Experiment: How Moonton measured the true impact of online video on app installs

December 2020

What we set out to test

Can brands measure the incremental impact of online video on app installs and revenue?

The background

Moonton is a Shanghai-based game developer and publisher known for popular apps such as Mobile Legends: Bang Bang (MLBB), a multiplayer online battle arena (MOBA) game. Since launching MLBB in 2016, the brand has used App campaigns to connect with gamers — first by optimizing its campaigns for installs and in-app actions, then by implementing target return on ad spend bidding.

Although MLBB has 100 million monthly active users and more than 1 billion downloads, Moonton noticed it wasn’t engaging as many first-time gamers as it wanted to. As gaming-related online video content picked up popularity, Moonton decided to use YouTube to connect with more players online while putting a new measurement strategy to the test.

How we set the experiment up

Moonton traditionally used YouTube ads to drum up excitement for its annual gaming tournaments, but it wanted to make the platform work harder when it came to increasing app installs and in-app purchases. To fuel gamers’ excitement leading up to its MLBB 515 eParty gaming tournament in Southeast Asia, Moonton launched an online video campaign to complement its existing App ads in May 2020. Then, it worked with the Google team to use Ads Data Hub for the first time. By combining its first-party data and Google’s event-level ad campaign data with the tool, Moonton pulled impression-level insights in a privacy-safe way.

The brand experimented with a variety of video formats ranging from TrueView in-stream ads to Masthead ads to engage prospective players online. Moonton further incentivized people to download MLBB by hiding a code in its creative that unlocked a free character in the game.

Moonton’s video ads encouraged viewers to download MLBB.

Minimizing other variables in its measurement strategy was one of Moonton’s top priorities, so it used Ads Data Hub to focus on the incremental impact of online video. After two weeks of running its App campaign and branding YouTube campaign simultaneously, Moonton measured impressions across three unique audience groups on more than 149 million mobile devices:

  • Group one: Served YouTube branding ads only
  • Group two: Served App ads only*
  • Group three: Served both YouTube branding ads and App ads*

*App campaigns were optimized for installs

What we learned

Brands can measure how implementing another channel in their campaigns drives conversions. When Moonton uploaded its first-party data to Ads Data Hub, including organic device IDs, it was able to assess which of its ads prospective users had interacted with on an aggregated level. With its consolidated data, the brand was able to gauge the incremental impact of its YouTube branding campaign.

By running its YouTube and App campaigns simultaneously, Moonton lifted its conversion rate by 27% while lowering cost per install by 55% compared to when it served only App ads. Moonton also increased its converted payer rate — the number of players who downloaded MLBB and made an in-app purchase after seeing a branded YouTube ad — by 10%, and it decreased its cost per in-app purchase by 34% after adding YouTube to its media mix.


Moonton was pleasantly surprised to discover its YouTube branding campaign boosted installs by 8%. The experiment and new measurement strategy inspired the brand to take its online video campaigns to the next level by increasing its investment in YouTube branding campaigns by 280% year over year. Its success story was also featured at Think Games in China.


“Adding YouTube to our media mix helped us connect with ready-to-play gamers in a budget-friendly way. We’ll continue to use online video as we work toward our goal of making MLBB the top mobile MOBA game in the world.” — Lion Cai, Marketing Director, Moonton

This case study is part of the Experiment with Google Ads program.

Experiment: How tap4fun reached gamers most likely to play and pay with automated bidding