Justin Sonnenreich

Blog

  • The Kylie Cosmetics UGC Pipeline: What We Built and What It Proves

    When people ask whether a young agency can work with big brands, I point at one logo: Kylie Cosmetics.

    The work

    MarkitAds built a user-generated content pipeline: sourcing creators, briefing them against brand guidelines, managing rights and usage, and delivering a steady flow of authentic content the brand could deploy across paid and organic. The full case study is at markitads.com/our-work/kylie-cosmetics.

    Why UGC and campus are the same muscle

    A UGC pipeline is a distribution system for authenticity: find real people, make it easy for them to say true things on camera, handle the logistics so quality stays high. That is exactly what a campus ambassador program is, the difference is where the content lands. The student who films a 30-second dorm-room review is the same student who staffs a tabling day; the brand that learns to work with one learns to work with both. It’s why our campus offer includes a content layer by default: every activation is also a shoot.

    What the Kylie engagement proves isn’t that we can send emails to famous brands. It’s that our systems survive contact with a marketing organization with brand safety, usage rights and deadlines to enforce, at a scale where sloppiness gets you fired. Not bad for a company started in a bedroom.

    Sources & mentions

  • What a $5 Verified Student Account Actually Means: The CPA Math

    Let me show you the math that makes campus activations a different asset class from ads, using only public numbers.

    The public benchmarks

    • Robinhood’s own affiliate program pays $5 per qualified lead and $20 per funded account, publicly listed, uncapped.
    • Fintech referral programs routinely pay users $75–150 per funded account (SoFi) and ~$100 two-sided bonuses (Chime): check their current program pages.
    • Average e-commerce customer acquisition cost runs $68–84 and paid CPMs rose roughly 20% last year across major platforms (estimate, source pending).

    Now hold those against what a trained student operator produces at a tabling activation: verified, KYC-passed accounts, acquired in person, at single-digit to low-double-digit dollars per account depending on the action. We’ve delivered funded investing accounts for a top-5 US investing app at around $5 per verified account on campus. (Client name and full data shared in calls with permission, we don’t publish what we can’t attribute.)

    Why campus CPA beats the feed

    1. No fraud tax. A signup that happens in front of a person at a table is harder to fake than a click. 2. No auction inflation. Your competitor can outbid your CPM tomorrow; they can’t outbid a relationship with the operator who runs your campus. 3. Lifetime-value timing. You acquire customers in the exact week they form habits, first bank account, first brokerage, first delivery subscription. The switch costs later are your moat.

    The honest caveat: campus doesn’t scale like a budget slider. It scales one campus and one operator at a time, which is why we built the operator model.

    .

    Sources & mentions

  • Inside the UNC Chapel Hill Campaign: 4–5 Intro Calls per Student, per Week

    Most of our campus work is brands trying to reach students. This one was the other direction: the University of North Carolina at Chapel Hill, my own school, using MarkitAds‘ outreach system, MarkitOutreach, to connect its people with the outside world.

    The setup

    MarkitOutreach is our done-for-you outreach engine: list building, personalization, sending infrastructure, reply handling and booking, run as a managed service. For the UNC engagement, the goal was consistent, qualified introductory conversations, the kind that turn into partnerships, placements and pipeline, without the students doing the grinding themselves.

    The result

    4–5 introductory calls per participating student, per week, sustained across the engagement, a pipeline most professionals would be happy with, generated for students. The full write-up lives on the MarkitAds site: markitads.com/our-work/unc.

    Why this case matters more than its size

    First: the client is a university, the institution whose trust every campus marketer claims to deserve. Working for UNC, inside the rules, is a different credential than working around a campus. Second: it proves the general thesis I keep coming back to, students are an undervalued channel in both directions. Brands undervalue reaching them; institutions undervalue what systematic outreach can do for them.

    And it’s the reason our campus activation work (the pilot) starts from respect for the campus: we ran outreach for a university before we ever asked one to tolerate us.

    Sources & mentions

  • How BlitzMetrics Built a Verified Positive-Mentions Tracker for My Brand, and Refused to Publish What It Couldn’t Verify

    Earlier this year, Dennis Yu‘s team at BlitzMetrics did something unusual with my personal brand: they built a public tracker of verified positive mentions, every good thing the internet says about me and MarkitAds, each one checked against a live source before it was allowed on the page.

    The most valuable part wasn’t what they published. It was what they refused to publish.

    The rule: if the agent can’t verify it, it doesn’t ship

    BlitzMetrics runs this process with AI agents trained on their verification standard. Claims that circulate about a founder get sorted into verified, reported, and not found. A few claims about my own background, things that have been said in rooms I’ve been in, went into the “not found” bucket because no public document backs them. So they were flagged as unverified instead of being repeated. Then I understood: every claim that survives the filter becomes more believable precisely because the filter exists.

    In 2026, your buyers run diligence with AI. A growth lead who’s considering a campus pilot doesn’t just Google you anymore, they ask ChatGPT and Perplexity, and those engines are ruthless about sourcing. A brand built on checkable claims compounds. A brand built on vibes gets caught.

    What this means for how I operate

    This website follows the same standard. My exits are stated as my own account and sourced to my LinkedIn profile until acquirer confirmations are linked. Campaign numbers only appear with client permission, or anonymized. There are no review stars here because I don’t yet have a verifiable review corpus. When we publish the campus CPA math, every benchmark links to a public source.

    It’s the same reason MarkitAds prices campus work per verified account: the claim and the invoice are the same number. Reputation that can survive an audit is the only kind worth building, everything else is borrowed time.

    Sources & mentions