Unfiltered takes on the biggest shifts in marketing technology. We spotlight what matters, who's leading (or lagging), and what's next. In Martech, clarity is power — and we're here to deliver it.
00:00:01.280 — 00:00:25.080 · Jacqueline Freedman
Welcome back to Making Sense of Martech. This is Part 2 of my conversation with Tony Byrne. In my eyes, your sharpest claim in the whole book was actually not about the stack. It was about the industry's structure. And so you wrote, "Thirty years in, we can declare a victor: composable MarTech / Digital stacks have won the debate."
Who do you think actually won?
00:00:25.120 — 00:01:28.460 · Tony Byrne
So the enterprise won. There has been really, throughout my entire career, this huge debate between suite versus best-of-breed. It's played out in all of the different technologies that we have covered. And on the martech and digital side, it was a particularly consequential debate. And you had, on the one hand, certainly Oracle and Microsoft and even IBM back in the day.
And now, of course, Salesforce, Adobe to some extent, Acquia, and HubSpot in the mid-market saying you can find safety and security and robustness by going all in with us. Since integration is always a challenge and it's a frustration for your non-technical leaders to even hear about integration challenges, what if you just turn it all over to us and lo and behold, your integration challenges will poof,
disappear. So it's a very powerful argument. These companies are also shameless about going to your board and your C-level and deceiving them about the benefits of going all in with them, and that's really hard for the martech people who are trying to get [bleep] done.
00:01:28.500 — 00:01:31.620 · Jacqueline Freedman
Nothing grinds my gears more than when that happens.
00:01:31.700 — 00:03:13.490 · Tony Byrne
It's fundamentally unethical, too. But that's that's who these companies are. And so we actually looked at, you know, the 30 some odd members among our corporate membership who are all Global 2000 firms or soon-to-be Global 2000 firms, 10 of them in the Fortune 20. And we said, okay, what's in your stack?
And yeah, they had some pieces from Salesforce, SAP, Adobe, but only one of them had gone all in and now is kind of actually pulling back from that. And they actually did it because their agency took them there. All the others had what we would think of as composable stacks. Yeah. Maybe they used this tool from Adobe, that tool from Salesforce.
But if you looked at their stack, it was necessarily a diverse collection of tools from a broad set of vendors that have different platforms and assumptions and integration. Yeah, is an issue. But it's no less of an issue than if you just went all in on a big suite. Because these are big, complicated companies and they have the resources to do this, but also life and leadership took them there.
And really smart enterprise architects understand this. And so who won? Composability won, we believe, at least at the large enterprise level. It won the reality on the ground. Now what's the reality out there in analyst land and on LinkedIn. And what's getting whispered into your CMO's ears? That's a different story.
But the composable model won. It won handily. There's evidence that there's actual metrics that prove why it's better. And so now let's just declare that and move on. I feel like, so who won? The enterprise won, as it should be.
00:03:13.530 — 00:03:30.930 · Jacqueline Freedman
I love that, and it makes my heart so happy. Because why wouldn't you want something modular, flexible that you can plug-and-play? Is it easy? No, but it is more easy than being stuck in one universe.
00:03:31.050 — 00:03:40.810 · Tony Byrne
And let's be clear, these are pretty [bleep] universes too. Like, yes, there are significant problems with them. Even if you did believe their story about integration and one belly button to push.
00:03:40.850 — 00:04:10.700 · Jacqueline Freedman
Yes. All right. Taking a step back for the industry at large. So maybe, you know, like the Reputation album. Your advice flashes red on Adobe, Salesforce, and Acquia. Yet suites are still very much the default shortlist in a lot of enterprise RFPs. If composable actually did win the debate, like you say, why is procurement still losing it?
And why are industry analysts saying otherwise?
00:04:10.740 — 00:06:06.850 · Tony Byrne
Yeah. So you know, we talked the last time about a couple of things. One is that these vendors have a lot of mindshare. They can afford to have a lot of mindshare. And they're also whispering in the ears of your CEO and very often your board. And there's questionable ethics behind that. But in any case, it's real.
When people turn to AI and say, who are the leaders? Of course these guys are going to show up because these analyst firms and others have declared them leaders. And so AI is giving people some bad answers. And then I think the other piece is in many cases, they're already incumbent in your environment. So somebody may already have one of theirs.
So they can't kind of not include them. And in fact, it's so interesting because, you know, a big part of our work is providing short lists to our corporate members. And sometimes we'll actually recommend them on the short list. And the reason why is that if you follow a proper empirical selection process, they're very likely to lose, right?
Which is why they push back against this process very often. But if you follow an empirical, agile test based selection of the kind that we would recommend or maybe guide you through, they're very likely to lose. Right. And so but you still you have to have that evidence that shows them losing. We call this internally.
We call this slaying the dragon. Right. Like they're there. And so let's just have them let's test them out. And that way later when inevitably you're going to run into some problems with some with the winner of the RFP. Right. There's no perfect software out there. And there's going to be an integration hassle.
