Name: Martech in the Middle (SAS Interview)
Uploaded: 2026-05-05T16:42:07.032Z
Duration: 26 min 8 s
Description: Martech in the Middle (SAS Interview)

Transcript for "Martech in the Middle (SAS Interview)": Alright. Franz and I are back again here with Jonathan Moran, from, Seth. Jonathan, like, you've been with us on this journey of the state of Martech, for years. We're so excited to have you back again. Yeah. Thank you so much for having me. I'm, I'm I'm always looking forward to this session and talking with you guys. So it's it's great to be talking about a a bit of a different topic today. So excited to jump into it. Alright. Well, yeah. Certainly, a lot is happening in the world of MarTech. Just in case, for those who might not be as familiar with, you and, the role you play at SAS, could you give a brief introduction? Sure. So at SAS, you know, we're a company that's mostly known for data and AI, but we also have a set of solutions that we sell into the market to cover areas like MarTech, AdTech, fraud, risk, IoT, etcetera. We have industry specific solutions, and we have industry agnostic solutions. And I'm fortunate enough to to work with a team of marketers across what we call our horizontal solutions that sell into a variety of different industries including, my first love, which is MarTech. Excellent. Excellent. Alright. Well, let's dive right in. So I would say that the Martech stack conversation has been dominated by two layers. There's data, particularly CDPs and warehouses, and then activation, channels, personalization. But is there a critical middle leg that's been underappreciated? You and I chatted a bit about this. What is the middle and why does it matter now more than ever? Yeah. So, you know, just as you kind of said in your question, you know, when we talk about Martech stacks in their simplest forms, we talk about three main layers. Right? As far as, you know, the SaaS perspective. The foundational data and data application layer. Right? That's where cloud data warehouses reside. The top panel delivery and activation layer where messages are sent from email, mobile, social, etcetera. And then this middle decisioning and orchestration layer. And that's really that insight layer that's needed to turn data into, action or activation. Right? In the form of offers, messages, interactions. Historically, brands have invested very heavily in those outside layers. Right? Whether it's a large cloud data warehouse or it's a big email messaging or social media platform. Why is that? Well, in my opinion, it's because that's where most vendors have traditionally excelled. Where brands have not invested as heavily over time is in that middle layer. The layer where orchestration happens based on decisions that are being made, around who gets what engagement, at what point in time. You know? And and I think in this age of AI, right, as brands are adopting technologies throughout their stack, they're realizing they've gotta have that middle layer. They've gotta have AI and analytics to make decisions, about who to contact. And a lot of brands don't have the chops in the middle layer that they need. Right? Most of the biggest martech vendors don't really excel in that middle layer. And there's a reason for that. It's because it's not really what they were founded out of and and who they are. Right? So, you know, we've got companies that are very good at CRM and Salesforce automation. We've got companies that are very good at digital content creation. We've got companies that are very good at mobile messaging. But don't have as many companies that are really deep and well versed in analytics and decisioning. And so, you know, you talk to a lot of folks, out in the industry and and they'll tell you that that that middle layer is becoming the new battleground as AI advances. Every vendor wants a piece of that middle layer and that's why we're seeing vendors like CDP vendors and others latch on to this term of AI decisioning. But here's the thing, you know, they can latch on to the term, but are they are they really truly doing it? And and do they have the tools and technologies, to make it succeed. Right? So, you know, we're seeing a lot of marketing teams that are drowning in data rather than acting on insight and that's because they're they don't really have the right tools for the job. So, you know, I I would I would pose and I think you guys would agree, you know, we've gone beyond simple adoption of AI to really figuring out brands are trying to figure out how to, you know, derive and and and use AI with precision and utilization. Let's double click on that decision area part. We've been tracking, the the evolution of CDPs morphing from collecting unified platforms into something more like what you just said, context ready decision layers. So how how does SaaS or you see this transition, and, where does customer intelligence fit in that picture? Yeah. I think, you know, I think the shift that we're seeing, this evolution, it's natural. Right? It's what the market is demanding, that move from traditional CDP capabilities to, you know, CDPs becoming that context ready decision layer. For me, you know, I'm seeing a bit of a squeeze, a change in the traditional CDP model, if you will. Right? Because the traditional model involved the CDP managing its own proprietary data store. Customer data was ingested, unified, activated from within the CDP's own database. But, you know, this worked for a while, but it created a bit of a frustrating