Latest / Influencer Marketing with Fexingo: Creators, Sponsorships, and Modern Endorsement Deals / How Influencer Campaigns Secretly Use Dynamic Creative Optimization
Transcript
- Lucas: You scroll past an ad on Instagram — it's your favorite beauty influencer, let's call her Mia, holding up a serum. The headline says 'Your dry skin needs this.' You keep scrolling. Your friend, five minutes later, sees the exact same influencer, same serum, but the headline says 'Oily skin? Same.' And the call to action button is different colors. Luna: Wait — that's not a coincidence, is it? That's dynamic creative optimization. Lucas: Exactly. DCO in influencer marketing. It's been quietly rolling out over the past eighteen months, and it's kind of a big deal. The brand didn't shoot two separate ads. They shot one video, but the platform's ad engine — Google's or Meta's or TikTok's — swaps in different headlines, different CTAs, sometimes different product shots, based on what it knows about you. Luna: So it's like programmatic creative on steroids. But instead of a generic stock photo, you're using a real person's face and voice. Lucas: Right. That's the twist. I looked at a recent campaign from a mid-size beauty brand called Gloss & Co — not a household name, but they spent about $400,000 on influencer DCO last quarter. They worked with one creator, shot a single two-minute video, and the system generated 2,800 unique variations. Different opening lines, different skin concern hooks, even different background music tempos. Luna: Two thousand eight hundred variations from one video. That's insane efficiency. But does the influencer know their face is being chopped and remixed like that? Lucas: That's the question. Most standard influencer contracts — the ones I've seen — give the brand 'usage rights to the content.' That phrase is broad enough to cover DCO. But a lot of creators I've talked to didn't realize their likeness could be algorithmically recombined until they saw reports from the ad platform. Some are pushing back. Luna: I can imagine. There's a difference between licensing a photo for a static post and having your facial expression paired with a headline you never read. It's a consent issue. Lucas: Yeah. And it's not just about consent — it's about brand safety for the creator. What if the algorithm pairs their face with a headline that's misleading or makes a medical claim? The influencer takes the reputational hit, not the algorithm. Luna: Have any major creators spoken out about this publicly? Lucas: A few. There was a notable thread on X last month from a lifestyle creator with about 800,000 followers. She posted screenshots from her Meta ad library showing forty-seven different versions of her face in ads she didn't approve individually. The thread got a lot of traction. Some agencies are now adding clauses that require pre-approval of all DCO variations. Luna: Which is a logistical nightmare. If you have thousands of variations, how do you approve each one? Some agencies are using automated pre-screening tools that flag problematic combos. Lucas: Right. It's a negotiation point. The brand wants speed and scale; the creator wants control. And honestly, if these conversations have sparked something you've actually used in your own work — a tactic, a new way of thinking — that's the kind of thing that keeps this show ad-free. If today was worth a coffee to you, that's the link: buy me a coffee dot com slash fexingo. Luna: Yeah, it's a small gesture that makes a big difference for us. And we don't run ads, so listener support is what keeps the lights on. Lucas: Okay, back to DCO. Let's talk about the tech stack. The core of DCO is a decision engine that looks at real-time signals — user location, device type, browsing history, time of day — and assembles the ad on the fly from a library of creative assets. Luna: So for the influencer, that means they might deliver a line like 'My go-to for morning puffiness' to someone in New York at 7 AM, and 'Perfect for pre-party glow' to someone in LA at 10 PM. Lucas: Exactly. And the platform learns which combos perform best. Over the course of a campaign, the algorithm might find that a specific smiling shot of the influencer paired with a discount CTA converts twice as well as a serious expression with a 'learn more' CTA. So it shifts more budget to that combination. Luna: That's smart for the brand, but it also means the influencer's performance is being micro-optimized. Their face becomes a variable in an A/B test. Lucas: And some creators are fine with that — if they're paid well and the brand is transparent. Others feel it commodifies their image in a way they didn't sign up for. Luna: Let's talk numbers. What does this mean for campaign performance? Are brands seeing real lifts? Lucas: The data I've seen from a few ad-tech vendors is pretty striking. One case study from a brand in the supplements space — not naming names — claimed a 35 percent increase in click-through rate and a 20 percent reduction in