Latest / Social Media Marketing with Fexingo: Instagram, TikTok, LinkedIn, and Platform Strategy / Why Brands Are Using Social Media to Run User-Generated Ad Tests
Transcript
- Lucas: So, there's a cosmetics brand I've been following — they're not huge, maybe forty thousand followers on Instagram — and last month they did something that felt almost too simple to work. They posted two versions of the same product video, one highlighting the clean ingredient list, the other leaning into the founder's personal story, and they let their followers vote in the comments which one felt more compelling. Luna: That's basically a focus group but with zero recruiting cost. Did they actually do anything with the results? Lucas: They did. They took the winning angle, turned it into a proper ad creative, and ran it on Instagram and TikTok. And here's the kicker — the organic engagement ratio on that post predicted the paid click-through rate better than any of their previous guesswork. The post that got more saves and shares, not just likes, was the one that outperformed in paid by about thirty percent. Luna: Thirty percent is meaningful. But I wonder if that's just because the content was already proven to resonate with their most engaged audience, which is a bit of a self-selecting group. Lucas: That's the fair critique. But it's also the point — you're testing with the people who already know you, so you're not looking for a broad appeal, you're looking for a signal on which angle to scale. And the cost is basically nothing. You're using your organic feed as a cheap creative lab. Luna: So this is turning into a trend — brands using their social comments and DMs as a testing ground before they spend a dollar on paid media. I've seen a few small brands do this on TikTok, but it seems like it's becoming more systematic. Lucas: It's definitely moving from accidental to intentional. There's a whole category of tools emerging that mine comments and direct messages for patterns — they'll flag which phrases show up repeatedly, which questions get asked, which emojis cluster around which types of posts. It's like social listening but focused on the creative itself. Luna: And platforms are making it easier, right? I think TikTok has a feature where you can repost a user's video as an ad, and Instagram has something similar for reels. Lucas: Exactly. Both platforms now let you take an organic post, one that's already performed well, and boost it with a 'create similar' or 'use as ad' option. So the loop is: post organically, see what sticks, then amplify the winner. That's the whole play. Luna: But what about the risk? If you base your paid creative on engagement, you might be optimizing for what's popular with your existing followers, not what converts new customers. Likes can be a vanity metric. Lucas: That's the key nuance. You have to look at the right engagement metrics. Saves and shares are way better signals than likes or comments. A save means someone wants to come back to it — that's intent. A share means they're vouching for it. Comments can be misleading because sometimes the comments are just arguments or spam. Luna: So what's the actual framework you'd recommend to someone listening? They can't just post two videos and hope for the best. Lucas: Right. So first, you pick one variable — either the hook, the angle, or the format, but not all three. Then you post at the same time on the same day, to control for timing. You let it run for forty-eight hours, and you compare not just total engagement but the ratio of saves to views. That ratio is the closest thing to a pre-paid proxy. Luna: And what if neither post gets meaningful engagement? Then you haven't lost much, but you also haven't learned much. Lucas: That's when you go back to the drawing board. But the nice thing is, you're testing cheaply. Every failure is a data point, and you're not burning ad spend on something that doesn't resonate. Some brands I've talked to are running these tests weekly, just iterating on hooks and thumbnail styles. Luna: There's something elegant about using the platform's own organic feedback loop to tune your paid strategy. It feels like a more honest way to create than guessing in a boardroom. Lucas: And it's a shift from the old model where you'd hire a creative agency, spend six weeks on a campaign, and then hope it worked. Now you can test five ideas in a week, and the market tells you what's worth scaling. It's a much faster feedback loop. Luna: But I have to ask — what about the platforms' algorithms? If you post two videos, the algorithm might just show one to more people, and then it gets more engagement, and it snowballs. That's not exactly a fair test. Lucas: That's a real caveat. To mitigate that, some brands use the same post with different captions, or they use a small paid spend to guarantee equal reach. You can set a tiny budget like five dollars per post to level the playing field. But honestly, most brands I've seen just accept the noise and run the test a few times. Luna: So the signal might not be perfect, but it's better than nothing. And for a small brand, it's a way to get data without a huge budget. Lucas: Exactly. And there's a broader point here — this is part of a bigger trend of brands treating social media as a direct line to their most engaged customers, not just a broadcast channel. They're using it to test everything from product names to pricing to ad creative. It's all about letting the market speak before you commit. Luna: And you mentioned comment mining tools earlier — do you have any examples of how brands are using those in practice? Lucas: I know of a small food brand that used comment mining to discover that their customers kept asking about 'snackable portions' and 'on-the-go packs.' They didn't have that in their messaging at all. They tweaked their ad copy to highlight those phrases, and their click-through rate jumped by eighteen percent in a week. Luna: That's a perfect example of listening to the language your customers actually use, not the language you think is clever. Lucas: Right. And that's the thing — the words your audience uses in comments are often the same words that work in paid ads, because they're already proven to resonate. It's like getting the exact search terms for your product but for creative. Luna: So you're basically saying that the comment section is a goldmine of ad copy, waiting to be mined. Lucas: It is, if you know how to look. And the tools are getting better at filtering out noise and pulling out the actionable phrases. It's not just about counting comments anymore — it's about understanding the sentiment and the specific language that signals intent. Luna: Honestly, this is exactly the kind of practical, low-cost strategy that I think our listeners can use. And I know you mentioned that this is a topic you've been exploring a lot lately. Lucas: Yeah, and if the conversations we have on this show spark something you've actually used, that's the thing. There's a small link in the show notes — buy me a coffee dot com slash fexingo — if today's episode was worth the price of a coffee to you. Luna: It's the smallest ask, but it genuinely keeps us going ad-free and independent. Lucas: It does. And we appreciate every single one. Now, back to the testing — the other piece I think is underrated is the 'after' part of the test. Once you pick a winner, you should still go back and look at why it won. Was it the hook? The color grading? The pacing? That's the learning that transfers to your next test. Luna: So it's not just about the winner, but about building a library of insights over time. Each test teaches you something about your audience that you can apply to future creative. Lucas: And that's the compounding benefit. After a few months of running these tests, you have a sense of what angles resonate, what formats work, what words to avoid. It's like having a creative playbook that's constantly updated by real data. Luna: I'm curious — have you seen any brands take this to the next level, like automating the whole process? Lucas: There are a few startups that are trying to automate the testing loop — they'll generate multiple creatives, post them organically, measure engagement, and then automatically boost the winner with ad spend. But from what I've seen, the human element still matters for interpreting the 'why' behind the numbers. Luna: So the tools can handle the mechanics, but the insight still comes from a human who understands their brand and audience. Lucas: Exactly. And that's probably a good place to leave it. The trend is moving fast, but the core principle is timeless: test cheap, learn fast, and let your audience tell you what they want to see. The question is whether we're heading toward a world where every ad creative is essentially user-tested before it ever runs. Luna: I think we are, and honestly, that might not be a bad thing. It makes advertising feel less like a gamble and more like a conversation. Lucas: A conversation backed by data. That's a pretty good way to think about it. Thanks for joining us.