Something's going to go wrong. It doesn't mean you made a bad choice. It's just that's life, right? And then you want to have that point where somebody says, well, if we had gone with, you know, mega vendor suite, we never would have had this problem, you say, well, actually they did lose and they lost for a reason.
And we have the evidence to show you that and the team. This multidisciplinary team that we put together. Ranked them fourth out of four. So shut the [bleep] up. Let's just keep going. Yeah. So sometimes you have to slay the dragon.
00:06:06.850 — 00:06:24.570 · Jacqueline Freedman
So we of course have been thinking about the enterprise a lot. And we've talked a little bit about mid-market, but for a lot of mid-market companies, a suite like HubSpot is their stack. Is red the right call for them too, or is the suite their foundation layer? How does it differ?
00:06:24.690 — 00:07:11.550 · Tony Byrne
Yeah. So I'm going to say yes on this one. And the key thing I'm making a subtle but I think important distinction between a suite and a kind of all-in-one box set. So a suite has all these different products that nominally are supposed to work together, but often target different personas or different use cases and end up being very disconnected.
And so HubSpot is different than like a CEP that we talked about a customer engagement platform where everything is all on the same dashboard. It's an integrated whole. You may not like all the pieces of that whole, but it is an integrated whole. HubSpot is not an integrated whole, it's a suite. You know, they like to call themselves a platform.
And so suites have just not done particularly well and have led to, I think, more than their fair share of frustration relative to the alternatives. So yeah, we're still kind of flashing red on HubSpot too.
00:07:11.590 — 00:07:21.870 · Jacqueline Freedman
I don't blame you. So is build just a new kind of lock in, like where your own headcount becomes the vendor and your own limitation?
00:07:21.910 — 00:10:11.960 · Tony Byrne
Yeah, I mean, that's a fair criticism. I think for two and a half decades now, we've been recommending to companies that, look, you're not a software vendor. Don't build stuff, right. If there's a buy solution, go buy it. Whatever the issues are, there's always exceptions to that. I think we've seen that typically we've seen really successful build in like B2B CDPs, right, which has been a hole in the market.
We've seen successful build for media companies with their own CMS, because if content is your product, you know, you have very special needs around it. So there's, but the case for that is narrow and narrowing each year. Like you shouldn't necessarily be building software. Having said that, this new MarTech 4.0 world is going to be about composing highly bespoke experiences for humans and agents from some base platforms, and all this middleware is likely to to dissipate.
And so what is it going to take to get there? It's going to take some build. That build is likely going to be around assembly and composability, and some vendors may end up stepping into that breach a little bit better than they are today. So there's always been a build around integration. There's always been a little build around assembly and that build is not going away.
And so what we're telling people is don't necessarily build this yourself, but the spend that you have with systems integrators is unlikely to go down. I mean, I'd love to tell you that it is, but I think that it's more the systems integrators are likely to to really survive in this new world. Separate story with your agencies.
We're hoping hopefully your agency spend will go way down. So to come back to that word build. And I sometimes say there's going to be a bit more build. But what I really mean is that there's going to be a bit more integration and assembly building in some controls over these things. You know, take one example, Jacqueline, and that's this so-called context layer that everybody's talking about and how important it is.
And it is important to AI. You need to own that layer. You don't want to outsource it to a vendor or to your agency. The context layer is going to have content, data and decisioning. And there's no platform out there, despite what your lakehouse vendor will tell you. There's no platform out there that can just be a plug-and-play context layer.
It doesn't exist. And so that will have to be at least in the year of our Lord 2027. That's going to have to be built at some layer. And what we're focusing on is design patterns for this. And so are we asking you to build heavyweight software or to build real software? No, but what we are saying is that you will be reducing your spend with software vendors.
You hopefully will be reducing your spend with agencies, and you'll want to realign some of that spend towards some glue work and assembly work, for which there aren't going to be really great package solutions in the near term at least.
00:10:11.960 — 00:10:20.240 · Jacqueline Freedman
But then I'm curious, where do you consider maybe custom APIs or custom MCPs as part of the build?
00:10:20.280 — 00:11:00.850 · Tony Byrne
Those are part, you know, API layers, MCPs, those are all part of the fabric of this new world where your experiences, particularly real time experiences, are going to be assembled from, you know, a number of piece parts. And again, the channel is going to kind of matter a lot less. And if you go all in with the CEP and a CMS and an e-commerce platform, they'll have the pieces for you to do that to some degree.
If you're an upper mid-market or an enterprise firm and you have, quite necessarily, a more composable stack where you've really invested a lot in the foundation. There's going to be more tech work there, but hey, that's why we call it martech. And my people, you can handle this.
00:11:00.850 — 00:11:23.650 · Jacqueline Freedman
So in your book, which I will put on the screen once again, Chapter 14 covers new supplier models. So if the suite relationship is completely out, what replaces it? I feel like we've talked about a couple of things, whether it's best of breed, you name it. But I guess what is is our new term. Is there a new way of thinking about it?
