Rethinking Marketing Team Structure in the Age of AI

Episode Transcript

Greetings. If you're a regular listener, you're probably saying, "You're not Andrea." And you're right, I am not. I'm Ross Libby, Chief Value Officer here at AgileSherpas, and you can think of me as another guide on the journey — joining Andrea, along with her colleague Monica, in future episodes.

With that clarified, I have a question. When's the last time you looked at your team or the org chart overall and thought, "Yep. That's exactly the right group of people doing exactly the right work in exactly the right way"? Probably next to never — and next to never just left town.

Why is this? Probably because most of us inherited a team structure. Nobody sat around and designed it, especially not us. It sort of amassed over time. A hire here, a reorg over there, a centralization someone stood behind back in 2019 that never got questioned again.

And for the last several years, the big structural argument in marketing has been: cross-functional team or center of excellence? Embed your people directly with the business, or build a hub of specialists that serves everyone. People have strong opinions on this. I have a few opinions on this.

But here's what I want to argue today: nobody won that fight. And it doesn't matter, because it was never actually a fight you needed to win. Cross-functional and COE were solving for different strengths. And AI just made it possible to get both of them out of the same team at the same time. It's not cross-functional or COE anymore. It's cross-functional and COE — with the powers combined. Cue Captain Planet.

And it also means you need fewer slices of the pizza to get more of the pie. Let's get into it.

---

Welcome to The Agile Marketing Edge, the first podcast dedicated to turning Agile theory into real-world marketing breakthroughs. I'm Ross Libby, Chief Value Officer at AgileSherpas, and this week I'm your guide on this climb to smarter, faster, outcome-driven marketing. Today, we're going deep on the fight that never actually needed a winner — cross-functional teams vs. centers of excellence — and the shifts that let the smartest teams stop choosing sides and start teaming up for the win.

---

Let's name the fight honestly, because both sides have a real point. And that's exactly why nobody won it.

We are squaring up on the "who" and the "how" of the Agile marketing operating system. Because team structure isn't a nice-to-have design question you get to when you have time. It's at the core of the infrastructure. And I think most of us are still approaching it in a way that no longer makes sense — or maybe never really did.

On one hand, there's the cross-functional argument. Embed a team directly with the business unit or product line it serves — writers, designers, analysts, maybe even a strategist — all in the room together, all accountable to the same outcome. The pitch is speed and context. Nobody's waiting on a ticket in someone else's queue. You are in control.

On the other hand, in the other corner, stands the center of excellence argument. Build a hub of specialists that serves the whole organization — one brand team, one analytics team, one content team. The pitch is depth and consistency. You don't reinvent your approach or your measurements five different ways across five different units.

Both of these are legitimate structures for solving real-world problems, which is exactly why this debate never resolves in most organizations. People just keep swinging back and forth between them every few years.

And I want to point to something from this year's State of Agile Marketing Report that shows how high the stakes are on getting this right. We asked marketers how confident they were that their team structure helps them get work done smoothly. Get this: 67% of Agile marketers said "extremely" or "very confident." For their non-Agile counterparts, only 61%. There's a gap there, and it's real. About two in five non-Agile marketers are walking around without much confidence that their structure even works.

But I want us to notice what's actually behind that question — what it's measuring. It's not measuring cross-functional versus COE. It's measuring whether people believe their structure was an intentional decision at all, or something that just happened to them. That's the deeper issue underneath this debate. Most teams haven't chosen a side. They've inherited whatever combination of speed and depth they happen to have. And when it doesn't work, they blame the model instead of asking whether anyone actually designed it for their situation in the first place.

So here's my reframe as we start: cross-functional and COE were never opposites to pick between. They're two different strengths — speed and context on one side, depth and consistency on the other. And for years, the only answer was to trade one for the other. You couldn't have both, not unless you had headcount you couldn't justify. That's what made it feel like a real fight.

Well, that stopped being true. AI — more on that in a moment. But first, who's hungry?

