Best of LinkedIn: MarTech Insights CW 06/ 07

00:00:00: This episode is provided by Thomas Algeyer and Frennis, based on the most relevant LinkedIn posts about MarTech from calendar weeks six-and-seven.

00:00:08: Frennes is a B to B market research company helping enterprises gain the market customer and competitive insights needed.

00:00:18: Yeah.

00:00:19: Today, we are unpacking the marketing technology landscape based on what we saw in calendar week six and seven of twenty-twenty-six.

00:00:26: And I have to say looking through the curation for this week The vibe has definitely shifted.

00:00:31: it really had you know normally when we open these updates It's very I don't know, feature focus.

00:00:36: It's like look at this shiny new generative AI tool.

00:00:40: or here is a slightly better way to sequence your outbound emails?

00:00:44: Right

00:00:44: very tactical

00:00:45: exactly but this week the conversation on LinkedIn just felt heavier.

00:00:48: it felt Like The Industry Is Having A Bit Of An Existential Crisis.

00:00:52: That is a really great Way To Put It.

00:00:54: We Aren'T Just Seeing People Ask What Software They Should Buy Anymore we were seeing marketing leaders asking Why Did We Build Our Infrastructure This Way?

00:01:00: In The First Place?

00:01:01: It's a huge move away from just shopping for tools and toward the fundamental rethinking of the whole architecture.

00:01:08: Right,

00:01:08: it is difference between deciding how to decorate house in actually checking foundation.

00:01:13: make sure that house isn't sinking into mud.

00:01:16: Exactly

00:01:16: So.

00:01:17: we have three big themes today.

00:01:19: really makes sense.

00:01:20: shift.

00:01:21: First look at rise vibe coding and how ops teams are replacing expensive enterprise software with tool they build themselves.

00:01:29: Then tackle great stack.

00:01:30: clean up why companies are ripping out these massive black box platforms in favor of their own data warehouses, really trying to cut through the tech clutter.

00:01:40: And finally we will discuss the human element because I mean you can have the best AI architecture in the world but if your sales and marketing teams are at war or If You've Hired The Wrong Type Of Rebops Leader

00:01:53: none of that technology matters.

00:01:54: nothing matters.

00:01:55: it's gonna be a dense one But i think is exactly what We need To dive into right now

00:02:00: Absolutely.

00:02:01: So let's start with the disruption part, there was one story that absolutely dominated The Feed this week.

00:02:07: it is a case study shared by Barry Latimer and frankly its kind of story probably keeps SAWS CEOs awake

00:02:13: at night.

00:02:14: Oh!

00:02:14: The four hundred sixty thousand dollar story

00:02:17: That's the one walk us through.

00:02:19: So Barry shared a breakdown of a project led by his team lead, Jennifer Bright.

00:02:23: And the context is pretty standard for large enterprise.

00:02:26: their client was paying roughly four hundred and sixty five thousand dollars a year For legacy sauce platform.

00:02:32: Wow Yeah!

00:02:33: This platform's main job was campaign execution so handling audience logic splitting volumes for A-B testing segmentation to standard marketing ops stuff.

00:02:44: Let's just pause on that number for a second.

00:02:45: Nearly half-a-million dollars a year, That is a massive line item.

00:02:49: and let be honest For campaign execution software A lot of the classes often just legacy bloat

00:02:54: Exactly!

00:02:55: And this exactly what Janessa thought.

00:02:57: She looked at price tag Looked at capability and effectively said I can build something better than This

00:03:01: Which is wild because usually when a marketer says I can build it myself, everyone in IT just rolls their eyes.

00:03:07: Right?

00:03:08: It usually ends with some massive spreadsheet that crashes the server or a shadow IT project that gets shut down and three months.

00:03:14: Well The classic Build versus Buy Debate Usually Ends With Buy.

00:03:17: right Because building's too hard Is Too Expensive To Maintain And You Aren't A Software Company But This Is Twenty-Twenty Six.

00:03:26: Jennifer Isn'T A Software Engineer By Trade.

00:03:29: She'S A Campaign Operations Leader But she used claw the AI model to architect and build a custom solution from scratch.

00:03:36: She didn't just write a script.

00:03:38: She built a tool that runs natively in The client's snowflake environment,

00:03:42: okay?

00:03:43: I want to clarify that for you if your listening And maybe aren't deep in the data weeds.