00:11:23.650 — 00:13:22.880 · Tony Byrne
Well, you'll still have a stack, and these stacks will still have vendor components in it. It's going to be more infrastructure heavy. That infrastructure will probably not be monolithic platforms either. So we talk about a customer data ecosystem with three hubs. They're logical hubs which is your customer 360, your customer activation, and your customer analytics.
These are managed by different people for different purposes. They relate all to each other. They could be 3 to 9 different software platforms, right? So there will still be software in your stack. It's just that the channel pieces that we have all been leaning so heavily on are going to become a lot lighter and ultimately probably go away.
So in terms of supplier models, you're probably dealing with a different set of suppliers. There will be, at least in the near term, probably through 2030. A lot of glue work that we're gonna have to do ourselves, hopefully following really good design patterns. Your systems integration spend is therefore likely to continue to be robust.
But your agency and vendor spend like we just mentioned ideally will be going down. And I think the agency piece is super interesting because agencies we've relied on for really three things. Three related things. So agencies provide creative spark and vision on demand. They provide production work also on demand.
Then they provide outsourced operations whenever of any kind, whenever you need it. And in exchange, what we've done is we've outsourced a lot of, in many cases, infrastructure to agencies as well. We use their DSPs, we use their analytics programs, we use their marketing mix models, tools.
We're building sometimes customer data and identity graphs within their world using their digital asset management systems. And it's going to be increasingly bad as we go into this AI era. So our sense about agencies in particular is that, yes, you want to continue to use them for their creative spark and the ability to create hopefully original creative assets.
00:13:22.960 — 00:13:25.520 · Jacqueline Freedman
Hopefully not AI-generated assets.
00:13:25.840 — 00:14:34.540 · Tony Byrne
You will still need AI-generated derivatives for all the different personalization you're doing, but that then becomes production work and you should capture all the benefits there, and it should happen within your world and your creative ops platform and marketing ops. And you get those production efficiency benefits.
And you also then own the analytics too. And then some people like the fact that they can have variable, you know, campaign specific, you know, rather than having their own in-house agency where they've got FTEs, they just do the agency and they're elastic and they can expand a level, you can still lean on an agency for kind of overflow work, but they should be using your platforms, not theirs.
But secondly, the AI learning loop is we know now where the ROI is, right? Going through this loop of AI insights to AI, decisioning to generative AI, and then figuring out what works and looping back and getting smarter. The value's in the loop and constantly getting smarter and constantly iterating.
And you. That can't be a black box for you, if only for compliance reasons, but also for value reasons. And so you need to own that loop and not have your agency own it. So in the future we predict a much lighter but higher impact spend with your agency.
00:14:34.540 — 00:15:22.790 · Jacqueline Freedman
I agree. And the whole time you've been talking about the more technical components, the more process components. I've also been thinking about the institutional knowledge that you outsource, and as a result, you have a worse outcome because internally you can't push back. If you are, maybe you're not in charge of the spark.
You're not the catalyst for things, but you oversee the general vision or how it fits into your broader goals as a company. Yeah, and the amount of outsourced knowledge I have seen, not just, you know, the creative kind, but also the technical when you don't even know how to do the things in your own platform that the agency does for you is something that I always am struck by.
00:15:22.830 — 00:16:15.850 · Tony Byrne
Yeah, it's an abdication at the end of the day. And yes, it's scary and I understand how it happened. So again, I have some empathy for a marketing team that's under a lot of pressure and can outsource some of that stress. But by the same time, and it's not even just knowledge of the technical platforms, although that's important and we've both seen that.
But even like knowledge of the customer and what works and things like what are the ways we have to manipulate the data to get the analytics to work out. We've seen that in spades. And what happens then is agency people have turnover too. And sometimes we've been in situations where our analytics got delayed by two months because the agency, the person who knew the Excel spreadsheet, where we normalized all of the metrics, left the company and nobody else knew how to do it.
You know, it was like, so yeah, that can happen to you as a company that can also happen to your agency. So bringing the knowledge in-house is really important.
00:16:16.330 — 00:18:00.640 · Jacqueline Freedman
Yeah, I think it's a very overlooked piece because yes, if you are strapped for time or a headcount and you can outsource certain things to an agency, that's perfect, when you're outsourcing more than just production is when it gets really tricky. Because to your point, as it relates to the analytics piece, I've been in-house and with clients where no one has the same definition or reporting structure because not everyone has all of the knowledge.
And that's some of my favorite projects is not interviewing, but asking every different department. Like how do you define what sounds so simple a free customer? How do you define it? This customer? You won't be surprised, but I think many of us would be surprised how unusual it is to have the same answer.
And it goes back just the fundamentals of not just documentation, but actual definitions and logic and filtering and considerations. When the business changes or new business departments new product lines come into the fold. You have to be mindful of, oh, this might shift a couple of things because they could be a free user of this product, but they're still not a free user.
