Reducing choice paralysis in a 120M-user rewards app.
Increasing onboarding activation and completion by 20%
A data-led onboarding redesign to cut acquisition cost and lift early retention. Swagbucks was paying to acquire users it lost in the first five minutes.
New sign-ups landed on an uncurated wall of tasks and left before completing one.
Set first-task completion as the north star, then designed a guided first earn, progressive profiling with a swipe feed, and streaks, quests and stories.
Six months after launch, four numbers moved.
The app opened with every option instead of a first step.
Swagbucks is the rewards platform under Prodege, with 120M members and more than $2B paid out across shopping, casual gaming, and survey research. New sign-ups landed on an uncurated wall of tasks. Leadership read the drop-off as an acquisition spend problem. The funnel data read it as a first-session problem.
A text-heavy, uncurated task list met new members before they had picked anything.
In a category crowded with scams, confusion reads as illegitimacy. Users left to verify the app was real.
Nothing in the first session gave a reason to open the app a second time.
We agreed on what success meant before we agreed on the design.
Not sign-ups, not screens seen. One completed, paid task in the opening session — the shortest honest proof the platform worked.
I defined it with Product Management and the data analysts, so design work was judged against the funnel rather than against taste.
Supporting measures · +
I ran the rewrite as a funnel problem across twelve weeks.
Intent to earn was high. Abandonment spiked before the first task.
I paired wide-scale surveys with unmoderated diagnostics, then read both against the funnel with Product and Engineering to find where the journey actually broke.
“Always a hunt involved... you're trying to pick up on how this actually works.”— Qualitative participant
Where the journey broke · +
Pay out before asking for commitment.
In a category crowded with scams, members arrive skeptical. Retention is won or lost in the first five minutes, and only a real, visible reward closes the gap. A guided first earn turned onboarding into the proof.
The three moves · +
- 1Show the payout up front. Task, effort, and reward stated plainly before a member starts.
- 2Land the first reward fast. A short, guaranteed task inside the opening session, no card, no wait.
- 3Celebrate the milestone. A high-polish completion moment anchors the peak-end memory.
Curation beats catalogue, and forms are the wrong way to earn it.
A curated feed needs preference data, and preference data usually costs a long form. To avoid trading one drop-off for another, the profile had to be gathered through engaging micro-moments instead.
Members categorised interests three times faster than the legacy form, and reported 40% less perceived effort.
How it works · +
One completed task is a transaction. A habit needs a next reason.
Members who earned once tended to stay inside the single vertical they arrived for, and then stopped. Retention depended on giving them momentum to carry forward and a path into the verticals they had never tried.
The flow shipped because five functions shared one metric.
Onboarding touches acquisition spend, the task marketplace, and the release train at once. I ran the working sessions that turned five separate mandates into one measurable outcome, and I owned the tradeoffs when they conflicted.
Who owned what · +
I made three tradeoffs to ship inside twelve weeks.
Six months in, the lift held from first session to lifetime value.
Great onboarding is the first conversation, not the finish line.
A fast, visible reward earned the trust the category withholds by default. Progressive profiling cut the wall of tasks down to a feed. Streaks and quests gave members a reason to return. The gain came from ordering those asks correctly, not from any one feature.