The Superwall Podcast · 2026-01-13

Roger Chen - How I Built the #1 App on The App Store (Twice)

Roger Chen on two #1 consumer social apps. Bro (AI companion floating over your texts): the team filmed a ProtoPie mock-up, bought TikTok ads on it to validate the concept and find a format (7.9M views on the brand account), then scaled hook-and-demo UGC with up to ~50 TikTok creators to #1 in the US. Lobby (drop-in video hangouts): flat for a year until it went viral in Israel, where 60% of new users had 5 friends on day 1 (vs ~20% elsewhere). The team then built invite features like a synchronized, watermarked Group Picture, and Lobby hit #1 in the UK, Italy, Germany, Egypt and Saudi.

Play the episode

8 tactics

Bro · B2C · AI companion app

Film a clickable mock-up and put TikTok ad money behind it before building

7.9M views; proved the concept and found a video format to scale

“when we made this video, like, uh we haven't actually finished building the app yet, but we want to like make a demo and uh see if people want to use it.”

Workedvalidationpaid-social

The Superwall Podcast · 2026-01-13

Bro · B2C · AI companion app

Pick one everyday moment, texting, as the way into a big AI companion idea

Texting wedge became the winning format; lock-screen versions also did well

“even though we found like the whole vision like just super, super exciting, we still need to find a wedge, like a concrete wedge, like a concrete scenario in people's lives.”

Workedpositioning-nichevalidation

The Superwall Podcast · 2026-01-13

Bro · B2C · AI companion app

Post lots of TikToks: a real person in a texting mishap, then the app helping

Very top-heavy: a few videos hit 2-3M views; the best one was a surprise

“it's basically it's the classic hook and demo. So it's like the hook is the scenario, you need a real person there, cuz you- you don't have a real person there, people swipe away.”

Workedugc-creatorsorganic-short-video

The Superwall Podcast · 2026-01-13

Bro · B2C · AI companion app

Find TikTok creators through topic hashtags like #relationship and email them to post

~50 creators at peak, down to 10-20; hits took 2-3 months; hit #1 in US

“what we did was we tried to get creators from TikTok by searching like hashtags, like relationship, and um and then, you know, you go into those people, and then you see you try to find people like those creators, and then you just like email them”

Workedcreator-opsugc-creators

The Superwall Podcast · 2026-01-13

Lobby · B2C · drop-in video hangout app

Add a group photo countdown to video calls that people share to Stories

Credited with more sharing and pulling in left-out friends; #1 in UK, Italy, Germany 3 months later

“So it's like you you press a button, and then everyone sees on their screen like 3, 2, 1, like a countdown, and the phone vibrates, and and then you take a boomerang of a picture.”

Workedproduct-viralityretention

The Superwall Podcast · 2026-01-13

Lobby · B2C · drop-in video hangout app

Track how many new users get 5 friends on day one, then build for it

60% vs 20% hit 5 friends in Israel vs elsewhere; D90 25-30%, about double

“And when we looked at it, like there's only really one key difference, uh which is the network density.”

Workedanalytics-attributionproduct-virality

The Superwall Podcast · 2026-01-13

Lobby · B2C · drop-in video hangout app

Invite friends to a 1-minute video catch-up at a random minute each night

“at night um, you know, at a random minute, you know, the app is going to you're going to see like a little little bar pop up”

Unknownretentionnotifications

The Superwall Podcast · 2026-01-13

Bro · B2C · AI companion app

Carry a tappable fake of the app and use it daily before coding

Founder says it cut a lot of wasted engineering; no numbers

“Like we want to we build a mock-up, and then we first like use it amongst ourselves, like the mock-up.”

Workedfeature-developmentvalidation

The Superwall Podcast · 2026-01-13

Numbers mentioned

Subject	Metric	Value	Quote
Lobby	users Daily active users (peak)	500K users/day approximate	“You got 500,000 users, daily active users on Lobby, and then what happened?”
Lobby	retention D90 retention in Israel during the 2022 breakout	25–30% approximate	“I think like the day 90 uh for uh Lobby in Israel back then was like 25 to 30%, uh so like super solid, and like the retention in other places like half that.”
Lobby	share Share of new users who got 5 friends on day 1, Israel (start of 2022)	60% approximate	“we tracked this like one number, which is like the percentage of people who got five friends on day one. Uh in other places it's like like 20%, but in Israel it's like 60%.”
Lobby	time to milestone Period with no real growth before the Israel breakout	12 months approximate	“we did not really grow for like a year”
Lobby	rank App Store rank in Israel; stayed at #1	1 rank	“but it suddenly like went viral, and became number one in the App Store, and just stayed there”
Lobby	rank App Store rank in the UK, then Italy, then Germany, ~3 months after the Israel breakout	1 rank	“Uh and uh 3 months later, Lobby became number one in the UK, and then Italy, and then Germany.”
Lobby	rank App Store rank in Egypt, then Saudi (spread from UK via international-school students in Cairo)	1 rank	“Lobby became number one in Egypt, and then Saudi, too.”
Bro	rank US App Store rank	1 rank	“But you hit number one on the App Store, all right? Like, how did this happen?”
Bro	views Mock-up demo video (texting scenario, floating Bro) on the brand TikTok account, boosted with ads before the app was finished	7.9M viewsper video	“This was a 7.9 million views. This was on your brand account”
Bro	count Other mock-up ad variations tested before going all in on the floating-window texting format (earlier ones focused on the lock screen)	5–6 videos approximate	“I think we had like five or six more.”
Bro	ad spend Rough cost of validation ads before launch, framed as trivial next to getting instant feedback	a few thousand dollars illustrative	“add a few thousand dollars of ads, it's nothing uh if you can get feedback instantly”
Bro	views A few outlier UGC videos; results were very top-heavy	2M–3M viewsper video approximate	“we do have a few outlier hits, like uh a few videos got like 2 2 to 3 million views. So it's like really, really top-heavy.”
Bro	views Creator video: 'I sent I love you to my crush by accident' autocorrect panic, Bro suggests a reply	2.3M viewsper video	“Yeah. 2.3 million views. That's really good.”
Bro	time to milestone Time for the UGC creator program to produce breakout hits	2–3 months approximate	“the ones the breakout hits that we saw like lately, like those took like 2, 3 months.”
Bro	count Creators at peak; many stopped posting, falling to ~10-20 active	50 creators approximate	“at our height, we have like maybe like 50 creators, and then a lot of them stopped posting, but so it went down to maybe like 10, 20.”