Case studies · Couples app · Named
Flamme: 53M organic views and D30 retention from 0.8% to 50%+, with zero paid spend
Published October 2026 · By An Nayal, Agniverse Media
In short
Agniverse Media is an operator-run UGC agency for consumer apps, and Flamme is the app its founders built the system on. Over seven months, from July 2025 to January 2026, 19 creators posted 4,228 videos for Flamme, cross-posted to TikTok, Instagram and Facebook, with zero paid spend. The result was 53M organic views, 200K users and Day 30 retention from 0.8% to 50%+. Arya then acquired Flamme.
- Client
- Flamme (our own app, acquired by Arya)
- Category
- AI couples app
- Window
- July 2025 to January 2026
- Platforms
- TikTok, Instagram, Facebook
- Volume
- 4,228 videos, 19 creators
- Paid spend
- $0
- Run by
- An Nayal and Lauriane Nayal
The app
Flamme is an AI couples app built around a daily ritual: a few questions a couple answers together, every day. An Nayal built it, and it is the one app on this site we can name, because it is ours.
Couples apps are hard to market. Nobody searches for one, the buyer is two people, and the category had no organic playbook to copy.
The starting point
- Flamme launched with 0.8% Day 30 retention: high churn and no organic traction.
- No paid acquisition budget, and no plan to buy installs.
- In July 2025 the first 187 videos earned about 0.4M views.
The challenge
A category with no playbook
There was no proven couples-app format to copy, so every angle had to be found by testing.
Installs that left
At 0.8% D30, more installs alone would not fix anything. The content had to bring in couples who wanted the daily habit.
Organic only
With no paid spend, every view had to come from content good enough to travel on its own.
What we did
Angles
9 audience angles, each a couple in a specific moment: long-distance, the couple who ran out of things to say, married with kids and lost the spark, new and want to go deeper, reconnecting after a rough patch, anniversaries, chores and money fights, hard conversations, and situationships.
Hooks
34 hooks written as bets against those angles. Each one got a batch of videos, then a verdict: kill it, mutate it into a sharper variant, or scale it across more creators.
Creators and distribution
19 creators produced the 4,228 tracked videos, each cross-posted to TikTok, Instagram and Facebook. Around them, the wider farm ran 50+ accounts posting 200+ pieces of content a day.
The scorecard
Hooks were judged on the users they brought in, not only on views. The winners describe the ritual the app is built around, a question at dinner or a nightly habit, so the people they attract already want that habit.
Results
| Metric | Before | After |
|---|---|---|
| Day 30 retention | 0.8% | 50%+ |
| Users | 0 | 200K |
| Monthly organic views | 0.4M (Jul 2025) | 19.8M (Dec 2025) |
| Total organic views | - | 53M (TikTok 26.0M, Instagram 17.7M, Facebook 9.3M) |
| Paid spend | $0 | $0 |
| Outcome | - | Acquired by Arya |
Views per month: Jul 0.4M, Aug 0.8M, Sep 3.5M, Oct 11.5M, Nov 12.0M, Dec 19.8M, Jan 5.0M. View counts come from the campaign board, rounded. Users and retention come from Flamme's own analytics.
The hooks that won, and the ones we killed
8 of the 34 hooks drove over 85% of the 53M views. Nobody could have picked them up front, which is why the system tests before it scales.
| Hook | Angle | Videos | Views | Verdict |
|---|---|---|---|---|
| “asking my husband questions I've never asked in 6 years (his answers shocked me)” | Ran out of things to say | 429 | 10M | Scaled |
| “POV: he's 8 hours ahead and you still know each other better than couples in the same bed” | Long-distance couples | 402 | 7.9M | Scaled |
| “3 months in and we already know each other like it's been years” | New couple, want to go deeper | 228 | 5.3M | Scaled |
| “we've been together 6 years and dinner was silent every night... not anymore” | Ran out of things to say | 322 | 5.3M | Scaled |
| “3 kids later and we finally feel like a couple again, not just co-parents” | Married with kids, lost the spark | 295 | 5.2M | Scaled |
| “the app that keeps long distance couples connected” | Long-distance couples | 29 | 6.5K | Killed |
| “Flamme is the app that fixed our conversations” | Ran out of things to say | 24 | 6.1K | Killed |
What we killed, and why
- Hooks that named the product. "Flamme is the app that fixed our conversations" earned about 260 views a video. The best problem-first hook earned about 23,000, roughly 90x more.
- The whole situationship angle. Its 3 hooks got about 0.2M views from 102 videos. Couples who are not sure they are a couple are not shopping for a couples habit.
- 9 hooks in total were killed. Each one only got a test batch before the verdict, so failed bets stayed cheap.
Takeaways
- Name the problem, not the product. Problem-first hooks beat product-first hooks by about 90x per video on Flamme.
- You cannot pick the winners up front. 8 of 34 hooks drove over 85% of the 53M views, and only testing found them.
- Scale means more creators on a proven hook. The 8 scaled hooks averaged about 300 videos each.
- Judge hooks on who they bring in. D30 retention moved from 0.8% to 50%+ over the same period, and the hooks we scaled sold the habit, not the app.
Frequently asked questions
How did Flamme get 53M organic views?
What was Flamme's Day 30 retention?
Which TikTok hooks worked for a couples app?
More case studies
Want this run on your app?
Agniverse Media is an operator-run UGC agency for consumer apps. The founders run every account, three clients a month, and programs start at $10,000 a month.
Bylined Agniverse Media, part of Agniverse - the growth universe for consumer apps. View counts come from our campaign board.