00:03:46: when we say it runs native Lee in Snowflake What does that actually mean for the business?

00:03:51: compared to the old way

00:03:53: It's a massive difference.

00:03:54: Yeah In the old model of the sauce model You usually have to copy your data out of your warehouse

00:03:59: right

00:03:59: send it to the vendor's cloud, let them process and then maybe sync back later.

00:04:03: Which is a nightmare!

00:04:05: It costs money...it creates huge security risks…and its slow.

00:04:09: Jennifer's tool sits inside the warehouse.

00:04:12: The data never leaves.

00:04:14: It handles metadata..does audience splitting And this is kicker.

00:04:19: Generates production ready SQL with single click.

00:04:22: So she effectively built her own user interface that translates.

00:04:26: you know I want to target people in New York who bought X into the complex database code that actually finds those people.

00:04:33: Yes,

00:04:33: and a result they saved the client four hundred sixty thousand dollars annually.

00:04:38: Unbelievable!

00:04:39: They cancelled the SAWS contract entirely And according to Barry The new solution is faster and aligns way better with clients specific workflow because it was custom built for them.

00:04:49: This feels like major tipping point.

00:04:51: We've been hearing about the democratization of code for years, but this is a high-stakes real world example.

00:04:57: If an ops leader can use an LM to build replacement for our half million dollar platform in few weeks... The

00:05:03: suddenly billed becomes cheaper faster option!

00:05:05: Exactly and This connects directly to concept John Miller was discussing this week.

00:05:09: he calls it vibe coding.

00:05:11: Yeah I saw that term floating around.

00:05:12: It sounds like something Gen Z would say on tiktok But definition actually really technical.

00:05:17: So, Vibe Coding essentially means writing code through natural language prompts.

00:05:22: You describe the vibe or function of what you want and AI handles syntax structure.

00:05:30: John's argument is that the Martek landscape is splitting.

00:05:33: on one side we have workflow layer user interfaces click around in.

00:05:38: he thinks vibe coding will eat entire market.

00:05:41: Because why pay Salesforce or HubSpot for a generic interface when I can just vibe code?

00:05:46: A form that looks exactly how my team naturally works.

00:05:49: Exactly, domain experts the people who actually know the problem Can now just build this solution.

00:05:54: However Miller adds a crucial nuance here and i think This is important.

00:05:58: so we don't Just assume all saws as dead tomorrow.

00:06:01: right

00:06:01: he argues That enterprises will still pay For what He calls systems of control.

00:06:06: okay let's unpack that.

00:06:07: if We are vibe coding our own apps What Is left to buy?

00:06:10: Safety.

00:06:11: Governance.

00:06:12: Trust.

00:06:14: You can vibe code a user interface, but you cannot vibe code compliance with GDPR or privacy laws.

00:06:21: You can't vibe code email deliverability reputation.

00:06:24: That makes total sense like I might trust my custom app to sort My contacts.

00:06:29: But i don't want To be personally responsible for ensuring Im not breaking international Privacy Laws!

00:06:34: I will gladly pay A vendor.

00:06:35: take that liability

00:06:36: Precisely.

00:06:37: You build the fun stuff, the workflows that creative tools.

00:06:40: The integration glue and you buy insurance?

00:06:43: Identity resolution?

00:06:44: Consent management?

00:06:45: Heavy infrastructure

00:06:46: And we are seeing this build mentality trickle down to smaller tasks too.

00:06:50: It's not just replacing massive enterprise platforms.

00:06:53: All Accenture Leasig shared some really practical examples of using low-code tools like N-Eight Me To do exactly this.

00:06:59: Yes!

00:07:00: Leasigg is a great example what I would call modern marketing engineer.

00:07:03: He isn't building replacement for Salesforce But he is building the automation glue that holds his day-to-day together.

00:07:09: Like for example, he automated The Creative Handoff process.

00:07:12: What does it look like in practice?

00:07:13: Well instead of downloading images from Trello and manually uploading them to MetaAd's manager which was just mind numbing work

00:07:20: Oh...the worst

00:07:21: He built a script That doesn't automatically

00:07:23: Which sounds small but if you have ever been the person dragging And dropping JPEGs For three hours on Friday afternoon You know that as a lifesaver!

00:07:31: ...and

00:07:32: It removes human error entirely.

00:07:35: He also built an automated alerting system for Google Ads budgets.