They're more in this stage. And so it's just one of those things that even when you don't outsource, that it can go awry internally. If that knowledge is not shared. So the godfather of martech, Scott Brinker, wrote the foreword to your book, and his landscape tracks a stack that never stops growing. It's kind of wild.
And while you argue for a smaller, more intentional core, where do you two disagree, and did that tension survive the edit?
00:18:00.720 — 00:19:24.580 · Tony Byrne
We maybe have a different lens on it. I mean, Scott has had many great insights. One of the great values is that he has charted successfully the superset of marketing technology players. Right. Or whatever. I forget what's up to 15,000 or something now, and that's really useful. Generates a certain degree of vertigo.
But but that doesn't mean it's not real. It is real. There really are that many. I think where our focus is, because we tend to focus on the enterprise, in particular larger enterprises is among those. What are the 160 to 200 that really matter and are real, you know, technology vendors and that kind of thing.
And so we have a little bit of a we're focused a little bit more on a subset. And Scott, I must say is also he's an excellent futurist. And we have although I wrote this book, we've tended to be, you know, a little bit more about what our clients are telling us. Having said that, Scott has also done, with his partners,.
Some really great survey work too. So, you know, the guy adds a lot of value to that extent. I don't think that there was any tension that that survived the editing and in fact didn't edit his intro at all. It was written from him. And and what's so great is that he has a very distinct voice, and that voice comes out in everything that he does, and he has this kind of this eagle eye vision across the whole landscape, across the whole world.
That's pretty unique. And I've always really appreciated that about it.
00:19:24.580 — 00:19:35.980 · Jacqueline Freedman
The inconvenient one. So Real Story Group works only for buyers. And your own podcast ran an episode called "Why IT Industry Analysts Suck."
00:19:38.020 — 00:19:48.310 · Jacqueline Freedman
Your revenue depends on keeping buyers happy with advice they want to hear. How is that not the same conflict aimed at a different audience.
00:19:48.670 — 00:22:52.040 · Tony Byrne
I think this is an evergreen challenge in advisory and consulting work. Like, who are you trying to help? Can your empathy go too far where you're not giving the best advice? And I think this is a boundary in advisory and consulting that we all have to kind of understand and know when we're crossing it. I will confess that our team is very focused on martech leaders and making them better leaders and helping them out.
And so that does put us at risk of sort of are we advancing this person's career or are we doing what's right for the enterprise and we try, at the end of the day, to do what's right for the enterprise? I remember when I was first starting out, this older consultant kind of would-be mentor to me, described this problem.
And he said, you know, at some point you have to stop advocating for the client and start advocating for the project. And I never really liked that because project success is not necessarily what we're looking for here. I understood the concept and he was kind of getting at what you were talking about. What I think it is, is at some point you have to start advocating for the business, assuming that the business is very clear about what they need out of this.
And, you know, there have been times as any consultant who's been doing this long enough will tell you when we have had to tell the people who brought us in that they were not going to like the answer that we were going to give them. We, for example, tend to lean very heavily on enterprise architecture. That enterprise architecture is your friend, that it's the basis for future decisioning, that with each passing year, it's more important than ever to martech.
And so there are times when an EA lens isn't necessarily the lens that a marketing person wants to align around. And we have to have a tough conversation then. And, you know, it goes back to your discovery has to be really, really, really. And sometimes we have to drag this out of companies get to to the extent there is a leadership vision and plan that is a plausible and coherent plan about how they want to make money going forward.
That has to be the Rosetta Stone behind all of this. And then you just, you know, by the way, if you go back and tell your leadership that you're going with a vendor that's not Adobe or Salesforce, you better be able to tie it back to that Rosetta Stone of top line and, you know, efficiency, effectiveness, all of the traditional kind of business levers.
So, you know, it's a challenge. I'll admit that we have a bias towards really wanting to support the martech leader. A lot of times they're sandwiched between an upper leadership that doesn't always understand what they're trying to do and and people underneath them who are frustrated and wanting to just go nuts with AI in ways that might or may not be, you know, productive.
And so it's kind of a sandwich generation, and we're really empathetic. And we've watched successful martech leaders move up the ladder in their own companies into other careers. And a big part of our business is wanting to support that person. But sometimes that's not. Occasionally that's not the right thing to do.
00:22:52.080 — 00:22:55.040 · Jacqueline Freedman
Good for a person within a company, they're very different.
00:22:55.080 — 00:23:35.080 · Tony Byrne
Yeah. And conversely, you know, we've been in situations probably even more common where the person was right and the company was wrong. And we've had to have this tough conversation. Like, my friend, you were in the wrong place. I understand why you may need and may want to stay, but there's nothing I can tell you that's going to make this better.
Your leaders are just screwed up and they have issues themselves with ethics, or confusing business model or conflict at the C-level. And you, my friend. There's nothing that we can put in a PowerPoint that's going to allow you to fix that. We can get in front of them and tell them, and it still may not fix that.
And those are tough conversations to have, too.
