One 17-second outfit-swap ad brought 380k installs, then it stopped: Outfityz and the 'one theme, one app' paid acquisition matrix

A small Singapore company called FuMaoMao, in late August, posted more than a dozen AI outfit-swap short videos on YouTube in one go. Over a month they brought about 380k installs to its new app, Outfityz. Once the creatives stopped in October, daily adds fell from tens of thousands to a thousand or two. The app itself has nothing worth copying. What's worth looking at is the machine behind it: the same ad channel, the same creative style, and the same paywall have already produced three apps with a million installs each. Below is how it started, why it collapsed, and whether you can still take this path.

Aug 27: a channel with only a thousand subscribers

On the morning of Aug 27, 2026, the YouTube channel 'Sign' uploaded a batch of vertical clips of 17 to 18 seconds. None of the titles named the app. They only described results: 'AI Style Visualization: See How Different Outfits Suit Your Body & Personality', 'Your face, any outfit. AI creates your perfect look.', 'One selfie, endless outfit possibilities. AI stylist.' The descriptions were a single line, such as 'Never buy clothes that don't suit you again.'

By Nov 6, the first clip had 9.5M views, the second 3.72M, and the third 1.71M. This batch of 12 clips totaled about 16M views, with almost no comments (public YouTube data, 11-06 0). Lots of views and no comments is the typical look of ad creative.

The channel was created in May 2026, with 519 videos and 535M views, but only about 1,000 subscribers. Its video themes line up one-to-one with FuMaoMao's apps: AI-generated babies, photos that dance, AI hairstyle swaps, AI outfit swaps. The channel name 'Sign' likely comes from 'Signature Master,' the first app the company listed in April. No official statement says the channel is FuMaoMao's own. This is an inference, but a strong one.

What it is

Outfityz: AI Outfit Swap, listed on Google Play in 2026-08-17, developer FU MAO MAO PTE. LTD., address in an office building on Thomson Road, Singapore (per the Play store page). It lists four features: swap clothes by uploading a photo, change clothing colors, change backgrounds, and generate outfits from a one-sentence description, plus hundreds of templates (Korean style, Y2K, Indian traditional wear, weddings...).

The model is pay-to-open: in-app purchase from $0.49 to $29.99, no ads. 11-06 The store page shows 11,000 ratings at 3.5 stars, and its listed install tier is "100k+"; third-party estimates put lifetime installs at around 460k.

The install curve is the ad schedule

The two lines almost overlap: the 08-27 creatives went live, daily adds passed 09-03k from 20, 09-10 and 11 peaked, fell back in late September, and collapsed in October. The 9-04 batch of 30 clips clearly didn't take off, and ads were largely stopped in October.

You can also see where the traffic was bought: 11-06 it charted only in Nigeria (#36 on the Art & Design free chart) and India (#69), and never entered the overall chart in any country. In September 10 an Indian creator posted a Hindi tutorial, 'outfityz app kaise use kare', which got 923 views. Searching for it on TikTok, X, Reddit and in the media turns up no related content at all. Apart from the ads, it had no other traffic.

What users say

The three most upvoted reviews on the store page cover the app's problems:

'I can't even try it once, I have to buy a subscription first. I don't buy something I can't see results from. The outfit feature won't open at all.' Agent Tomcat, 10-09,17 people found it helpful

'Scam app, don't download! I downloaded it, couldn't use it, deleted it, and my money was still charged.' Akubueze Oluchi, 10-09

'The outfits are trendy, it works well, I like it.' Rebika Singnarpi, 08-28, the second day after the ad went live

The paywall sat before the first generation. Users saw results in the ad, came in, and couldn't try it. The ads were bought in places like India and Nigeria, where users have low willingness to pay. A third-party model estimates October revenue at under $5,000 (AppGazers, model estimate). The ads most likely didn't pay back, which is why they stopped. This is an inference.

