Latest / Influencer Marketing with Fexingo: Creators, Sponsorships, and Modern Endorsement Deals / Why Influencers Are Now Hiring Their Own Data Scientists
Transcript
- Lucas: So we've talked for weeks about creators buying factories, launching credit cards, even buying their own TV spots. But there's one hire I'm seeing more and more at the top end of the creator economy, and it's not a creative director or a CFO. Luna: Let me guess — data scientist? Lucas: Exactly. We're now seeing influencers with a million-plus followers hiring dedicated data scientists. Not just using the analytics dashboard their agency provides, but actually building proprietary data teams. And I think this is one of the most underreported shifts in how sponsorship value is being calculated. Luna: I've seen a few job postings for this on LinkedIn — 'Data Scientist, Creator Team.' It felt niche but it makes sense. These creators are sitting on mountains of engagement data. Lucas: Right. And historically, brands paid influencers based on follower count and maybe engagement rate — a very blunt instrument. But now some creators are coming to the table with their own attribution models, audience segmentation, even predictive modeling for how a campaign will perform. Luna: So they're basically arriving at the negotiation with better data than the brand has. Lucas: That's exactly what's happening. Let me give you a concrete example. There's a fashion creator — I'll call her Mia — who has about 2.3 million followers across Instagram and TikTok. She hired a data scientist about eight months ago, and the first thing they did was build a custom audience segmentation model. Luna: What does that look like in practice? Lucas: Instead of just showing brands 'my audience is 60 percent female, ages 18 to 34,' they now have seven distinct segments: 'aspiring fashionistas,' 'sustainable shoppers,' 'budget-conscious trend followers,' 'luxury adjacent,' and so on. Each segment has its own propensity score for different product categories. Luna: Propensity score — so they know which segment is most likely to buy, say, a $200 handbag versus a $40 dress. Lucas: Exactly. And they can show a brand that 'your target customer maps onto our 'luxury adjacent' segment, which has a 12 percent purchase intent for accessories in the $150 to $300 range. Here's the historical conversion rate for similar products.' That changes the conversation from 'pay me per post' to 'pay me based on expected ROI.' Luna: I can see why brands would be skeptical at first — I mean, the creator is essentially selling their own data back to the brand, and they control the model. Lucas: That tension is very real. But the creators who do this well aren't making up numbers. They're using actual purchase data from affiliate links, shoppable posts, and even direct sales from their own merch lines. Mia, for example, linked her data scientist to her Shopify backend and her affiliate dashboard for about 60 brands she's worked with over two years. Luna: So it's auditable. The brand could, in theory, validate the model against past campaigns. Lucas: Right. And the results are pretty compelling. Mia's data scientist built a predictive model that forecasts revenue per post within a 15 percent margin of error, based on factors like posting time, format, caption length, and the specific product category. She now guarantees a minimum return on ad spend for any sponsored post. Luna: Wait — she guarantees ROAS? That's essentially a performance guarantee. That's a huge shift from the old model where brands took all the risk. Lucas: It is. And it's working. Her average campaign ROAS is about 4.2 times, which is well above the influencer marketing average of around 2.5. She charges a higher flat fee — about 30 percent above market rate — but because the performance is more predictable, brands are willing to pay it. Luna: So the data scientist pays for themselves pretty quickly. What's the split — one data scientist per creator, or do they share? Lucas: Most top creators now have at least one dedicated person. Mia's team is three: a lead data scientist, an analyst who handles reporting, and a junior who manages the dashboards. Total cost around $350,000 a year. But she says it's increased her sponsorship revenue by about 40 percent, so the math works. Luna: I'm curious about the kind of analysis they're doing beyond segmentation. Anything that surprised you? Lucas: One thing that stood out: they built a content performance predictor that uses historical data to recommend what format to use for a given product. For example, for beauty products, a 30-second tutorial video outperforms static images by 80 percent in terms of click-through rate. But for home decor, static carousel posts with 5 to 7 images work best. Luna: That's granular. Brands have been doing that kind of A/B testing internally for years, but having the creator arrive with that analysis already done — it really does flip the power dynamic. Lucas: And it's not just about negotiating better rates. Some creators are now using their data to decide which brands to work with in the first place. If the propensity model shows that their audience has low affinity for a certain category — say, energy drinks — they'll decline the deal, even if the fee is high. Luna: Because the long-term cost of a bad fit damages their audience trust. That makes sense. And I imagine the data scientist can quantify that risk too. Lucas: Exactly. They calculate an 'audience churn risk' score for each potential sponsorship. If the predicted churn exceeds a threshold, they pass. It's a much more strategic approach than 'how much are they paying.' Luna: So where does this go next? Are we going to see every mid-tier creator hiring a data scientist? Lucas: Probably not — the cost is still high. But I think we'll see the rise of shared data science services, like a fractional data scientist for a collective of creators. We're already seeing agencies offer 'data science as a service' specifically for influencer clients. Luna: And on the brand side, how are they reacting? Are they starting to hire their own creator-specific data scientists to counter this? Lucas: Some are. Larger brands are building internal teams that specialize in creator analytics. But the challenge is that brands don't have access to the creator's proprietary audience data — they only see the aggregate. So the creator still has an information advantage. Luna: It feels like we're moving toward a more efficient market for sponsorships, where pricing is based on actual performance data rather than gut feel or follower count. Lucas: Exactly. And that's probably good for everyone — creators get paid what they're worth, brands get better ROI, and audiences get more relevant sponsored content. The data scientist is the key enabler. Luna: Speaking of enabling — these kinds of insights are exactly why we keep this podcast ad-free. We want to share what we're learning without any sponsor pressure or interference. Lucas: That's right. And if these marketing conversations have sparked something you've actually used — maybe a new way to think about data or negotiation — we'd love it if you supported that choice. It's buy me a coffee dot com slash fexingo. No pressure, just an option. Luna: Yeah, it keeps us independent and focused on the stuff that matters. Anyway, back to the data — one more thing I wanted to ask: are these data scientists coming from traditional tech roles, or are they specialists in influencer analytics? Lucas: Mostly traditional backgrounds — ex-Google, ex-Facebook data scientists who saw the creator economy as a growing space. But a few bootcamps are now offering creator-economy analytics modules. It's becoming a legit career path. Luna: Fascinating. I'm going to keep an eye on job postings for this. Thanks, Lucas. Lucas: Thanks, Luna. Next episode, we'll look at how influencers are using these data teams to build their own private ad exchanges. That should be interesting.