00:23:35.120 — 00:23:44.370 · Jacqueline Freedman
They are. But I think the person on the receiving end is grateful. Either it's validation or confirmation of, okay, I have to accept the terms and conditions.
00:23:44.970 — 00:23:45.730 · Tony Byrne
That's right.
00:23:45.770 — 00:23:50.130 · Jacqueline Freedman
Yeah. If only we could figure out a way to go upward and get that sorted.
00:23:50.570 — 00:24:21.690 · Tony Byrne
Well, we find ourselves doing that more often, so we're pushing that like, hey, let's do a workshop with the leadership. Let's have a chat with the three board members who are really dialed into marketing. And in many cases, these are very smart, flexible people who just needed to hear something from someone else and maybe framed in a different way.
I pay this advisor to tell me what I really wanted, and sometimes there's some truth to that. I won't say otherwise. Sometimes there's a little bit more to it than that.
00:24:21.730 — 00:25:12.150 · Jacqueline Freedman
Yes, to your point, it's usually the packaging, the delivery, the approach. Because yes, whenever I approach a new client and I'm like, all right, tell me what's going on. They already know the answers. They already know what the right thing is. They just need some guidance or bumper cars to get there. And it's just trying to avoid those landmines.
This episode is brought to you by our sponsors. Hightouch is an agentic marketing platform that reinvents the lifecycle marketing workflow, making it AI-first and warehouse-native. Brands like Thumbtack, MoneyLion, and Fullspan Health have reduced campaign timelines from weeks of work across 6 to 10 teams to days or even hours with Lifecycle Studio.
If you're not sure what to make of all of this acquisition news in the email marketing space, they'd love to talk to you and show you how they can help. See what Hightouch can do for you at hightouch.com/msom.
00:25:13.990 — 00:25:52.680 · Jacqueline Freedman
Brace yourself, email marketers. Salesforce is buying Braze, and change might be coming whether you plan for it or not. The good news is you get to choose what comes next. If you're a retailer, choosing an AI company for personalized customer engagement is a must. Bloomreach powers brands like American Eagle, Pandora, and Benefit Cosmetics who win the inbox.
The secret is Loomi, Bloomreach's intelligence layer, in every email you send. It personalizes each message to the person reading it, delivers it when they're most likely to open, and can build entire campaigns from a single prompt. The result is email that feels 1:1 at any scale. See it in action at bloomreach.com/msom.
00:25:54.040 — 00:26:34.720 · Jacqueline Freedman
And now back to the episode. So we've been talking a lot about the perhaps internal mechanisms of any singular company. I'm curious about how a CDP is always a never-ending project in your eyes. You've also described CDP owners whose marketing partners can't yet supply the logic or content to use that very data.
So if enterprises couldn't execute the last version of a foundational layer, what's the evidence that they will be able to execute data, content and decisioning in this new layer and semantic layer and all the things that are coming and evolving.
00:26:34.760 — 00:29:34.250 · Tony Byrne
You're getting to what is an incredible dilemma that we see right now, and possibly the leading issue as we try to transition from 2.5 to 3.0. For those that are trying to do that. And I would say that's where a lot of our own corporate members are at. And so, yeah, CDPs and I would say customer data projects in general, because CDPs typically are just sort of one part of your customer data ecosystem, they're never-ending.
And any data person will tell you this, like, because data is the best source of the truth that you have in that moment, but it's never perfect. And so you're always trying to improve on it and, and make it better and put more pieces into the picture and have better predictive models and all that kind of stuff.
So to that extent, having said that, I think of the three pillars data, content and decisioning. Generally speaking, within a large enterprise, they've been working longer on data. It is the more mature pillar, not always but usually. And so you run into this block. And in fact, we've seen in some cases enterprises literally decommission their CDP because of this issue that you raise secondarily, which is, okay,
I have much better access to customer data, and I can activate it in all these different channels, and that's wonderful. But do I have the content to provide personalized messaging that this now affords me? Maybe not. Can we agree that these people should be retargeted and these other people shouldn't or.
And why? And who's going to pay for it and which promo are we going to fire on our website. You know, and so there's some math there, but usually there's a lot more governance that comes into play. And so this all really speaks to the issue that you can mature one of your three pillars, but the other two can hold you back in a world where you need to have insights, drive decisions, which then tell you which content and messaging to display.
My colleague Jarrod wrote a book that also just came out. Jarrod Gingras called Getting to DAM 4.0 and DAM 4.0 is basically the MarTech 4.0 equivalent of just zeroing in on the content pillar. And what does it take to be able to do that, and there's a maturation that's required there. Decisioning is the hardest.
It's always been the hardest of the three pillars. We've been surveying our corporate members for eight years now. And they without exception, when we dig into what really what's the degree of maturity that they have, the decisioning pillar is the least mature, particularly if you think about decisioning from a customer omnichannel perspective, as opposed to the personalization that I'm just doing on my website, and even the personalization I'm doing my website, it's not always a lot of evidence that it's driving value.