The real star: one theme, one app

Outfityz is one of the weaker apps in this company's matrix. The same developer has more than ten apps on Play, with about 10M cumulative installs according to data sites:

The playbook is the same: pick a theme of 'AI turns your photo into X,' build the app in three to five weeks, batch-post 17-second effect demos on the same channel, and title them only with results. Once a batch of creatives performs, add budget; when it stops performing, cut it and move to the next theme. From May to August it launched 9 apps, and 3 of them passed a million installs. AppGazers estimates the whole matrix's monthly revenue at $130k to $690k, mostly from BabyGenie (model estimate, unverified).

So Outfityz didn't take off because the product was good. It took off because the company had a ready-made paid acquisition machine. It collapsed because this theme couldn't sustain that machine's costs.

Who it's competing with

AI outfit swapping is one of the most crowded AI photo themes of 2026. Apps in the same direction on Google Play (cumulative installs, early November):

The bigger rivals are free: Gemini's image editing, ChatGPT image generation, Google Doppl, where one sentence swaps the outfit. A Gemini AI outfit change tutorial on YouTube India alone has 130 million views. An app charging $0.49 per outfit change is competing for the exact same Indian users.

Notably, Hit Studio and SBS Tech: same theme, same market, and they're still climbing in October. That suggests outfit-swap demand isn't dead; what died is Outfityz's conversion in this version. My inference: the paywall may have come too early, or creative fatigue set in without a fresh supply.

If it were you

Verdict: Outfityz itself isn't worth copying, and the outfit-swap theme is already crowded, with free substitutes like Gemini pressing on it. What's worth studying is FuMaoMao's portfolio playbook, but that's a paid-acquisition business, not an AI business.

Who it fits: a small team of 2–3 people who have actually run Google App Campaigns or Meta, who track cost per install and first-week payback, who can ship an app in three to four weeks, and who have a budget they can spend on creative tests without losing sleep if it fails.

Who it doesn't fit: anyone who only codes and has never bought ads (these apps have no technical moat; generation is just calling off-the-shelf models), anyone hoping store search or organic social will build traction (Outfityz collapsed the moment spend stopped, proving these apps have no organic flow), and anyone wanting to compete head-on in the outfit-swap theme.

How to do it: if you insist on outfit swap, fix the step Outfityz got wrong. Let the first generation be free and watermarked so people see themselves in the outfit before paying, and sell per use rather than rushing into subscriptions. Better still, copy the playbook, not the theme: find an "AI turns your photo into X" concept that Gemini can't yet do well in one sentence, where the output is shareable (multi-photo, multi-step edits, or specific templates), and build one app per theme.

How to acquire users: follow the "Sign" channel's format. Vertical video, 15–20 seconds, before-and-after in the first three seconds, headlines that state only the result ("Your face, any outfit"), no app name. Test 10+ creatives at once and only scale the one or two that work. Don't chase just India and Nigeria for cheap installs: installs are cheap, but the paywall doesn't hold there. Put part of the budget into Southeast Asia, the Middle East, and Western markets to test paid conversion.

How long you have: there's no hard deadline, but the window for AI photo themes usually lasts a few months. Once big companies' free features do the same thing well, paid conversion drops. Outfityz went from launch to collapse in about two months.

What would mean I got it wrong: ① Outfityz or a new app from the same company regains installs after spend stops (that would mean organic retention exists, not just paid traffic). ② Hit Studio or SBS Tech, similar apps, are still scaling fast in December with stable ratings (meaning the outfit-swap theme can pay back, and the problem is only Outfityz's paywall design; then the theme is still worth building). ③ Conversely, if FuMaoMao's new apps keep dying after one or two months, the portfolio machine itself is failing, so don't copy it.

Data and sources

• Installs and ratings: Google Play store page (11-06). Monthly new installs are third-party estimates (AppstoreSpy); 10-29 to 11-06 is daily-sampled data.
• Charts: Google Play Art & Design category, Top Free and Top Grossing (from 10-29).
• Revenue: modeled estimate from the AppGazers developer page, not official figures.
• Ad creatives: public data from the YouTube channel "Sign" (video titles, upload dates, view counts, 11-06). Channel attribution is inferred.
• Reviews: "most helpful" reviews on the Google Play store page (11-06).
• Developer: Google Play developer info, Singapore company registration (FU MAO MAO PTE. LTD.).
• No ads for it found in the Meta or TikTok ad libraries. Not finding them doesn't mean they weren't run.