00:07:39: If a campaign is overspending, or conversely if it's performing incredibly well the system alerts the team in slack and can even autoscale budget by thirty percent.

00:07:50: That moving from reporting to acting.

00:07:52: Usually you get report on Monday saying oops we overspent Saturday.

00:07:55: this catches at real time.

00:07:57: Even as set up for SEO content One AI model writes a draft, and second separate AI model acts as the editor reviewing it against style guide before human ever even looks at.

00:08:08: It's like having digital interns checking work of your digital writer!

00:08:12: This all points to future where we aren't just logging into software for working anymore... We are orchestrating agents that do this for us which brings perfectly Matthew Sweezy post about the agentic era.

00:08:28: His argument is that for twenty years we've designed customer journeys.

00:08:31: We built websites and landing pages assuming a human being with human eyes, and mouse would navigate them.

00:08:37: And now?

00:08:38: Now it might not be the humans visiting your site at all It's their AI agent.

00:08:43: If I tell my AI hey find me a CRM software that costs under fifty bucks a month and integrates with X That AI will go out to scan the web.

00:08:51: So if your website is full of fluffy marketing speak and giant pop-ups, the AI might just skip

00:08:57: it.

00:08:57: Worse you may not be able to read at all.

00:08:59: Sweezy says brands need become callable infrastructure.

00:09:03: You need a resource that an AI can ping get.

00:09:05: pricing availability and specs via API.

00:09:08: If brand isn't callable effectively disappear from consideration set.

00:09:12: That is a terrifying thought for brand marketers.

00:09:15: We've spent so much time optimizing for eyeballs, and now we have to optimize for algorithms!

00:09:20: It

00:09:20: all comes back to data structure.

00:09:22: whether you are Jennifer Bright building a tool on Snowflake or Matthew Sweezy preparing for AI agents the conclusion is exactly the same The pretty user interface matters less...the underlying data architecture matters more

00:09:36: which is the perfect segue to our second theme today.

00:09:39: Because if data architecture is new gold, we have a lot of cleaning up-to do because right now most stacks are a complete

00:09:47: mess.".

00:09:54: There's this tendency to think, oh our conversion rates are down.

00:10:10: Let's buy an optimization tool.

00:10:11: Our leads are weak.

00:10:12: let's buy a data enrichment tool.

00:10:14: It is retail therapy for CMOs

00:10:16: Exactly.

00:10:17: But Bujaj argues that digital success Is about connected systems.

00:10:22: If your system aren't talking with each other if the Data Enrichment Tool isn't syncing cleanly With CRM You are just generating noise, you aren't generating revenue.

00:10:31: We saw a radical example of someone decluttering this week too.

00:10:34: Lars Petterblokam shared the story about ripping out at customer data platform a CDP.

00:10:39: This was such bold move.

00:10:41: For context A CDP is supposed to be brain your marketing.

00:10:45: It collects data from everywhere, web email sales and creates a unified profile of your customer.

00:10:51: In theory in practice they can be incredibly complex and very expensive to implement

00:10:56: right?

00:10:56: And Lars called their CDP black box.

00:10:59: They were feeding data into it But they didn't fully control the logic of what happened inside and getting the data back out was a massive pain.

00:11:06: So they just replaced it.

00:11:07: They moved to a fully owned architecture on BigQuery And

00:11:09: the timeline was shocking Two days.

00:11:12: I have seen CDP implementations take two entire years.

00:11:15: It

00:11:16: really speaks to power of modern cloud data warehouse.

00:11:19: By moving into big query, they eliminated their technical debt.

00:11:23: They aren't reliant upon vendors proprietary roadmap anymore.

00:11:27: If want change how they calculate customer lifetime value Just changed SQL query.

00:11:33: They don't have to file a support ticket and wait three weeks.

00:11:36: This aligns so well with the small core model that Jeff Smith was discussing?

00:11:39: Yes,

00:11:40: Smith's view is we need stop buying tool for every single edge case.

00:11:45: He suggests a small core of powerful orchestrated platforms like your CRM in your data warehouse And then using those vibe coded apps or specialist agents To handle specific edge-case problems.

00:11:56: So instead of a stack of fifty disconnected tools, you have two or three massive pillars and then a bunch of lightweight scripts that just do specific jobs.

00:12:04: Exactly!

00:12:05: And this shift is forcing the CDP vendors to actually change their approach.