So that's hard. It's really hard. We have there's some thoughts in the book. We have a whole chapter about what you can do to get to decisioning 3.0, but the larger theme here is you're absolutely right, or get close to perfection in our customer data. And if we don't know how to leverage that data, if we can't execute in terms of leveraging that data, what did we get to?
00:29:34.290 — 00:30:32.670 · Jacqueline Freedman
Very true. And also there's so many different ways the term decisioning can be played and used. Whether it's a platform itself, it's homegrown, it's a it's a wide open pasture and also intimidating. Stacks definitely can change on a roadmap and people oftentimes change on a very different clock. So sometimes 2030 is a little too close for evolution.
But I want to transition into maybe some things that keep you up late at night, whether it's Midnights, I don't know. So every one of these stack changes, they always land on a person and they're responsible. And you say email production managers and paid media production roles largely will go away. We've kind of alluded to that with the agency component. Who will get reskilled, and who doesn't?
And lastly, is there optimism?
00:30:33.110 — 00:32:18.540 · Tony Byrne
I'm incredibly optimistic in general that you know, the smart companies are not going to hollow out their martech team or their marketing operations teams. Having said that, and there's a couple of chapters in the book on this. The roles are going to be very different and the environments that these individuals work in, it's going to be very different.
And people talk about a cockpit, and I do too, which is more a 4.0 world. Who gets reskilled and who doesn't is a really interesting question. I mean, I think it starts first with what are the capabilities you think you're going to need? And we tend to think that a lot of them are going to be around marketing, data operations, journey orchestration, journey analytics and these sorts of things.
And so who gets reskilled and who doesn't? Our answer to that is find your best people and start having them work in a more cross-channel way. Bring them to the extent that you have the capabilities, bring them more to sort of base roles and have them work on the hard problems that are not channel specific, and begin thinking about taking your metadata manager and converting her into a graph manager, which you are likely to need.
And so these are the sorts of, you know, roles that are likely to be needed in the future and start thinking about who's going to be your input manager into your decisioning subsystem, because it's coming for sure is hard to suddenly switch the gears overnight. But I think identifying your best people and getting them up to speed on some of this stuff is really the way to go.
So I don't think about it in terms of who gets reskilled and who doesn't. Hopefully your whole team can reskill at some level, and for a lot of them it will be natural and they'll want to do it. For some people it may be natural, it may not be natural or they don't want to do it. In which case, you know, you have these conversations about the world changing.
00:32:18.580 — 00:33:20.830 · Jacqueline Freedman
Yeah, I, I see a really easy schism in this. The way I learned marketing was not by channel, was not by segment. It was how do we do all of these things as a very small team and. you know, you would rotate that, oh, I have capacity. I'll take this segment and do these things. And I have a feeling to your point, the reskilling will be a lot more holistic.
It will be by no means. You need to be a jack of all trades and know one of everything. But it's like, all right, the channel doesn't matter so much as if we're getting the person at the right time in their journey based off of their actions and really honing in on what is best for the customer, hopefully. Yeah, because nothing grinds my gears more when there's just a certain segment.
That's all I do. It's like, well, this individual might become something else at some point. Don't you want to continue?
00:33:20.830 — 00:33:22.550 · Tony Byrne
Yeah, there's a person on the other end of this screen.
00:33:22.590 — 00:33:54.410 · Jacqueline Freedman
We're multifaceted. We're not one segment. We're not one, you know, very targeted list. And so I'm hopeful in maybe the, the soft landing of the reskill is like, all right, let's just at least think beyond a singular segment, a singular channel and way of communicating. So if we're going to be really practical about this, we've been very pie in the sky of what is coming.
What can a journey overseer, a journey doer, a lifecycle person, an agent controller actually do tomorrow?
00:33:54.450 — 00:35:44.960 · Tony Byrne
They can do tomorrow. You know, you said, like, on a Tuesday. And in fact, I think we have a Tuesday, you know, where a travel business was disrupted by a bad storm and whatnot. But let's just talk about the loop for a second. So we have a loop of constant learning and adaptation that's going on in real time.
And also there are longer term loops. And so, you know, you need if you're a journey overseer and for that matter, an agent controller, you need to be thinking about and working with your peers on what's before the loop, which is what are the inputs and the constraints and controls that we want to place on this, and in particular, what's the outcome we're looking for?
Then you may be in the loop, which is making real time decisions rather than an AI. You may be over the loop or on the loop, which is you're just watching what's happening and maybe adapting, right? And then there's after the loop. So there's four phases here. After the loop is AI is saying that this happened.
Is that really what happened? What takeaway should we make? It's recommending that we make the following four adjustments. One of these doesn't sit right with me. I think it's hallucinating. Whatever. So there's an after the loop, too. And so that's what journey overseers are doing on kind of a day to day basis.
What are they doing on Tuesday? That's what they're doing on Tuesday. On Fridays they may be hopefully that to your point, taking kind of a longer picture, working with their peers, what's really happening with our customers? Are we serving their best interests. What are the different loops that they're on?
