Latest / Community Marketing with Fexingo: Discord, Slack Groups, and Brand Communities / How a 75-Person Community Tested 24 Ad Creatives in a Weekend
Transcript
- Lucas: You know that moment before a product launch when the marketing team has twenty-four ad variations — different headlines, different hero shots, different offers — and nobody knows which three to put budget behind? Luna: Oh, I've lived that. You end up running Facebook tests for two weeks, burning through five figures, and still arguing over whether the winner was just noise. Lucas: Exactly. So a DTC brand I've been watching — they sell a premium kitchen gadget, call it a smart peeler, around eighty bucks — decided to try something different. They have a brand community, about seventy-five people, in a Slack group. Mostly early adopters, some repeat buyers, a few people who'd never bought but were interested. Luna: Seventy-five people. That's tiny. Most brand communities I see are chasing thousands. Lucas: Tiny is the point. On a Thursday afternoon, the community manager posted rough cuts of all twenty-four ads. Not final — just storyboard frames with placeholder copy. She asked: which three would make you click? And which two would you scroll past without thinking? Luna: So it's a forced-choice exercise. That's smart — you're not asking people to rate on a scale, you're forcing them to pick winners and losers. Lucas: Right. By Sunday night, fifty-three of the seventy-five members had responded. That's a seventy percent response rate, which is absurd for any survey, but in a community where people feel invested, it's normal. Luna: And what did the data tell them? Lucas: The top three ads by community vote — clear consensus, each got around forty votes — were the ones they'd been leaning toward internally. But the interesting part was the bottom two. The brand's creative director loved one of them. Thought it was clever. The community hated it. Twenty-seventh out of twenty-four — only seven votes. Luna: Ouch. So the community saved them from a bad bet. Lucas: They ran the Facebook test anyway, because they had the budget and wanted to validate. And the two that the community rejected? One had the worst click-through rate of the whole batch. The other was the only ad with a negative purchase intent lift. Luna: That's a pretty strong correlation for a sample of fifty-three people. What's the margin of error on that? Lucas: It's wider than you'd want for a political poll, but for creative direction, it's extremely useful. With fifty-three responses, assuming a simple random sample, the margin of error is around plus or minus thirteen percent. But here's the thing: the community wasn't a random sample. They were people who already had context — they knew the brand, they knew the product category. Their feedback was more predictive than a general population sample because they were closer to the actual target audience. Luna: So you're saying a biased sample is better if the bias is the right bias. Lucas: Exactly. A community of enthusiasts is a biased sample — biased toward people who care enough to engage. If your product is for enthusiasts, that's the bias you want. You don't need a representative sample of all Americans; you need a representative sample of the people who will actually buy. Luna: That makes sense. But I've also seen communities where people give feedback that's too nice. They don't want to hurt feelings. How do you avoid that? Lucas: That's the design challenge. This brand did two things. One, they asked for the bad first. The question was: which two would you scroll past? Not: which ones do you like? They primed people to be critical. Two, the ads were rough cuts, not polished. It's easier to be honest about something that clearly isn't finished. Luna: That's a good insight. So the medium itself encourages honesty. Lucas: And speed. The whole exercise took a weekend. The Facebook test took two weeks and cost about twelve thousand dollars. The community test cost zero incremental dollars — the community already existed. The only cost was the community manager's time to post and compile results. Luna: But there's an ethical question here, right? You're getting free labor from people who think they're part of a community, not a focus group. Lucas: It's a real tension. The way this brand handled it was transparency. They told the community: we're testing ads, your feedback will directly shape what we launch, and we'll share the results with you before anyone else. So members saw their input matter. Also, the community wasn't sold as a research panel. It was sold as a place for early access and behind the scenes looks. Research is a byproduct, not the purpose. Luna: That feels fair — as long as you're not exploiting the relationship. I think the moment it becomes extractive is when you ask for feedback constantly without giving anything back. Lucas: Right. The brand also gave a small token — a discount code for the new product, same as any early-access member, but they made sure to emphasize that the discount was unrelated to participation. You didn't have to give feedback to get the code. Luna: So the feedback was truly voluntary. That's the gold standard. Lucas: I think the bigger lesson here is about scale. A lot of marketers think they need a community of thousands to get useful data. But for rapid iteration — testing a specific creative direction or a single claim — a small, engaged group can be more efficient than a large, passive one. Luna: You're measuring depth, not breadth. Lucas: Exactly. And if you're running