00:12:10: Shashi Bellumkanda had great analysis on it – he noted that Warehouse Native is becoming mainstream for CDPs now.

00:12:16: Meaning… Instead of the CDPs saying, send me all your data … The CDP says I'll come where your data already lives.

00:12:22: Correct The trend is moving away from storing a copy of your data in a vendor's cloud.

00:12:28: The data stays on your snowflake or BigQuery, the CDP just sits on top to provide identity resolution and activation layer.

00:12:37: Shashi also touched on AI in this context didn't he?

00:12:40: He did!

00:12:40: And it was harsh reality check.

00:12:42: He said that simply claiming you're platform as AI powered no longer differentiated.

00:12:47: It has table stakes.

00:12:48: Everyone has AI.

00:12:49: The real differentiator in twenty-twenty six is governance and trusted customer context.

00:12:55: Can I trust the data?

00:12:55: Is it compliant?

00:12:57: that is what people will actually pay for.

00:12:58: It's funny.

00:12:59: We keep coming back to trust and governance as the only things left That are worth paying a premium for.

00:13:04: everything else is just becoming a commodity.

00:13:06: Speaking of paying we should definitely mention Moritz Nicholas Wolfe here.

00:13:10: He's pushing who he calls this zero dollar migration movement.

00:13:13: This

00:13:13: is a push for transparent pricing, right?

00:13:15: It's a reaction to the hidden costs that drive everyone crazy.

00:13:19: User caps, data storage fees implementation costs.

00:13:23: Moritz is arguing.

00:13:25: these hidden fees are exactly what drives people build their own stacks in first place.

00:13:30: If vendors want survive they need remove friction of switching

00:13:34: it all connected.

00:13:34: if you make too hard or expensive use your software.

00:13:38: Jennifer Bright is just going to build a replacement with Claude.

00:13:40: Precisely, but here's the thing.

00:13:42: you can have the leanest most beautifully vibe coded stack in the world running on a pristine data warehouse and it will still fail if your team is dysfunctional.

00:13:51: which brings us To our final theme today organization and strategy because You cannot fix A people problem With a software patch.

00:13:59: This Is The hardest part.

00:14:01: Natalie Furness broke this down beautifully regarding RevOps revenue operations.

00:14:06: RevOps is one of those buzzwords that everyone uses, but I feel like every company defines it differently.

00:14:11: That's exactly the problem.

00:14:13: Natalie argues that companies fail because they don't understand the ladder of Revov's roles.

00:14:17: They think Revovs is just one homogenous thing, but she outlines very clear levels level-one as an analyst this someone who builds reports and configures fields on a specific platform.

00:14:29: so the tactical fixer right?

00:14:31: But Level three is systems manager or architect.

00:14:34: This is someone who translates business requirements into a system design across platforms.

00:14:40: And level five as a director, Someone with strategic board-level alignment Who talks about revenue not API keys.

00:14:47: and I'm guessing companies are hiring Level ones and expecting level five results

00:14:50: or the exact reverse.

00:14:52: they hire A high-level Director and then get incredibly frustrated when that person isn't in HubSpot fixing a broken form field.

00:14:58: You have to match The talent to the maturity of your organization.

00:15:01: Speaking of broken relationships, we have to talk about the eternal war between marketing and sales.

00:15:06: Drew Nyser posted about The Blame Game.

00:15:08: Oh!

00:15:08: We all know this game.

00:15:09: Revenue misses the

00:15:10: target."The

00:15:11: CMO says,"Hey I sent you ten thousand leads.

00:15:14: Sales just can't

00:15:15: close.".

00:15:16: And Sale said those leads were garbage, I need Glengarry Leads.

00:15:19: It's a tale as old is time.

00:15:21: so how do actually break that cycle?

00:15:27: Literally, stop talking about how many leads you got.

00:15:29: No one cares if he generated ten thousand leads and zero dollars closed.

00:15:33: instead start diagnosing conversion.

00:15:36: He suggests a revenue reset sprint.

00:15:39: What does that actually look like practically?

00:15:40: It's a weekly standing meeting with sales leadership.

00:15:43: But the agenda isn't.

00:15:45: here are my numbers.

00:15:46: The agenda is let's look at why these specific deals didn't close.

00:15:51: You co-owned the failure you work together to fix the messaging in the targeting.

00:15:55: I love that phrase, stop defending start diagnosing.

00:15:58: It changes the dynamic from adversaries to allies.