You've probably heard the two-pizza rule — Amazon's old guidance that if a team can't be fed by two pizzas, it's too big. It's become shorthand in a lot of org design conversations, including in marketing, for: keep teams small and nimble. There's real thinking behind it.

Look at Team Topologies, by Matthew Skelton and Manuel Pais. Their work builds a whole framework around team size and team boundaries, and marketing organizations have started borrowing from it. It dives directly into the conversation we're having about what goes into a team and how those dynamics play out.

But here's something worth sitting with: the two-pizza rule was never about how many specialists or skills you need. It was about communication overhead. The math is mechanical. For every person you add to a team, you add more possible communication pathways — not just one more, but one more for every other person already on the team. Small teams don't just move faster because they're small. They move faster because there are fewer lines of communication to maintain, fewer meetings required to keep everyone aligned, fewer places for a decision to get diluted on its way to being made.

So the two-pizza rule answers "how many people can coordinate effectively" — not "how many skills do I need represented?" They're related, but they're not the same question. And marketing org design has been treating them as if they were interchangeable. That's a gap.

Because if the answer to "how many skills do I need in the room" changes — if you suddenly need fewer people to cover the same range of expertise — then the whole calculation of what a right-sized team looks like changes with it. And that's exactly what's happening.

So let's get concrete, because this is not only theoretical.

Let's look at MetLife Stadium — home to two NFL teams and a packed calendar of concerts and major events. MetLife Stadium runs its entire marketing and communications operation with a two-person team. You heard that right: two people managing a subscriber list of nearly 800,000 fans, sending 30 to 40 email campaigns a year across ticket pre-sales, on-sales, and event-day logistics.

And they're not just a marketing function. They partner with HR for onboarding, guest services for post-event surveys, and partnerships for sponsor reporting — multiple departments all connecting through a single platform, with a lean two-person team at the center. Their quarterly fan newsletter averages a 60.7% open rate. Those are exceptional numbers by any standard, but at a list size of 800,000, they're remarkable.

So let's look at what this two-person team actually is, structurally. It's embedded directly with the business alongside it — getting the same speed, context, and accountability you'd get from a cross-functional team. At the same time, they've got the depth and consistency you'd normally only expect from a full center of excellence: design standards, analytical rigor, optimization discipline. Click mapping and send-time analysis that used to require dedicated analysts. Drag-and-drop builders that don't need a designer and developer to run. Built-in AI handling subject line optimization automatically — consistently, on every single send.

That's the combination. Not cross-functional beating out COE. Not COE getting automated away. Two people getting cross-functional speed and COE-level depth out of the same small team, because the system is doing work that used to require choosing one structure or the other — or hiring enough people to have both.

This is why the old fight is over. And why nobody actually won it. The trade-off that made you choose speed or depth, embed or centralize, isn't a fixed trade-off anymore. With powers combined, this two-person team delivers what used to take two entire departments. You don't need the whole pie — because the slice you do have can do double duty, or more.

---

Alright. The team is smaller, leaner, more concentrated. That does not mean it's unstructured. If anything, it means it needs to be more disciplined — not less — because there's less redundancy to absorb a bad decision or a misstep.

I want to ground this in something real, and share a rhythm I've come to rely on. I didn't invent it at a whiteboard. I lived it. I learned it as a young grasshopper going through Target's Dojo — yes, that's a real thing, and not just a martial arts reference. I do have the samurai sticky note holder to show for it.

A dojo is an immersive, hands-on coaching experience. You don't read about agility — you do the work differently for a stretch, with a coach right by your side. I went through it as a team member first, then became one of the coaches guiding other teams through it. And here's the important part: it didn't stay confined to IT. It started there, sure, but over time I found myself coaching more and more business-side teams through the same rhythm — including many marketing teams. Same coaching structure, wildly different day-to-day work. That's usually a good sign you've found something closer to a principle than a mere tactic.

So, the rhythm I coach teams through every single time, regardless of size or department: four questions, in this order.

