← Back to work
Redesigning onboarding for a dating app so more users reach a complete profile
Client
FilipinoCupid
Year
2024–2025
Role
Product Designer
Initial market
Philippines
FilipinoCupid is a dating app with 5M+ users. I owned the onboarding and activation work shown in this case: research, profile setup, safety and notification settings, and UX copy. I worked in a team of 11, alongside four other designers. The initial focus was the Philippines, with country selection designed to support expansion to other markets.
28% → 43%
completed onboarding with interests
39% → 64%
profiles with 3+ photos
30% → 24%
early D1–D3 churn
Role
Product Designer
Scope
Research · UX/UI · UX copy · A/B test
Platform
Mobile app
Year
2024–2025 · Philippines
Research
Finding the friction behind the drop-off
I used Amplitude to locate registration drop-off, then reviewed around 100 app reviews, user comments and support requests. The feedback helped explain repeated setup, long forms and uncertainty about whether personal details could be changed later.
82%
71%
60%
52%
45%
28%
Base: everyone who started registration. Source: product analytics, before the redesign.
USER FEEDBACK
Three barriers between signup and the first swipe
Three recurring themes shaped the redesign: a second setup stage after payment, too many fields at once, and uncertainty about editing personal details later.
PROBLEM 01
The setup started twice
People completed registration and paid, but still needed to enter preferences before they could start swiping.
PROBLEM 02
Personal questions lacked context
Users could not tell why age, location and preferences mattered — or whether those answers could be changed later.
PROBLEM 03
Too many decisions at once
Long forms mixed account details with personal preferences, making the remaining effort difficult to judge.
One journey to a match-ready profile
Essential matching details stay in onboarding, so the first action after the paywall can be discovery — not another setup form.
BEFORE · TWO SETUPS
1. Long registration →
2. Paywall →
3. Preferences again →
4. Start swiping
AFTER · READY AT PAYMENT
1. Guided profile setup →
2. Paywall →
3. Start swiping
DESIGN HYPOTHESES
Three hypotheses behind the redesign
The goal was not fewer questions at any cost. It was less effort per decision, more trust and a complete profile before purchase.
HYPOTHESIS 01
Smaller steps, visible progress
A focused question on each screen makes the journey easier to follow without dropping information the product needs.
HYPOTHESIS 02
Make personal questions understandable
Explain why details are requested and what can be edited later, instead of asking without context.
HYPOTHESIS 03
Finish setup before payment
A completed profile lets subscribers start discovering matches immediately instead of entering the same information twice.
How the onboarding experience changed
Explore four friction points in the original flow, where people struggled and what needed to improve.
Explore the redesigned screens
The redesign breaks registration and profile setup into focused steps, with clearer progress, guidance and feedback. All steps are required; there is no Skip action.
A/B TESTING
Testing the choices that shaped activation
I tested three activation decisions: photo requirements, purchase CTA placement and interest selection. Each A/B test ran for one month with approximately 50,000 users, randomly split 50/50 between variants.
TEST 01
Hypothesis: Asking for at least three photos could create unnecessary friction. We compared that rule with a screen inviting users to upload up to seven photos, then tracked both step completion and how many photos people added.
OPTION A
At least 3 photos
OPTION B
Test results
Philippines · New users reaching photo upload
Primary metric
Photo-step completion
Minimum 3 photos
85.2%
Flexible upload
77.1%
Quality check
Share of profiles with 3+ photos
Shipped: minimum 3 photos, with higher photo-step completion.
Paywall: sticky vs inline purchase CTA
Hypothesis: Keeping Get Premium visible while people scroll through their personalized match preview, benefits and plans will increase paid conversion compared with a button placed only by the plans. These screenshots show different positions of the same sticky-CTA paywall, not the two test variants.
SCREEN 01 · TOP
Personalized match preview
SCREEN 02 · LOWER
Plans with the CTA still visible
Test results
Philippines · New users reaching the paywall after profile setup
Primary metric
Purchase conversion among paywall viewers
Inline CTA
8.2%
Sticky CTA
9.5%
Supporting signal
Purchase CTA engagement
Shipped: sticky purchase CTA, with higher paid conversion.
Interests: minimum one vs minimum three
Hypothesis: Letting people continue after selecting one interest will improve completion, while requiring three could provide stronger matching signals. The list layout stays the same; the minimum selection rule and number of checked interests change.
OPTION A
At least 1 interest
OPTION B
At least 3 interests
Test results
Philippines · New users reaching the interests step
Primary metric
Interests-step completion
Minimum 1 interest
75.9%
Minimum 3 interests
81.6%
Quality check
Match-ready profiles and interests selected
Shipped: minimum 3 interests, with higher interests-step completion.
RESULTS
Results after the onboarding redesign
We compared two product snapshots: six months before release and one month after release. Completion with interests and profiles with 3+ photos increased, while early churn decreased.
Completion and 3+ photo rates use everyone who started registration as the base. D1–D3 churn is the share of registered users who did not return during the first three days after registration.
Onboarding with interests
+15 pp
28% → 43%
Profiles with 3+ photos
+25 pp
39% → 64%
D1–D3 churn
−6 pp
30% → 24%
What changed, before and after
Onboarding with interests
Before
28%
After
43%
Profiles with 3+ photos
Before
39%
After
64%
D1–D3 churn
Before
30%
After
24%
REFLECTION
What I learned
01
One journey can address several problems
Bringing profile setup into one guided journey addressed repeated setup, long forms and uncertainty together. This showed me how a single flow can connect several product problems.
02
Detail works when its value is clear
I learned that onboarding can ask for detailed information within reasonable limits. Personalization and a sense of safety make that effort feel worthwhile.
03
Shared foundations support parallel design
The challenge was delivering one coherent app with five designers under tight deadlines. I owned onboarding and activation, while our shared UI kit provided a common foundation across the team.