these tests regularly, you start to build a track record. You learn which community members are most predictive. You can weight their responses. Over time, your seventy-five-person community becomes a more reliable signal than a thousand-person survey panel. Luna: That's a really powerful argument for investing in a community even if it's small. I think a lot of brands hesitate because they think, we'll never get to ten thousand members, so why bother? Lucas: And the answer is: because seventy-five passionate people who trust you can tell you things in a weekend that cost you twelve grand and two weeks to learn otherwise. That's not a small return. Luna: It's actually a huge ROI when you think about the speed alone. In a product launch timeline, two weeks is an eternity. If you can get directional feedback in three days, you can iterate before the final assets are even produced. Lucas: That's exactly what this brand did. The community feedback came in by Sunday night. Monday morning, the creative team re-cut the top three ads based on the community's critique of a specific headline. They had final assets ready by Wednesday. Launch was Friday. That's a one-week turnaround from test to launch, instead of three. Luna: And the launch performed? Lucas: The brand said it was their best launch ever by revenue in the first week. Now, that's not purely because of the community feedback — the product itself was strong — but the ads were the best they'd ever put out. The creative director told me she would have launched the ad the community hated if they hadn't spoken up. So yeah, they credit the community with a meaningful lift. Luna: You know, if these conversations about community-driven marketing have sparked something you've actually used — maybe a testing idea like this, or a way to engage your own members — and you've found value in the show, there's a way to support it that keeps it exactly like this: no ads, no sponsors, just us talking through the cases we find interesting. It's buy me a coffee dot com slash fexingo. One-time, no subscription, no pressure. Just if it's useful to you. Lucas: And we mean that. The show stays free either way. But listener support is what lets us spend the time digging into stories like this one — and keeps us from having to read ad copy for some mattress company. Luna: Right. So — back to testing. One thing I want to push on: the brand tested ads, but could this work for other types of marketing decisions too? Pricing? Positioning? Lucas: Yes, and we've seen examples. Another brand in the same community — different product, a subscription snack box — used their seventy-person Slack to test three different pricing page layouts. They showed mockups and asked: which one makes you feel like you're getting the best deal? The winner was not the one their pricing analyst preferred. They launched the community's choice and conversion rate improved by eight percent. Luna: Eight percent on conversion is huge. That's a direct line from community feedback to revenue. Lucas: It is. And it didn't require a statistician. It required a community manager who knew how to ask a clear question and a group of people who felt invested in the brand's success. Luna: So what's the bottleneck? Why don't more brands do this? Lucas: I think two things. One, they don't have the community in the first place. Building a seventy-five-person community takes time and consistent value. Two, even if they have the community, they don't trust the data. They think: seventy-five people, that's not statistically significant. But they're comparing it to the wrong benchmark. The benchmark isn't a perfect survey; it's the alternative, which is making decisions based on gut or on a single opinion from the CEO. Luna: Or on a focus group that costs ten grand and takes three weeks to recruit. Lucas: Exactly. So the question isn't: is the community data perfect? The question is: is it better than what you'd have otherwise? And in most cases, the answer is yes — and it's faster and cheaper. Luna: I think there's also a cultural shift needed. Marketers are trained to want large sample sizes. But when you're iterating fast, you don't need precision. You need direction. Lucas: That's the right framing. Direction over precision. You're not publishing a peer-reviewed paper; you're deciding which ad to put fifty grand behind. If the community says this one's a dud, you can test it at small scale to confirm, but you've already saved yourself from a potential disaster. Luna: So the community becomes a cheap, fast filter. It's not the final answer, but it's a very good first pass. Lucas: And over time, you calibrate. You learn that when the community is split fifty-fifty, it's a coin flip. But when there's a clear consensus — like the ads that got forty votes out of fifty-three — you can bet on it with confidence. Luna: I love that this is a repeatable process. It's not a one-off hack. It's a system. Lucas: It is. And the brands that do it well treat it as a muscle they exercise regularly. They run a test like this every month. They build a library of what the community likes and doesn't like. Over a year, that's twelve data points per ad element. That's a real corpus of knowledge. Luna: Alright, I'm convinced. Next time I'm working on a campaign, I'm finding a community to test with. Lucas: Start small. Find fifteen people who love your product. Ask them one question. See what happens. You might be surprised. Luna: And if it works, you've just built the beginning of something that can scale in value even if it never scales in size.