What do we do with people who are simultaneously on 2 or 3 journeys? Our governance isn't telling us which one to prioritize. We need to work on that. So there's a lot of strategic work that has to happen too. And so at a tactical level, it's very different. And it's going to rely a lot on human judgment and monitoring and and checking.
It's going to be a lot less copy-pasting from Word into Iterable.
00:35:45.000 — 00:36:24.440 · Jacqueline Freedman
That makes sense. So if we think about 2030 and the numerous digital transformations we have to think about and actually prepare and implement, not just talk about and hope about and dream about, do you think an effective way to frame this is this is the cost of doing business, even if there's not a very distinctive ROI on this quarter, this year?
How if you were a senior leader, if you're in the C-suite at one of these enterprises, how do you think about this or what would make you think, ah, this makes sense. This is an investment with dividends.
00:36:24.440 — 00:37:50.310 · Tony Byrne
You can show an ROI for a number of this stuff, both with respect to effectiveness, which should drive revenue, a big one. There is going to be speed, like speed of adaptation, which marketers have always wanted. There is an efficiency story, although it may not be quite as profound as we might wish. With respect to AI, there's definitely some business model transformation opportunities.
And then there's issues that probably wash out around risk reduction. Those are typically the four things that people look at from a business case perspective. But there are some things that are going to be more cost of doing business and cost of doing business is a tough one, because CFOs don't like that, because everybody says, well, this is a cost of doing business, you know, whereas the CFO is going to say, well, email is a cost of doing business, but I'm not sure about, you know, taxonomy work.
Right. So I think this is where, again, customer agents are going to come into play and really call the question on that. So if you have thousands or millions of customer agents coming in and your sense is you're not able to address those, then this becomes a cost of doing business the same way that people started emailing you.
People started visiting your website. There wasn't necessarily an ROI on that. That was a cost of doing business. And I think that restructuring your stack to deal with agents, which also restructures it, to deal with customers more generally, that's going to be a cost of doing business, for sure.
00:37:50.670 — 00:38:16.430 · Jacqueline Freedman
Without a doubt, there's nothing harder to sell than maintenance and hygiene. Whew. It's hard to make that look great when it is literally the cost of doing business. That's right. As an example, I guess to double click on that, do you see potentially efficiencies within headcount as part of the equation or just a potential bonus?
00:38:16.470 — 00:39:25.850 · Tony Byrne
No, I don't. And so this is maybe where we're a little bit more contrarian. I think that we're going to have different roles and people are going to be in different roles, and they're going to be higher value roles. And this has been the story of martech from the very earliest days when you had like the email manager, who then became the campaign manager, who then became the journey orchestrator, the webmaster who then became the content, you know, like.
We've gone from generalized work where you had kind of a generalist doing some stuff to specialists who were really driving a lot of value. And I think this is just going to continue. So you're going to get more out of your employees, but it's not clear you're going to need less of them. So it's going to be a lot more value that's being driven in a channel that's going to be increasingly important to you.
And I'm not convinced that there's necessarily headcount reductions. And by the way, this story is being told over and over again in a lot of enterprise departments and not just martech. There are some efficiency stories here, for sure, and but a lot of that value is going to be recycled into engaging the customer better in what's going to be an increasingly competitive environment where you're going to have a lot of lower end, AI fueled companies nipping at your heels.
00:39:25.890 — 00:39:44.980 · Jacqueline Freedman
That makes me very happy, because nothing makes me more mad than "we'll need fewer folks in martech." And this is not even a bias. The more complex, the more individuals you need to make it simple. That's so true. And it's so hard. Making things simple is the hardest part.
00:39:45.020 — 00:40:07.220 · Tony Byrne
This famous quote, who justified the enormous amount of money that they were spending on the back end. I mean, I think they spent like close to $1 billion just on their metadata regime, which is considered like the most sophisticated in the world, what the executives said. I may be misquoting it, but something to the effect of we eat complexity on the back end so our customers don't have to deal with it on the front end, and that's the whole ballgame right there.
00:40:07.420 — 00:40:13.660 · Jacqueline Freedman
I wish that was more commonplace, but also not everyone has $1 billion to spend on it.
00:40:14.060 — 00:40:14.980 · Tony Byrne
Well, that's true too.
00:40:16.020 — 00:40:18.260 · Tony Byrne
But Netflix sets the bar whether we like it or not.
00:40:18.300 — 00:40:31.520 · Jacqueline Freedman
Okay, well, as we wrap, I want to do a premortem. So it is January 1st, 2030 and the book got something wrong? What was it?
00:40:31.560 — 00:42:14.610 · Tony Byrne
Yeah, I can tell you that right away. So, you know, we have a chapter on content. It's one of the longest chapters, and Jarrod has built it out into this whole book. And one of the themes that we get on is this notion of a content warehouse, which is kind of an analog to the data warehouse, where you have your best data and it's best organized and best curated, and you could leverage that.