dynamic. Right? Because the golden customer profile lived inside of the CDP and was a bit siloed from the broader ecosystem. And so then, we saw the shift, right, towards composable or warehouse native native CDPs when, you know, the CDP vendors started sitting on top of cloud data warehouses, Snowflake, BigQuery, Redshift, etcetera, rather than replacing it using the warehouse as kind of the system of record for customer data. So, you know, the CDP provides the identity resolution, segmentation, decisioning, activation layer on top. So, you know, the squeeze is the pressure CDP see from the cloud data warehouse coming up and the activation layer solutions coming down. And that's that's forcing CDPs to become what they have become. Right? And and how they they started positioning themselves. And I think it makes sense because organizations don't want a second copy of their data living in a CDP silo. Right? They want, their data science and analytics teams to be working in the warehouse. And, you know, reverse ETL vendors have accelerated that concept by proving you could activate warehouse data without a traditional CDP. So, you know, the spectrum today looks like, you know, you've got your package traditional CDPs that have their own data layer, but increasingly, you know, offer that that very good warehouse connectivity. You've got the composable CDPs. They're purely warehouse native with no proprietary store. And then you've got the hybrid approaches. Right? These are warehouse vendors building CDP like capabilities directly into their platforms. So in in the SaaS context, you know, our embedded CDP capabilities fall fully into composable. Right? Because customer intelligence three sixty can connect to external data sources and cloud environments and doesn't, you know, require or maintain a proprietary store, which furthers our position as a context ready decision layers with with a, with a bunch of capabilities that surround it. Alright. So decisioning is very closely related to orchestration and, you know, particularly as you're saying here in the the AI app, you know, lots of people are now claiming orchestration. But also, it sounds like the term is applied with very different meanings from one vendor to another from, you know, I mean, simple journey builders to full cross channel coordination. Yeah. Can you walk us through what what orchestration looks like in practice when the middle layer this middle layer is doing its job well? What are the moving parts? Yeah. Great question. And and bear with me on this one because this is a this is gonna be a bit of a lengthy answer. So, you know, orchestration in its simplest form is the coordination of decisions along a customer journey across channels, time, context. So that every interaction is the right one, not just the scheduled one or the one that the brand forces upon the consumer. Right? So there are many moving parts that work simultaneously in harm harmony to get orchestration right. And, you know, I've gotta use an analogy here, right, to to to bring orchestration to life. And I'm not gonna use the orchestral one because everyone likes that one, but I'm not gonna use it. I'm gonna use the human body one. So, you know, if we think about and and tie the the the components of orchestration to the human body. Right? First up, unified customer profile. That's the food or the fuel for the body. Everything depends on that continuously updated identity resolved profile. Not batch, not a batch refresh snapshot, a live view that incorporates behavioral signals, transactions, interactions, so forth and so on. That's the first part of orchestration. Next part, decisioning logic. Decisioning logic is the brain. And the brain, you know, is obviously very important to the function of the human body. You know, real orchestration requires a decision layer that can evaluate for any given customer at any given moment. What is this person eligible for? What's the best action offer experience across all competing actions? What channel and timing maximize the likelihood, of the right outcome? Right? And this is where, you know, logic, business rules, suppression rules, fatigue controls, AI driven propensity models, all of these things have to coexist. Right? The true journey builder, it, you know, has all of these things included and if it doesn't, it's just a simple flowchart within a GUI. Channel connection and or integrations with bidirectional feedback. This is the nervous system. Right? Good orchestration doesn't just push through channels, it listens back. So, you know, did the did did an action occur? Did an email get opened? Did a push notification get dismissed? Did the customer call the contact center right after receiving an SMS? I think a lot of brands would love to have that information and know that and, you know, that's gotta be present for true orchestration, to happen. Right? If if if it's just a one way integration, that breaks the feedback loop and that that basically, you know, nullifies, true orchestration. Event driven triggers, those are the reflexes. Right? Those are, you know, you know, the ability to, you know, look at and respond to real time events, you know, a a browse, an abandoned, and a failed payment, a loyalty tier change, a geofence entry, and fire the right response in seconds or milliseconds depending on, the use case. And so what this requires is event streaming, which some organizations leverage, some don't. But being able to stream in data in real time and then, you know, transform