Latest / The CMO Podcast with Fexingo: Marketing Leadership, Budgets, and Executive Strategy / CMOs Are Breaking Their Own Brand Campaigns Into Micro-Experiments
Transcript
- Lucas: Luna, I've been watching something happen at the CMO level that sounds almost counterintuitive at first — CMOs are deliberately breaking their own flagship brand campaigns into dozens of tiny, sometimes contradictory micro-experiments before they commit to scale. Luna: Contradictory? That feels like a risky word for a brand steward. Lucas: It is. But the logic is actually pretty disciplined. The old model — one big creative brief, one big production, one big media buy — is dying because it's a binary bet. You either nail it or you don't, and you don't know which until you've spent millions. Luna: So the alternative is to run, what, twenty different versions of the same campaign in parallel? Lucas: Exactly. And sometimes more. Let me give you a concrete anchor — Unilever's Dove. In Q1 of this year, 2026, the Dove team ran 47 distinct creative variants across 14 markets as part of the ongoing 'Real Beauty' franchise refresh. Luna: 47 variants. That's not a campaign — that's a laboratory. Lucas: That's exactly the mindset. Each variant had its own media mix — some were purely social-first, some included connected TV, some were built around influencer seeding versus paid search. They even varied the call to action: some pushed to a brand site, some to a retail partner, some to a user-generated content prompt. Luna: And the brand consistency piece — how do you maintain a coherent Dove identity when you're testing fourteen different messages in fourteen countries? Doesn't that risk confusing the consumer? Lucas: That is the central tension. And the CMOs I've spoken to — including Dove's global marketing lead — say the key is that the experiments are not about changing the core brand promise. They're testing execution variables: tone, imagery, media context, audience segment, offer structure. Luna: So the 'what' stays the same — real beauty, confidence, inclusivity — but the 'how' varies wildly. Lucas: Right. And the data coming out of those 47 variants is already reshaping the next wave of investment. They found, for example, that in Southeast Asian markets, a variant featuring intergenerational family dynamics outperformed the standard individual-portrait version by 34 percent on purchase intent. That's a finding they never would have discovered with a single global campaign. Luna: 34 percent is huge. But that also means 33 other variants probably underperformed. What's the cost of all that 'learning'? Lucas: That's the honest question. The up-front production and media cost is higher — you're essentially funding multiple campaigns instead of one. But the argument is that the total cost of finding the right approach is lower than the cost of launching a single wrong approach at full scale. Luna: So it's insurance against a big miss, plus you get a roadmap for the next twelve months. I can see the logic. Lucas: And it changes the skill set needed on the marketing team. Several CMOs I've talked to are now hiring for a role they call 'experiment designer' — someone who comes from a data science or behavioral economics background, not traditional brand strategy. Luna: An experiment designer. What does that person actually do day to day? Lucas: They design the test matrix — which variables to isolate, what sample sizes are needed for statistical significance, how long to run each variant before declaring a winner. They also build the dashboards that let the CMO see, in near real time, which experiments are generating signal and which are just noise. Luna: That requires a pretty sophisticated data infrastructure. Most marketing orgs I know still struggle with basic attribution. Lucas: Absolutely. And that's the barrier. The CMOs who are succeeding with this approach have invested heavily in a unified measurement layer — usually a combination of a customer data platform and a multi-touch attribution model that can handle the complexity of dozens of concurrent experiments. Luna: So it's not something you can just decide to do next quarter. You need the plumbing first. Lucas: Right. And that plumbing investment is why we're seeing this mostly at large CPG companies — Unilever, P&G, Nestlé — and some dtc native brands that were built on A-B testing culture from day one. Luna: But I wonder if there's a risk of over-optimization. If you keep running micro-experiments, you might end up with a perfectly tuned message for every micro-segment — but no big, shared cultural moment for the brand. Lucas: That's a real concern. Some CMOs call it the 'death by a thousand optimizations' — you improve every metric by two percent, but you lose the brand heat that comes from a single, bold, unified statement. Luna: So how do they balance it? Do they reserve some budget for the big swing? Lucas: The best practice I've seen is a 70-20-10 model. Seventy percent of the campaign budget goes to the proven winner from the experiment phase. Twenty percent goes to a few promising variants that showed signal but need more scale to confirm. And ten percent is reserved for what they call 'blue-sky' — a truly bold, non-experimental creative that the team believes in, even if the data doesn't yet support it. Luna: That 10 percent is the safety valve for creativity. I like that. Lucas: And it's not just creative. The same logic applies to media mix, pricing, even packaging. One beverage brand I know — I won't name them because the data is still confidential — ran 12 different package designs in test markets last year, each with a different shelf placement strategy. They found that a specific matte-finish label with a lower shelf position outperformed the glossy, eye-level version by 18 percent in repeat purchase. Luna: That's fascinating. And it's the kind of granular insight you'd never get from a focus group. Lucas: Exactly. The whole premise is that real-world behavior is more reliable than stated preference. So you let the market vote with its wallet, across dozens of micro-tests, before you decide where to place your big bet. Luna: I want to ask about the organizational change this requires. You mentioned the experiment designer role — but what about the agency relationship? Traditional agencies are built to produce one big campaign, not 47 variants. Lucas: That's a huge friction point. Several CMOs have told me they've restructured their agency contracts to pay for 'creative capacity' rather than 'creative output' — meaning the agency is compensated for the volume of testable assets they produce, not just the one final execution. It changes the incentive structure completely. Luna: So the agency gets paid more for producing more variants, even if many of them never see the light of day at scale. Lucas: Right. And that requires a lot of trust. The agency has to be comfortable with the idea that most of their work will be 'disposable' — it's raw material for the experiment, not the finished product. Luna: That's a big cultural shift. Agencies are proud of their work. Seeing a beautiful campaign run for two weeks in a low-sample-size cell and then get killed because the data didn't support it — that's got to sting. Lucas: It does. But the CMOs who are doing this well say they've built a shared learning culture with their agencies. The agency becomes a partner in discovery, not just a vendor of executions. And the best ones actually get excited about the data — they start proposing their own experiment designs. Luna: That's a healthy dynamic. So where do you see this going in the next year or two? Is micro-experimentation going to become the new normal? Lucas: I think it will for any brand that has the data infrastructure and the organizational appetite for learning over ego. But I also think we'll see a backlash — some CMOs will overcorrect and lose the brand magic. The ones who succeed will be the ones who treat the experiments as a tool, not a religion. Luna: That's a good note to end on. Before we wrap, I just want to say — if these marketing conversations have sparked something you've actually used in your own work, or if you just appreciate that this show stays ad-free and focused, consider supporting us. It's a small gesture that makes a big difference for us. You can find that at buymeacoffee dot com slash fexingo. Lucas: Yeah, and we mean it — no pressure, no perks, just a way for listeners who find value here to keep the lights on. And honestly, it helps us know we're on the right track. Luna: Exactly. So back to the future of brand experimentation — Lucas, what's one thing you think every CMO should do differently starting tomorrow? Lucas: Start with one campaign. Take the next brand initiative you have planned and deliberately run three to five creative variants — not just different media placements, but genuinely different approaches to the message. Measure them against each other, not against a control. And be prepared to throw out your favorite version if the data tells you to. Luna: Hard advice, but good advice. I think that's where we'll leave it for today. Thanks, Lucas. Lucas: Thanks, Luna. See you next time.