Viral Apps: Why Some Spike and Fade While Others Stick

Every app in this piece reached No. 1 or close to it on the US App Store, and most of them were back outside the top 100 within three months. The ones that survived did not have a bigger spike; they had something for people to do on day eight.

What a chart spike looks like

A viral launch has a recognisable shape in rank data: downloads concentrate into a few days, the app jumps 50–150 places, sits in the top 5 for one to three weeks, then decays as installs fall back toward baseline. Rank follows install velocity, not installed base, so an app can hold No. 1 with a fraction of the audience of the apps below it, and lose it just as fast.

Lapse is the textbook case. In September 2023 it went from No. 118 overall to No. 1 in the US after making new users invite five friends before they could use the app; Appfigures estimated daily installs rose from about 8,000 at the start of the month to about 210,000 at the end, peaking at 218,000 in a single October day ^13,14. Eight weeks later it was averaging around 63,000 installs a day, down roughly 70% ¹⁴.

The seven cases

BeReal: the spike that lasted a summer

BeReal held the top of the US free chart from late July 2022 into September (dates from secondary trackers, unverified) and was named Apple's iPhone App of the Year that November ⁵. By October 2022 Sensor Tower counted more than 53 million installs, with monthly installs still rising from 12.3 million in August to 14.7 million in September. The warning sign was already visible: only about 9% of active Android installs opened the app on a given day, against 39% for Instagram and 29% for TikTok ¹.

A year later Similarweb estimated global monthly actives at about 16 million, down from roughly 20 million, with US monthly users falling from a 3.7 million peak in November 2022 to just over 3 million in August 2023; BeReal disputed the figures and said it had more than 25 million active users ². Voodoo bought it for €500 million in June 2024 ³, and by April 2025 the company was claiming 40 million monthly users and turning on ads ⁴. BeReal's mechanic, one notification a day, was original but gave people nothing to do between prompts.

Clubhouse: scarcity without a product underneath

Clubhouse had the opposite problem. Its invite-only design made the February 2021 spike enormous relative to its size: App Annie estimated global downloads rose from 3.5 million on February 1 to 8.1 million by February 16, with 2.6 million of those in the US ⁶. The app only opened to everyone in July 2021, with 10 million people on the waitlist by the company's count ⁷. By then the lockdown conditions that made live audio compelling were gone. In April 2023 the company cut more than half its staff ⁸.

Threads: the biggest drop, and the stick

Threads signed up 100 million people in five days in July 2023, riding Instagram's account graph ⁹. Then engagement collapsed: Similarweb's Android panel showed worldwide daily users falling from 49.3 million on July 7 to 10.3 million on August 7, a 79% drop, with US time per user falling from about 21 minutes to 3 ¹⁰. On a rank chart this looked like every other fade.

The difference is what Meta did in the following two years: DMs, custom feeds, fediverse integration, communities and a constant drip of features. By August 2025 Threads reported more than 400 million monthly users, and Similarweb's June 2025 estimate put its mobile daily users at 115 million, up 128% year over year ¹¹. The spike-and-fade shape is not destiny; it is the shape of a launch that has not yet earned its second month.

Lensa: a feature trend, not an app trend

Lensa's Magic Avatars launched in late November 2022 and took the app to No. 1 in the US Photo & Video chart within days; Sensor Tower estimated 1.6 million downloads in November, up 631% from October, with Brazil alone providing 31% of them ¹². The product was a $3.99 one-off purchase for 50 AI portraits ¹². That is the tell: people bought a trend, not a habit. Once the avatar wave passed on social media, the reason to reopen the app went with it.

Lapse: forced invites, fast decay

Lapse's five-invite gate is the purest example of buying rank with a growth mechanic. It worked for roughly six weeks, then daily installs fell about 70% from the peak ¹⁴. Forced invites fill the funnel with people who did not choose the app, so the invited cohort retains worse than organic installs. The company claimed millions of daily users, which the reporting could not verify ¹⁴.

Yik Yak: a nostalgia spike

Yik Yak relaunched on iOS in August 2021 and briefly hit No. 1 overall in the US the next day, driven by people who remembered the 2014 original ^15,16. By March 2023 it had about 3.5 million installs from the relaunch, sat at No. 89 in Social Networking, and had quietly been absorbed into Sidechat's developer account ¹⁶. A spike built on recognition has nowhere to go once the curious have looked.

NGL: a spike with a rotten core

NGL became the No. 1 app on the US App Store in June 2022 by letting Instagram users collect anonymous questions, and passed 15 million installs and $2.4 million in revenue by July ^17,18. Part of that engagement was manufactured: the app sent fake questions when a user's link got little traction, then sold $9.99-a-week "hints" about the sender. After reporting exposed the practice, NGL labelled the fake messages and cut the price ¹⁸. In July 2024 the FTC and the Los Angeles DA settled with the company for $5 million and banned it from serving under-18s, the first order of its kind ¹⁹. The spike was real; the retention was partly fiction, and the regulator eventually priced it.

What retention decided

Three things separated the stickers from the faders.

A reason to return that does not depend on the trend. Threads had the conversation graph of Instagram; Lensa had a one-off purchase. BeReal had a daily prompt but nothing between prompts.

Shipping through the trough. Every viral app loses most of its launch cohort. Threads kept shipping for two years; Clubhouse spent its window on a waitlist.

Honest engagement. Forced invites (Lapse) and fabricated messages (NGL) raise install velocity and lower the retention of the people they bring in. The chart cannot tell the difference for a few weeks. The retention curve can, and so, eventually, could the FTC.

How this shows up in the charts

The two curves in the figure are visible in rank history. On the US iOS free chart a sticker holds a stable rank for months with small daily moves; a fader shows a vertical climb on the movers page, one to three weeks near the top, then a long slide that is usually faster than the climb. Because rank reflects install velocity rather than installed base, the fade begins the moment launch coverage, invites or paid spend stop, which is why we suggest reading how App Store rankings work alongside the retention benchmarks before judging any launch by its peak position.