and normalize that data is required for real orchestration. It can't be nightly batch ETL jobs. Right? And so, next up in orchestration, suppression and arbitration. This is the referee. Right? Within the within the human body. When multiple journeys campaigns and triggers are firing simultaneously at scale, who decides what what actually wins. Right? This is decisioning arbitration requires good governance, that ensures a customer doesn't get three messages in one hour from three different campaigns from the same brand. And I know all of us have experienced that. Cross channel continuity, that's the that's the memory. Right? A customer who ignored an email then saw a paid ad then visited the website, that should be treated as one continuous experience, not three separate interactions in three separate channel tools. And, you know, orchestration really requires that coordination across channels and departments. And that requires brands to break the traditional business model of setting up functions, engagement functions by channel. I've got an email department. I've got a social media department. I've got a mobile messaging department. Those those can't be, you know, siloed. And then finally, optimization. Optimization, in the human body and and for orchestration takes the form of, you know, AI and ML integration at decision in time. So can your decisioning logic call live models? Can they call churn propensity models, next product to buy, optimal send time, message variance selection at the moment of decision? Again, not in batch or not based on a score that was computed perhaps last week. So this is the difference between model informed orchestration and then true adaptive learning and and learning, based orchestration over time. So point here, orchestration is hard. True, deep, advanced orchestration requires a lot of tools, technologies, processes, capabilities to be working, simultaneously all at the same time. Quite a metaphor there, Mike. Well done. Yeah. Yeah. I think you can build a health check on that one. Yeah. AI agents are one of the biggest themes in our state of the market 2026 research. How does an intelligent middle layer interact with the autonomous AI agents? Is the middle layer where, agents get their context, constraints, or rules of engagements, if you will? Yeah. I mean, I I think that the the middle layer is absolutely critical for autonomous agents to function. Right? This because this middle layer is the brain that provides intelligence to those autonomous agents. We know that autonomy without direction often results in in mayhem. We've seen that in society. I mean, we've seen that since the fall of the Roman Empire. Right? But I would argue that, you know, even more than direction, agents need grounding and rather continual monitoring, at least right now. What happens when you give an agent a direction? Right? Like, optimize click through rates and and the agent then starts hammering a single customer with messages. Right? Cannibalizing other marketing efforts or campaigns, violating suppression rules, making offers that a business can't honor. Right? That's that's trouble. So the middle layer is the place where, you know, agents get four critical components. Right? They get first context. Right? All relevant data data about the customer is three stage preferences, interactions, constraints. Right? So constraints, the middle layer defines eligibility rules, suppression rules, fatigue limits, channel preferences, business rules. And then, you know, the third c here is is compromise. Right? Which is essentially arbitration. In a multi agentic world, these agents are gonna have to make deals with each other. Right? And because they'll all be competing to act on the same customer simultaneously. The next best action agent might really wanna talk to a customer and get to that customer before the retention agent does. And then the email agent comes in and says, he or she can save the day with the message instead of the push notification agent. So you get the idea. The middle layer becomes the referee to ensure the customer receives a coherent experience. And then the fourth c is cognizance. Right? After an agent acts, this is this is kind of the memory or closing the loop. The middle layer records what happened and updates the customer context accordingly. And this is very crucial because, you know, agents have to be able to learn from outcomes, you know. And if that's not happening with a middle layer that closes the loop, you know, the agent acts, the outcome is not recorded, context is not updated, and the agent is not well informed for the next interaction. So, you know, it really, you know, it really requires the four c's context, constraints, right, compromise, and cognizance. Wow. That's that's a critical role. And I can imagine where where do you see this middle layer heading over the next two to three years? I mean, traditional middle layer was plumbing, you know, invisible, utilitarian type of stuff. You're describing something much more strategic. So does the middle layer become the new center of gravity of MarTech stacks? Yeah. I I I honestly think that it it does. Right? It's it's moved from being an afterthought to being not just the center of the MarTech stack, but the really the intelligence layer itself. And in the age of a Gentec AI, this intelligence layer becomes more than a decisioning engine, but a sensory nervous system, back to our analogy, that coordinates the entire customer experience. So, you