The CMO Split: What Happens When Brand and Commerce Stop Agreeing
For the last 30 years, brand discovery in CPG followed a familiar shape: shelf and search. You built brand equity through above-the-line marketing, fought for shelf position in retail, and optimized for keyword search online. Marketing organizations, agency structures, and CMO job descriptions were all built around that shape.
That world is being replaced faster than most organizations are willing to admit.
Recommendation engines now decide what a consumer sees before the consumer decides what they want. Retail media has moved from a line item to a strategic budget. TikTok Shop and other social commerce surfaces have collapsed the discovery-to-purchase journey into a single scroll. And with LLMs and AI agents entering the picture, we're approaching a world where the buyer isn't always the consumer — sometimes it's a machine acting on the consumer's behalf.
That shift is pulling marketing leadership in two directions at once. On one side: brand equity, clarity, and distinctiveness matter more than ever, because when an algorithm is doing the deciding, the strength of your brand signal is what gets you picked up. On the other side: the real value creation is happening in commercial execution — retail media optimization, content velocity, marketplace mechanics — at a level of granularity traditional CMOs were never trained for.
Most CPG organizations haven't figured out how to bridge the two. And the executives caught in the middle — CMOs, CCOs, CGOs — are the ones asking the hardest questions in my search conversations: Do I need to split my marketing organization? Do I need a Chief Commercial Officer alongside a Chief Brand Officer? Am I hiring the wrong CMO profile entirely?
I recently brought this tension to an LS Elevate roundtable with two people who live it from opposite sides: Heike Linnemann, a brand and marketing consultant who spent over 20 years in consumer goods — including at P&G and as CMO of Kiko Milano — and now advises founders and CEOs on brand-led business turnarounds; and Bas van Kesteren, EVP and Global Head of eCommerce at Opella, Sanofi's newly carved-out consumer healthcare business, whose career spans Heineken, Ahold, InBev, and Reckitt.
Here's what came out of that conversation
SEO to GEO — the mechanics change, the discipline doesn't
Bas framed the shift in AI search as a reframe of something marketers already know: SEO to GEO — generative engine optimization. For two decades, brands optimized to be found by search engines. Now the question is where ChatGPT, Gemini, or Claude actually go to gather their information — and how a brand earns its way into that answer.
His read on what that takes in practice:
Know your sources. Depending on the market, the engines lean most heavily on e-commerce marketplaces or high-traffic forums like Reddit. Figure out which, and show up there deliberately.
Write for humans, not keywords. The more complex the customer journey — the more research, comparison, and conversation involved before a purchase — the more the engines reward real language: actual sentences, real Q&As, not keyword density.
Volume and depth of reviews matter more than they used to. Product pages now need to work for two audiences at once: the human converting, and the model scraping for an answer.
Heike’s addition: brands need to start with the end in mind. The goal isn’t just to be tagged correctly — it’s to generate enough authentic trust and conversation that the model has a reason to recommend you. "It cannot just be about the short term," she said. "It cannot just be about the tagging." The fundamentals of standing out, differentiating, and offering real value haven’t changed — they’re just being read by a different judge now.
Can brand and performance both be right?
This is where the panel got interesting. Heike argued that clarity, distinctiveness, and consistency matter more than ever in an algorithmic world. Bas argued that human involvement in buying media and iterating content is going to drop sharply over the next two to three years, as the cost of producing infinite creative variants approaches zero.
Both are true — and the tension between them is the whole point.
Bas's view: the technology is often ahead of human acceptance. AI-generated creative is already doing the heavy lifting on execution — localizing ideas, producing volume, testing at a scale no human team could match. But the strategic layer — the big idea, what a brand stands for — is still, for now, a human judgment call. The skill isn't resisting the shift; it's knowing which parts of the value chain still return better ROI with a person in the loop, and moving talent there as the rest gets automated.
Heike's counter, which is less a disagreement than a guardrail: you can test as many asset variants as you want, but if the underlying strategy isn't right, you're optimizing for the short term at the expense of the long term. Without clear, iconic, well-defined brand assets, neither the machine nor the human consumer will recognize you — no amount of testing fixes a strategy problem.
One role or two?
On whether the CMO function needs to split, Heike was direct: brand and marketing are tools for one outcome — profitable sales growth. A brand lead responsible only for positioning, or only for communication, “doesn’t get it” in her view. The same goes for a growth function operating in parallel with an unclear relationship to marketing. Functional expertise in sales is real and valuable — but ownership of the P&L and the budget can't sit in two places at once. Split ownership, she argued, is “carnage for the organization.”
Bas agreed — with a caveat. As the scope of the role grows, the organization may genuinely need multiple people to lead different pieces of it. What can't be split is the thing they're optimizing for. Without a shared attribution model and shared guardrails, you end up with brand, performance, and sales all separately maximizing their own slice — and the sum of the parts ends up smaller than it should be.
Why most measurement models are already broken
Bas didn't hold back here: many marketing mix models in use today are, in his words, effectively decades out of date. They were built for a media landscape that no longer exists, and they're still being trusted by teams with tens of millions in budget to allocate. His view is that the companies that win from here are the ones building attribution models that are channel-agnostic from the start — reading a wide range of input signals rather than forcing everything through a single legacy proxy, and weighting long-term return alongside short-term lift.
Heike's experience tracks with this from a different angle. At P&G, MMM was foundational. By the time she was at Kiko Milano, measurement was already harder — and in her current work with DTC and scale-up brands, MMM often isn't used at all. Her read: that's not incompetence, it's that the model doesn't fit the business anymore. And notably, a growing share of category growth is coming from smaller challenger brands for whom traditional measurement was never built to begin with.
The parameter that decides who wins
Asked what determines which brand an AI shopping agent picks three years from now, Bas's answer circled back to something Google solved over a decade ago: relevance. The mechanics have gotten more complex since — more data sources, more signals — but the underlying question a model is answering hasn't changed: what's the most relevant response to what this person is actually asking?
The brands that win will be the ones that study what consumers are genuinely looking for, and build the content, reviews, and language that answer it — not for a search box, but for a system trained to recognize authentic, well-evidenced relevance.
The through-line
Strip away the AI framing and this is a conversation CPG leaders have been having for years, just with higher stakes and a shorter runway. Strategy still has to come first. Brand trust still has to be earned, not tagged. And the organizational question — one leader or several — matters less than whether everyone under that structure is optimizing for the same outcome.
What's changed is the cost of getting it wrong. The organizations still running a 2019 playbook won't have the luxury of catching up slowly.
If you're looking to make sure you have the right people in the right seats — or the right team to achieve your goals, including navigating today's AI and technology shifts — get in touch. Reach out at lauren@ls-international.com
By Lauren Stiebing, CEO at LS International