**First: Why?** Looking at the outcome. Before anything else — what are you actually trying to achieve? Not the activity, not the backlog. The outcome. I've watched teams argue for weeks about tools and structure when nobody in the room could articulate the why in a single sentence. If you can't do that, everything downstream is guesswork dressed up as strategy.

**Second: What?** The people and the skills. Once the outcome is clear, who actually needs to be in the room to deliver it? Map to skills, not just job titles. I coached IT teams where this meant realizing they had three people doing overlapping work, or a critical gap that was missing entirely. I later coached marketing teams through the exact same exercise, and the pattern held: the org chart and the skills map are almost never the same document, and most teams have never actually compared them side by side.

**Third: How?** Ways of working. This is where a lot of teams want to start — and it's exactly why I put it third. How you work: your rituals, your cadence, your ceremonies, your tooling — it only makes sense once you know why you're doing it and who's actually doing it. I've seen teams adopt a way of working wholesale from another team, skip the why and the what, and then wonder why it didn't stick. It didn't stick because it was never theirs to begin with.

**Fourth: When?** Iterative and timely. The final question, and the one teams skip fastest once the excitement of the first three wears off. Are you actually revisiting this? A dojo isn't a one-time event. The whole model is built on the idea that you keep coming back to the mat to get better. Same here. The rhythm isn't set-and-forget. It's a loop you keep running — to see what's working, what's not, and what needs to improve.

I'll be honest about one of the biggest lessons from that transition from coaching IT to business teams: the framework didn't need changing. What changed was how much support people needed on the "why" question specifically. Technical teams were often more comfortable saying out loud, "I don't know why we're doing this." Business teams — marketing teams especially — would sometimes rush past the why and the what to get straight to the how. Because we invented customer centricity. We were doing personas before product was even part of the org chart. The how felt more productive to talk about.

It wasn't. It just felt that way — or maybe it felt safer to avoid looking directly at an organization that was missing clear design and direction for its people.

---

I want to be really clear about something before we close, because I don't want this episode to land as "go fire half your team, AI is here to do the work." No.

What's shrinking is the coverage problem — how many people you need physically present to cover a range of specialized execution. What's not shrinking, not even a little, is judgment, reasoning, and wisdom. Someone still has to decide which outcome matters most this quarter. Someone still has to read a stakeholder relationship and know when to push back versus when to give in. Someone still has to decide what not to automate — because not every decision should be handed to a tool just because that tool can make it efficiently. Efficiently and effectively are not the same thing.

The failure mode I'd actually worry about isn't teams getting too big. It's teams getting small without anyone deciding, on purpose, what the remaining humans are actually there to do. If your two-person team is just as reactive and fire-drill-driven as the 12-person team was, you haven't gotten leaner — you've just gotten more fragile, because now there's no redundancy left when something goes wrong. The slice is smaller. The judgment, reasoning, and wisdom required to run it just got more concentrated, not less important.

---

Here's what I want you to do with this, in order — because the order matters.

**One:** Name the actual outcome your current team structure is organized around. Not what the org chart says — the real one. If you can't name it in one sentence, that's your first problem. And it's not a headcount problem.

**Two:** Map skills against that outcome, not against existing job titles or the org chart. Where are you paying for redundant specialization that a platform or tool could absorb instead? Where are you at risk of being too lean, where the redundancy actually benefits you?

**Three:** Pick one piece of the work — one thing you're currently throwing a full team at — and run it as a one- or few-person experiment this quarter. Not a transformation. An experiment. MVP.

**Four:** Build in a real review point before you scale the change. Proactive doesn't mean set-and-forget. Check whether the outcome, the skills matrix, and the work still match each other.

In the age of AI, your LLM isn't the differentiating factor. Your people still are. And scaling still requires people, process, data, and technology working together for sustained success — unless your idea of a good time is sending a hallucination across a multi-channel campaign.

Until next time, I'm Ross Libby. And remember: the struggle is real, but so is Agile marketing.