know, as I mentioned before, the middle layer has long been overlooked due to the dominance of certain players in the space that were born in and of the outside layers in the stack. And so, you know, as agentic comes more into view, it's going to change that activation layer. Right? Because that activation layer is gonna be more heavily agentic in the future. And so the middle decisioning and orchestration bit, becomes critical. So I do see it as the new center of gravity. And I think there's I think there's others out there that that would agree with me. Alright. One of SAS's historic strengths, has actually been code level adaptability, with your platform. You know, now there may have been a, you know, some folks who found that to be a barrier for, like, okay. What can buyers actually, you know, do with the code? But it kinda feels like that's becoming more and more of a superpower because with AI making custom code easier than ever, I'm curious, like, how does this position itself across that spectrum from, like, out of the box UI, an amazing UI, you know, for the product, but also to more fully custom crafted solutions. I'm kinda wondering, is this like a have your cake and eat it too moment for you? To some degree, I I think it is. So SAS has long positioned itself as providing capabilities for both do it yourself persona as well as the do it for me persona. Meaning, if you wanna code, great. And if not, that's okay too. And what's what's awesome is you can create custom code and solutions outside of SAS, you know, Python or what have you, and drop it right into SAS solutions for use. So we definitely position ourselves as very much open and agnostic, working with providers of all types. And that's one of the reasons so many brands look to SaaS to provide that middle decision in orchestration layer is because we work well with both upstream and downstream technologies. You know, we often say that the market is coming to where SaaS has been, which in essence is that, you know, have your cake and eat it too moment. Alright. So let's let's close this out, you know, for the marketing ops and MarTech practitioners who are just now reading the state of MarTech 2026 report, you know, and yeah. Picture the people who actually have to implement, and manage these stacks, what's your advice for how they should be thinking about this middle layer today and where should they start? Yeah. So, you know, I I feel like a lot of our conversations go back to the data. Right? And we know that, you know, as with traditional MarTech, you know, you gotta have good data for good campaigns. You've also have have to have good data for good AI. And, you know, identity resolution is a is a critical component to MarTech data management, You've gotta get that right before anything else. So, you know, the orchestration layer is only as smart as its ability to recognize the same person across channels. And most decisioning failures that we see, they're not really logic failures. They're identity failures that make logic look broken. And so this has gotta be fixed first. I think the the next big component is, treating this middle layer as infrastructure, not as a a campaign tool, a bolt on, an afterthought. And this orchestration layer should hold the rules and logic, you know, and and campaigns and journeys should consume those rules. Suppression logic, frequency governance, channel priority, should be standing infrastructure, not things that are configured by campaigns. And I think the separation is what lets you scale without having to rebuild over and over and over again. And then, you know, my last kind of piece of advice, and I've been thinking about this a lot is, you know, the process component of decisioning and orchestration. You know, you've gotta be able to know who owns the that middle layer. Right? It it it's it's it's gotta be named. It's gotta have an owner. And the processes have to exist where, you know, handoffs are very explicitly detailed and, you know, people are held account accountable for kind of those handoffs. So, you know, the place that orchestration most commonly falls apart is at the boundary between marketing automation and the sales and and service systems. So if the answer to, you know, what happens in the CRM when an orchestration action fires is it goes into the activity log, right, and gets forgotten about, then the loop is broken. Right? There's gotta be a a closed loop. So that hand off is a relationship design problem with, you know, sales operations and others before it's a technology problem. So I think documenting and really understanding processes as far as how decisioning and orchestration is actually going to work inside your organization is is so important before you start buying tools and technologies and and trying to to fix any decisioning and orchestration problems that you might have. Awesome. Thanks, Jonathan. Alright. So, the theme for 2026 here, it's not Malcolm in the middle. It's, Martech in the middle. Yes. Okay. Fantastic. Anyways, love that. Thanks again for joining us on this. Thank you again for your your support, for State of MarTech. And, I'm very excited to see this what the middle enables so many companies to be able to do moving forward. I I really like the way you frame this. It's almost like an infrastructure, type moment, from our tech. Yeah. For sure. For sure. Thanks so much for having me, guys. It's great to to speak with you once again, and, hope to talk again soon. Sounds great. Bye bye. Yep. Bye bye.