All workEsc

Game UA.Beyond theinstall.

I helped turn rewarded advertising from a single paid action into a journey through a product. Game UA, our approach to bringing new players into games, was where I designed and helped prove that model.

Two illustrative mission-page screens connecting a game introduction to rewarded in-game progress

A mission-page concept with sample game content.

The product idea

From paying for an actionto designing what happens next.

In a cost-per-action campaign, or CPA, the advertiser pays when someone completes a defined action. The campaigns we wanted to move beyond centered on that single moment. But an install alone did not mean someone had discovered the game or found a reason to keep playing.

The starting point

One action.
One reward.

Install the gameReward

The campaign reached its goal. What happened inside the game remained the harder question.

The shift we designed

A sequence of steps
inside the experience.

  1. 01Get started

    Enter the game.

    Reward
  2. 02Discover it

    Reach a first milestone.

    Reward
  3. 03Go further

    Progress toward the next goal.

    Reward
The campaign model, illustrated.

The campaign became
part of the product journey.

We connected rewards to steps inside the game. Players could see what to do next and what they would earn; advertisers could guide participation beyond the initial install. My early design work made that sequence understandable and helped establish that the model was worth developing.

Beyond games

The reusable idea was the journey,not the game.

We then separated this approach from the game-specific experience and carried it into brand and shopping experiences. The content could change; the underlying idea stayed the same: help someone discover a product through a sequence of clear, rewarded steps.

Games

From install
to participation.

Introduce the game, make progress visible and give the player a next goal.

Brands

From seeing a campaign
to discovering a product.

Give people an experience with the brand, beyond a view or a click.

Shopping

From browsing
toward a purchase.

Organize the experience around steps that bring someone closer to the advertiser’s goal.

These describe the design approach across experiences, not a claim that every journey led to a purchase.

Explore the later research into personalized brand experiences

A larger idea.
Specific problems to solve.

A sequence can also create friction. Could people get into it easily? Did choosing their rewards help? Did they understand and trust their progress? The following studies examine those questions, across my early work and the team’s later experiments.

Zoom in · Getting into the mission

Removing a screenmade the first step easier.

The explanation page repeated information already on the mission page. I proposed going straight to the experience, with a campaign switch to make the change reversible.

Web UA / entry journey
Ad
Explanation
Mission

One more explanation before the player could begin.

Reached the mission

35%Before67%After

People reaching the mission page
as a share of people clicking the ad

Installed the game

7.8%Before10.4%After

People installing the game
as a share of people clicking the ad

Advertiser’s target actionsEssentially flat.

Easier entry left the harder question inside the experience.

KakaoBank, 2024. Before/after comparison; other campaign changes may have affected the result. Journey illustrated.
Zoom in · Choosing a reward structure

Choosing a reward pathdid not improve conversion.

Getting people into the mission did not guarantee meaningful play. In a separate experiment, a feature called Mission Mode tested whether choosing a reward structure would build commitment. The first option dominated; few people explored the alternatives.

Original Mission Mode design asking the player to choose a reward style in a horizontal carousel Original confirmation screen listing the selected missions and warning that the choice cannot be changed
Original designs: choose a reward structure, then confirm it. Sample missions and rewards.

The test did not support
adding that decision.

Control 52.18%824 events / 1,579 installs
Mission Mode 40.59%772 events / 1,902 installs

One Android campaign, 2025: target actions per install. Results are directional; test assignment is unconfirmed.

All three campaign/platform slices declined on this measure. Some purchase counts moved the other way. The result called for investigation, not a success claim.

Zoom in · Trust and reasons to return

Progress needed a goal.Rewards needed trust.

The early feedback study exposed uncertainty about rewards. Later, the team tested a different part of engagement: giving people a nearby goal to pursue.

Original Ranking Board design showing nearby ranks, a next-rank prompt and an attendance entry
The team’s later ranking-board design, shown with sample values.

More visitors opened the game experience

January–February 2026
8%Without the board24%With the board

Share of daily visitors opening the game-rewards experience. Team comparison, 2026; rounded figures. Group sizes and test assignment remain unconfirmed.

More people opened RewardMon. That finding does not establish longer sessions, repeat visits or revenue growth.

Back to the product · Reach and investment

Growth exposed the limitsof partner inventory.

The early product work helped make the case for greater investment in 2025. By 2026, the strategy had to address a new constraint: how many more players the existing placements could reach.

Game UA revenue

2024
₩788M
2025
₩3.209B

2025 target ₩5B

Revenue grew about fourfold, but fell short of the ₩5B target. A shared product, sales and engineering outcome.

The reachable audience was narrowing.

~42kH1 2025
~32kH2 2025

Average monthly new users. The team saw the decline as a sign that existing partner placements were reaching their limit.

Reaching previous visitorsReferralsStandalone service

The 2026 strategy widened the journey beyond partner placements. These were directions to develop: the first-half review said external retargeting had not progressed as intended.

Mido, our designer, took over.My focus moved to new systems.

I handed ongoing design to Mido: mission hierarchy, progress and rewards, retention experiments and the standalone experience. My role moved toward guidance and review.

The recorded QA build's genre picker, offering six game categories and up to three choices
The team’s later test build: rewards and genre selection lead into game discovery.

My focus shifted to Buzzvil’s AI-native organization and web readiness ahead of a planned stock-market listing. The next design problem was helping more people move from intent to implementation, with a clear way to review what ships.

Continue with Buzzvil mediums

What I learned.

Access is not commitment.

Removing an entry step helped more people reach the mission. Advertiser’s target actions stayed essentially flat in that comparison.

More choice can add friction.

Mission Mode looked plausible, but all three campaign/platform slices recorded lower advertiser target actions per install.

Prove it, then give it room to grow.

My early design work helped establish the model. Mido took ownership of ongoing design as the product and its distribution strategy expanded.

Image detail

Loading full-size image…

Open original