And I think what's becoming very, very clear to me, even since the time I started writing this book, which was late last year, that the content warehouse isn't likely to exist, and if it does exist, it's going to be in a virtual world. And so what we're talking about now is more of a content orchestra. So on January 1st, 2027, not necessarily 2030, I could say, look, if this is we're going to change the way we describe this.
And it's important because labels matter. Now, all of the rest of the analysis, and certainly the recommendation tips at the end of that chapter all still hold true. But I think content warehouse as a term. Even if you say it's a virtual warehouse, it's a distributed warehouse, the federated warehouse.
It's not really the right term. You definitely don't want a content lake. I'll tell you that. Speaking about content, I think if there's going to be a noun and it's going to describe what it looks like, I think we're really thinking about a content orchestra. And then you need to orchestrate where the right content, the best content.
Now your DAM, or even more advanced, what we call an omnichannel content platform might be the most important part of that orchestra, right? It could be like the entire woodwind section. And so we see many organizations kind of centering their content pillar on what we call an omnichannel content platform, which is an advanced 3.0 DAM.
But it's not 100% going to be your content pillar.
00:42:14.650 — 00:42:24.050 · Jacqueline Freedman
That makes sense. Well, I will be sending you a note January 1st, 2030, because I want to know if the prediction of the premortem came true.
00:42:24.930 — 00:42:25.610 · Tony Byrne
Okay.
00:42:27.110 — 00:42:34.150 · Jacqueline Freedman
Well, Tony, thank you so much for so much of your time, and thank you for putting a date on the future.
00:42:34.190 — 00:43:01.710 · Tony Byrne
Thank you, Jacqueline, for having me on board. I'm a big admirer of yours. I think you're playing a really positive role in this ecosystem, particularly around keeping everybody honest. And in the marketing world that's so critical. Marketers are aspirational and we get excited about stuff and that we shouldn't not be excited, but we also have to keep it real if we want to be effective.
Happy to see everything that you're doing. So awesome to spend some time with you.
00:43:01.870 — 00:43:07.910 · Jacqueline Freedman
Right back at you. And before I let you go, who is someone we should have on the podcast?
00:43:08.110 — 00:43:10.590 · Tony Byrne
Yeah. So I made a short list here.
00:43:10.830 — 00:43:13.110 · Jacqueline Freedman
Sure. I love how prepared.
00:43:13.110 — 00:44:00.320 · Tony Byrne
You know, the first one is, my former business partner, Theresa Regli, who is off now on her own, doing all kinds of interesting things. She's at Vox Veritas, and she can speak a lot to, content governance and and also kind of what does this mean with respect to the future and our climate and how we should be connecting these things together, which is really interesting.
The other person speaking of metadata and taxonomies is Stephanie Lemieux of Dovecot Studio, who is, I think, the world's leading expert on the semantic layer and what that means for AI and what that means for the context layer. And what's great about it is that she's really in the trenches working with companies.
You know, sometimes you can get really abstract data, sciency type perspectives.
00:44:00.360 — 00:44:01.720 · Jacqueline Freedman
She's a little too esoteric.
00:44:01.960 — 00:44:54.900 · Tony Byrne
Yeah. Can get esoteric. And she's so smart and keeps it so real. the other two are people that, you know, that we've worked with over the years. Danielle Williams, who heads up martech for Hilton in London and has an unusually broad purview across digital paid, owned and operated and everything else.
That scope of work has given her a lot of wisdom. The other one who came to mind is Aurora Minter-Cleaver, who's at Santander Bank, who you see on probably many street corners. She heads up their North American martech. But it's a Spanish company and she's done. She's worked in startups, born digital brick and mortar and now finance, and has such a really interesting perspective on all of this, including the all important kind of analytics dimension of this and how hard that is and what you have to do to get good at it.
So, yeah, I've given you a whole roster for next spring.
00:44:54.940 — 00:44:59.540 · Jacqueline Freedman
Yeah, I will gladly take introductions. Thank you so much.
00:45:01.540 — 00:46:17.270 · Jacqueline Freedman
Before we wrap, here's what I want you to walk away with. The context layer of data, content, and decisioning can't be outsourced to a vendor or agency, and no plug-and-play version exists, which means the build in MarTech 4.0 is assembly work, like APIs, MCPs, and integration glue, not writing heavy software. Also, enterprises have handed agencies not just production, but infrastructure and knowledge.
And the AI learning loop can't be a black box, which also means agencies should work in your platforms and the definitions and know how behind your data have to live in-house. Otherwise, one departure can stall analytics for months. Lastly, production roles are going away, but the work is shifting, not the headcount, which means your reskilling plan should be channel-agnostic: journey orchestration, marketing data ops, and judgment before, in, on and after the AI loop.
Thanks for tuning in to Making Sense of Martech. See you next time! A special thank you to Christine Murtagh, who edited this episode. And an extra special thank you to Jenna Carter for believing in this passion project-meets-business. Stay curious.