00:16:01: Woder Delement had a very specific tactical way of helping with this in HubSpot didn't he?

00:16:05: Yes and it's something you can go implement tomorrow.

00:16:07: Wodder says Stop creating deals as soon someone fills out form.

00:16:11: But

00:16:11: wait.

00:16:11: marketing wants show massive pipeline for CEO.

00:16:15: That is the trap.

00:16:16: If created deal immediately your forecasting accuracy just tanks because most people aren't ready buy yet.

00:16:23: Your pipeline looks huge But your close rate looks terrible.

00:16:26: So what's the alternative?

00:16:28: Wouter suggests using a lead pipeline or status field first, you move them from new to connected to qualified.

00:16:36: You only create formal deal record when product fit is confirmed and dollar amount is actually clear.

00:16:42: That keeps sales forecast completely clean.

00:16:45: It seems like small database change but it has huge impact on trust.

00:16:49: If the pipeline contains real deals sales will pay attention.

00:16:53: Trust is the key word there.

00:16:55: Joel Harrison and Robert Nicholson had a fascinating discussion about integration failures.

00:17:00: They pointed out that when integrations break, it is rarely because of the API failed.

00:17:05: The technology usually works fine.

00:17:07: So why do

00:17:08: they fail?

00:17:09: Because of lack of trust.

00:17:11: If sales doesn't trust data marketing is sending over.

00:17:14: if they think its inaccurate or duplicated then stop using this system.

00:17:17: Then go back to their own spreadsheets.

00:17:19: The integration is effectively broken even though wires are technically connected.

00:17:23: That's a profound point.

00:17:24: The break isn't in the software code, it is on social contract between teams

00:17:28: Exactly!

00:17:29: And this leads us to really powerful closing thought from Bill Hobbib.

00:17:33: He poses huge question about future of competition.

00:17:36: Let's hear it.

00:17:37: He asks if AI abstracts away the software interface If we are all just typing prompts into box and get our work done Then tool itself is no longer the competitive advantage.

00:17:49: If I have Salesforce and you have Salesforce, we both have AI agents running them.

00:17:53: who wins?

00:17:54: Right

00:17:54: it completely levels the playing field.

00:17:56: if the tool's just a commodity box that AI runs my tech stack isn't my advantage anymore.

00:18:01: Bill says The new Advantage Is Contextual Precision.

00:18:04: It's not about Who has the Better Tool.

00:18:06: Its' About Who Has the Deeper Understanding of their Customer To Distinguish A Real Buying Signal from the Noise Using Those Exact Same Tools.

00:18:14: Can You

00:18:14: Give Me an Example Of That?

00:18:15: Sure Two companies might see the exact same signal.

00:18:20: Company X visited The Pricing Page, the company that loses will just spam them with automated emails.

00:18:26: The company that wins.

00:18:28: we'll have a context to know why they've visited who the decision maker actually is and what specific problem their trying to solve.

00:18:36: They would use AI to execute a specific relevant strategy.

00:18:40: So technology becomes commodity And human insight the context become

00:18:45: the moat.

00:18:46: It's no longer the code, it is context.

00:18:48: That is a powerful place to land!

00:18:50: It feels like we've come full circle today.

00:18:52: We started with Jennifer Bray using AI To build her own tools which commoditizes software.

00:18:57: And when software is commodity The only thing that actually matters Is how well you know your strategy and customer.

00:19:03: It'a major call-to action for twenty twenty six.

00:19:06: Stop obsessing over which tool to buy.

00:19:08: Start obsessing on data structure Your team alignment and customer context

00:19:12: Which really leaves something huge to think about.

00:19:16: If AI agents are the ones navigating the web, and AI is the one writing code.

00:19:21: And our MarTech stacks are just these invisible layers of agents talking to each other... maybe the ultimate B-to-B advantage moving forward isn't digital at all!

00:19:29: If human context is the only true moat left, maybe the most radical marketing strategy for the next few years is just literal face-to-face human connection because that might be the only thing an AI can't scrape or vibe code.

00:19:43: I love it back to basics but powered by the best tech we've ever had.

00:19:47: if you enjoyed this episode new episodes drop every two weeks.

00:19:50: also check out our other editions on account based marketing field marketing channel marketing ai and btb marketing.

00:19:56: go to market and social selling.

00:19:58: Thanks for joining us on this deep dive and make sure to subscribe so you never miss an update.