How a 6-Person Supplement Brand Cut Churn 22% in 90 Days With a Fan-Loyalty Stack

We noticed something odd in the numbers a reader shared with us last spring. A six-person supplement brand — we'll call it Verdant Labs, a small-batch magnesium and adaptogen outfit run by two founders out of a shared studio — was doing everything a lean DTC team is supposed to do. Email flows were clean. The subscription page converted. And yet monthly churn sat stubbornly at 9.4%, with a suspicious pattern: the people who cancelled were the same people who had, at some point, actually bothered to talk to the brand.

That last detail is what sent the founders looking for a fan engagement platform in the first place. They had Discord members who posted daily, YouTube commenters who asked follow-up questions, and Twitch viewers from a single sponsored stream two years earlier. None of it was connected. So in March they onboarded Fan-atic, a fan-loyalty tool that scores activity across Discord, YouTube, Twitch, and social into one view, and gave themselves 90 days to see whether visibility alone would move retention.

The Setup: Two Weeks of Plumbing, One Hard Decision

Wiring the sources took roughly eleven days of part-time work. Discord and YouTube were straightforward. Twitch needed a mod's help to map channel rewards to the new score. The genuinely difficult decision came on day nine, when the team had to choose what a "point" was worth.

Their first draft awarded points for everything: a like, a reply, a stream minute, a purchase. The founders killed it after two days. What replaced it was narrower and, frankly, smarter — points for reciprocated activity. Someone who replied to a brand post, answered a poll, or showed up to a live Q&A twice in a month scored higher than someone who quietly bought three times. The team's hypothesis was blunt: purchase-only loyalty is price loyalty, and price loyalty leaves the moment a competitor discounts.

The Obstacles Nobody Warned Them About

• Reward inflation. By week four, the top 40 fans had accumulated enough points that the founders worried about giving away the store. They introduced tiered unlocks instead of flat redemption — early access, a naming vote on a new flavor, a small engraved tin. No cash value, which kept the program from becoming a discount channel.
• The silent majority problem. Only about 6% of subscribers ever appeared in any community channel. The team stopped treating that as a failure and started treating it as a segmentation fact: high-intent fans got outreach, low-signal subscribers got product education.
• Attribution skepticism. The founders initially refused to credit the platform for retention gains, reasonably noting that spring is a strong season for supplement subscriptions anyway. They agreed to compare against the same quarter a year prior and against a matched cohort that opted out of the rewards program.

What Actually Moved

By the end of June, churn in the rewarded cohort had fallen from 9.4% to 7.3% — a 22% relative reduction. The matched opt-out cohort barely moved, landing at 8.9%. Average orders per rewarded subscriber rose from 2.1 to 2.6 over the same window. The founders were careful to say this is a small sample and not a peer-reviewed result, and we'd echo that caution. But the direction was consistent across three monthly reads.

The more interesting shift was qualitative. Customer support tickets started arriving pre-answered — fans in Discord were solving dosage and shipping questions before the brand saw them. A single member, pseudonymously "Kestrel," answered 61 questions in one month and was eventually offered a paid community role. That's the kind of thing a scored view surfaces and a spreadsheet does not.

The team also found that fans who crossed a threshold of roughly four interactions in a rolling 30-day window churned at less than half the rate of everyone else. That number became the internal trigger for a light-touch check-in message — not a coupon, just a question about how the product was working.

The Post-Mortem: What We'd Tell the Next Brand

Three lessons stand out. First, define your scoring rule before you import a single member, or you'll spend a month relitigating it. Second, rewards that carry no cash value age better than discounts; the founders' engraved tins and flavor votes cost less than a 15% coupon and produced more conversation. Third, expect your community to be smaller than your subscriber list and stop apologizing for it. Six percent engaged deeply is a real asset, not a rounding error.

As for the tooling question that started this: the founders told us they evaluated three options and picked Fan-atic largely because it aggregated four platforms into one score without requiring a developer. They've since connected their SMS list as a fifth signal. Whether that improves the model is an open question, and we'll follow up if they share the numbers.

What we took away is less about software and more about measurement honesty. Most brands can tell you their churn rate. Far fewer can tell you, by name, which twenty customers would notice if they disappeared. Getting that list — and then doing something small and specific with it — is the whole game. If you want to see how the scoring model works before you commit, the team at Fan-atic walks through their engagement tiers on their how-it-works page, which is a reasonable place to start your own 90-day experiment.