How we grew WhoseWhere from 600 to 10,000+ app installs, and built the data to see what happened next
WhoseWhere is a social app for people in Bali: find who's going where, connect with other people, join events and use venue deals. Gufy ran its paid acquisition on Meta and TikTok, set up its analytics and attribution, and built the push and email journeys that bring users back. Here is what we did, what worked, and what the data told us.
Featured case study · WhoseWhere · Mobile app (iOS and Android)
The starting point
When WhoseWhere's founder Jason came to us in October 2024, the app had around 600 downloads. Most came from in-person events, which worked (60 to 75 downloads in a single four-hour event) but couldn't scale. Radio hadn't worked, social posts got reach but few installs, and the previous agency's Meta ads sent people to a landing page rather than the app store, with no install tracking in place.
The goal was 150 to 500 new installs a week, ahead of conversations with large Bali venues.
Fix measurement before buying installs
We didn't increase spend until we could see what spend was buying. In the first phase we:
- Added install tracking: the Meta SDK in the app, and AppsFlyer to attribute every install to the campaign and creative that drove it.
- Wrote a lean tracking plan: only the events needed to make decisions (sign-up started and completed, profile created, ticket purchase started and paid, event interest, friend requests, carousel swipes, chats), to keep tool costs down.
- Piped events once, used them twice: Segment sent the same events to Mixpanel for product analytics and to CleverTap for push, email and in-app messaging. The app is built in Flutter, which Segment's CleverTap bundle doesn't support, so we worked with the app's developers to install the CleverTap SDK natively on iOS and Android.
- Handed it over properly: we trained WhoseWhere's own team on CleverTap and Mixpanel so they could send campaigns and read the numbers without us.
Split acquisition by platform, not just by audience
Meta app campaigns started in April 2025 and scaled up from October 2025. The data quickly showed that iOS and Android needed different plans:
- Android: Facebook placements brought cheap installs, but only 30% to 40% of them went on to create a profile. Moving Android spend to Instagram-only fixed the quality.
- iOS: Meta struggled to reach iOS users in volume, so we added TikTok for iOS. It cost more (about $4.50 per install, several times Meta's) but brought users with higher day-7 retention.
- Creative: a script-first UGC process with a local creator, then testing angles. An anti-dating-app angle ("Not Tinder") was the clear winner on Meta and drew a 7.41% click-through rate on TikTok, against 0.4–0.6% for most other creatives. Video drop-off data showed most viewers left between the 25% and 50% marks, so new scripts focused on the first half of the video.
From April 2025 to May 2026, Meta app campaigns delivered about 15,000 installs for A$12,909, an average of A$0.86 per install. The best month, February 2026, brought 3,097 installs at A$0.36 each. On the app stores' own count, WhoseWhere passed 10,000 installs in May 2026.
Bring users back automatically
Installs only matter if people come back. Using CleverTap we built:
- A win-back journey: push notifications that show lapsing users new people to connect with. It brought back about 30% of users who were drifting away.
- Event and ticket reminders: pushes 24 and 48 hours before events users had shown interest in or bought tickets for.
- Connection notifications: alerts when someone wants to connect at the same event. "Find Your Paddle Buddy" alone drew 187 users.
- Deal pushes: reminders about venue discounts for cost-conscious users in Bali.
Let the data steer the product, not just the ads
Once installs were flowing, Jason and Gufy reviewed what users actually did in the app, not just how many arrived. Each finding led to a change:
- Find the one feature people come for. WhoseWhere launched with events, venue deals and a people-matching carousel. Mixpanel showed the carousel was the clear favourite: two-thirds of weekly active users swiped, while far fewer used events. That gave the team the confidence to make connecting people the core of the app, put the carousel on the home screen, and carry that one message through ads, onboarding and push.
- Rebuild matching for a bigger user base. The original matching was built for a small community and ranked people by distance. At 8,000+ installs, it could surface someone nearby who hadn't opened the app in weeks. The data made this visible (on one busy day, 465 right swipes led to 5 matches), so the team moved to ranking by recent activity. That way new users meet people who are actually around.
- Smooth the first minute. Funnel data showed 26% of new users stopped at the profile-photo step straight after sign-up. That step was reworked so people reach the app before being asked for a photo.
- Back the channels that pay off. TikTok brought users who stayed longer, but at several times Meta's cost per install. Weekly reporting let Jason compare the two quickly. He moved that budget into product features and kept Meta as the main install channel.
- Turn the audience into revenue. With 10,000+ users and the strongest engagement around social events like "Find Your Paddle Buddy", WhoseWhere began running its own paid events. Gufy set up a two-path campaign: ads straight to a web ticket page for immediate revenue, and install ads followed by push reminders to buy tickets. The goal is to let ticket sales pay for user acquisition.
The growth system behind the result
Capabilities used
Mobile app growth · Performance marketing · Product analytics consulting · Lifecycle marketing · Marketing automation
Stack: Meta Ads, TikTok Ads, AppsFlyer, Segment, Mixpanel, CleverTap
How the numbers are measured: amounts are Australian dollars. Install counts and costs come from WhoseWhere's Meta ad account; the 10,000+ total is the app stores' count. Win-back rate is the share of lapsing users who returned after the CleverTap push. Reported by Gufy, not independently audited.