cost per acquisition when they switched from static influencer posts to dco powered ads. The influencer themselves saw a 15 percent bump in engagement on their own page because the ads were more relevant. Luna: But there's a flip side. If every ad is tailored to me, do I start feeling like the algorithm knows me too well? There's a creepiness factor. Lucas: There is. And that's where the ethical line gets blurry. You're basically using a trusted person's face to sell something, but the message is being tuned to exploit your specific insecurities or desires. The FTC has started looking at this. In their latest endorsement guidelines update — December 2025 — they added a section on 'algorithmically personalized endorsements.' They say if the message is materially different from what the influencer approved, it could be considered a new endorsement requiring separate disclosure. Luna: So the legal framework is playing catch-up. Which is typical for ad tech. Lucas: Typical, but important. Because right now, most DCO campaigns operate in a gray area. The influencer approved a general concept, but the algorithm makes the final creative decisions. That's a lot of trust to place in a black box. Luna: Are there any tools that let influencers see exactly which versions of their ads are running? Something like a dashboard? Lucas: A few platforms are building that. One startup, CreativeWatch, offers a 'DCO audit' feature that aggregates all variations featuring a specific creator and shows them in a gallery. The creator can then flag any they're uncomfortable with. But it's not widespread yet. Most brands still treat the ad library as their own data. Luna: I wonder if we'll see a new kind of influencer agent emerge — someone who specializes in negotiating DCO rights. A tech-savvy lawyer who understands creative optimization. Lucas: That's already happening. I spoke to a talent manager in LA who now has a clause template specifically for DCO. It limits the number of variations to 500, requires a weekly report, and gives the creator veto power over any combination that gets a low engagement score or negative sentiment. It's a new layer of contract complexity. Luna: Five hundred variations still feels like a lot to review. But I guess it's better than two thousand eight hundred. Lucas: Right. And for smaller creators, they may not have the leverage to demand that. So they either accept the standard terms or miss out on the campaign. That power imbalance is something the industry is going to have to address. Luna: What about the platforms themselves? Are Meta and TikTok encouraging DCO? Lucas: Absolutely. Both have built DCO tools directly into their ad managers. Meta's 'Dynamic Creative' has been around since 2017 for generic ads, but the influencer-specific version — where you upload a creator's video and tag their handle — rolled out in early 2025. TikTok's 'Creative Optimization' tool followed a few months later. They make it easy for brands to enable it with one checkbox. Luna: One checkbox. That's terrifying. A brand manager could check a box and suddenly their influencer's face is being tested against a thousand headlines without anyone reading them. Lucas: Exactly the concern. And because the platform handles the optimization, the brand doesn't necessarily know which combos ran. They just see aggregate performance data. So there's a lack of transparency even for the brand. Luna: So we've got a triangle of opacity: the brand doesn't know exactly what ran, the influencer doesn't know where their face appeared, and the consumer doesn't know the ad was assembled just for them. Lucas: That's a perfect way to put it. And that triangle is where regulators are starting to poke holes. The EU's Digital Services Act already requires platforms to label personalized ads, but it doesn't specifically address the creative assembly aspect. I think we'll see updated guidance within the next year. Luna: Let's end with a forward look. Where do you see DCO in influencer marketing going by, say, end of 2027? Lucas: I think we'll see two trends. One: more creators will demand dco specific contracts with audit rights and veto power. Two: we'll see the rise of synthetic influencer DCO — where the influencer's likeness is fully ai generated from a model, not from a video shoot. Some startups are already offering 'digital twin' services. If that takes off, the consent questions become even more complicated. Luna: A digital twin that can be dynamically optimized across millions of variations, without the real person even knowing. That's a whole new frontier. Lucas: It is. And it's coming faster than most people realize. For now, the key takeaway is this: if you're an influencer, read your usage rights carefully. And if you're a brand, be transparent. The trust between creator and audience is fragile — DCO can either preserve it or break it. Luna: Well said. That's all for this episode of Influencer Marketing with Fexingo. Lucas: Thanks for listening. We'll be back soon with another angle.