselves out of a job.

00:08:10: That's the fear but actually Human creativity becomes way more valuable.

00:08:14: The bottleneck isn't content creation anymore.

00:08:16: We solved that!

00:08:17: The bottlenecks is orchestration and context.

00:08:20: Meaning agents can build house in dark But they still need us to tell them where their property lines Are.

00:08:24: They Need Context

00:08:25: Exactly And To get that context They have connect directly with our internal data

00:08:30: Which brings this to most alarming part of deep dive.

00:08:33: Jena Mystery has been talking about MCP model context protocol.

00:08:37: It's the new standard for how AI connects to marketing data.

00:08:41: Yeah, MCP completely changes the game.

00:08:43: it closes the gap between insight and action.

00:08:45: instead of rigid APIs MCP acts like a universal skeleton key so the AI can just read your files And pull what it needs.

00:08:53: but the adoption speed is wild.

00:08:55: Diacosta noted that it took commercial Martek fifteen years To hit fifteen thousand products.

00:09:00: MCP hit twenty-nine thousand servers in just eighteen months.

00:09:04: And that is exactly why we have a massive security risk right now.

00:09:09: Clark Barron ran an independent security audit on these servers, and the warning he dropped was heavy!

00:09:14: There are twenty nine thousand MCPs out there with practically zero

00:09:19: verification.

00:09:20: What did he actually find in the audit?

00:09:21: He

00:09:21: found HubSpot MCP servers secretly collecting IP addresses, silently migrating access tokens and caching live CRM data as plaintext JSON on local file system.

00:09:30: Wait, plain text JSON so zero encryption.

00:09:32: just sitting there on a local drive...

00:09:34: It's sitting there unencrypted!

00:09:35: So if our internal data is a polluted river, MCP is essentially high-speed pipe pumping that toxic water straight into AI agent brain.

00:09:43: Yep

00:09:43: And this leads to my favorite quote of year from Komalzi.

00:09:46: She said oops your stack is showing.

00:09:48: Huh I love that because AI agents don't have human intuition to smooth over your bad data.

00:09:54: If you're internal architecture is broken, the AI exposes that directly to the customer.

00:09:59: The inner game is now the outer game.

00:10:01: So as a marketer do i need to freeze everything?

00:10:04: Do we need to spend eighteen months building the perfect Customer three sixty dashboard before We let ai touch our data?

00:10:12: no definitely not.

00:10:13: yogido wadwa advises strongly against That.

00:10:16: Skip the multi-million dollar customer three sixty obsession.

00:10:20: Okay, so what do we do instead?

00:10:21: You

00:10:21: start with a next decision.

00:10:23: Let's say you want to reengage a thirty day non purchaser.

00:10:26: figure out The absolute minimum data needed for that specific action.

00:10:30: clean just That tiny slice of data and build from there.

00:10:33: incremental execution

00:10:34: starts small.

00:10:35: I love that.

00:10:36: Well, we've covered a ton of ground today.

00:10:37: if you enjoyed this episode new episodes drop every two weeks.

00:10:40: Also check out our other editions on account-based marketing field marketing channel Marketing AI and B to be marketing.

00:10:46: go to market and social selling.

00:10:48: Yeah Thank You so much for joining us on the steep dive.

00:10:50: make sure to subscribe

00:10:51: And we want leave it with one final provocative thought.

00:10:54: Rupert Steffner and Stephanie Jane in both pointed out That as search evolves into answer engine optimization AI agents are going to start making buying recommendations in the background.

00:11:05: Right,

00:11:06: before a human ever even sees your site?

00:11:08: Exactly!

00:11:09: So you're website's primary audience is no longer human it's an AI crawler.

00:11:13: so ask yourself how were you designing your digital presence?