使用集合让一切井井有条
根据您的偏好保存内容并对其进行分类。
借助 Remote Config 个性化,Halfbrick 将收入提升了 16%
简介
Halfbrick 拥有一支由 70 位完全远程工作的开发者组成的团队,
致力于开发
令人们终生难忘的游戏。他们早在 2001 年
就开始开发许可游戏,
很快就通过开发具有复古怀旧感的
原创游戏而声名鹊起。Halfbrick 专门开发
Jetpack Joyride、
八十年代风格的动作平台游戏
Dan the Man、
时下经典游戏
Fruit Ninja
(下载量现已超过 10 亿次)等热门游戏。
挑战
由于 Halfbrick 的大部分收入来自于使用
AdMob 投放广告,因此在不降低用户留存率的情况下尽量提高广告放置率
一直是一个问题。
之前,该团队运行了大量手动实验和 A/B 测试
以找到理想的广告数量,但这样做占用了大量
工程时间和资源。他们希望以较少的手动输入量
来运行更细致的实验,
从而快速轻松地优化广告放置率。
他们还希望通过在玩家尽情享受游戏的时刻
(例如在教程或检查点结束时)
显示评分来尽量提高用户互动度。这是
该团队希望自动执行和优化的另一个手动流程。
解决方案
Halfbrick 许多游戏的
后端都利用了 Firebase
产品,例如
Cloud Functions、
Cloud Storage、
Realtime Database。此外,他们还使用
Remote Config 来动态
控制和更改游戏的行为和外观,因此很快就采用了 Firebase 的最新功能,
即 Remote Config 个性化。
这项新功能可让开发者指定目标,
然后使用机器学习技术为每个用户找到较佳参数,
以尽量实现这些目标,本质上是
执行持续改进的个性化 A/B 测试。
Halfbrick 使用 Remote Config 个性化功能来优化
三个关键指标:广告频率、解锁新级别的时间、
评分提示的时间。
通过使用个性化功能来决定显示
插页式广告的时间和频率,Halfbrick 的游戏现在会根据玩家的行为和
偏好自动向
玩家显示合适数量的广告,
从而提供更好的用户体验,
并提高留存率和广告支出回报率。他们能够
将每位用户的平均广告数量从 3.4 增加到 4.2,
从而使日活用户人均创造的收入 (ARPDAU) 提高了 16%,
而且没有影响互动度、
留存率或评分。此外,个性化功能还帮助 Halfbrick 确定
让每位玩家等待解锁新级别的
合适时长,从而增加
其观看广告的总时长。
此外,Halfbrick 还使用个性化功能确定
较佳时机(即用户尽情畅玩游戏时),
以请求用户为其应用评分,
而无需团队进行任何手动设置或猜测。个性化功能帮助他们
将用户与评分提示的互动度增加了 4%,
并将积极评分的数量提高了 15%。
如果他们手动运行这些实验,
则需要由一位专门的团队成员花费一周的时间来配置、测试、分析、
发布结果。现在,他们可以在一两天内
设置 Remote Config 个性化功能,
然后让其在没有人工定期干预的情况下一直学习和优化。
这让 Halfbrick 节省了资金和时间,从而利用节省下来的
工程时间为其游戏构建新功能。
效果
16% - 日活用户人均创造的收入 (ARPDAU) 增幅
15% - 应用商店积极评分增幅
“通过 Remote Config 的个性化功能实现的精细程度是人工难以达到的。个性化让我们对于如何优化广告策略有了新的见解,甚至帮助我们验证了自己的假设,即玩家不喜欢太多广告。令人难以置信的是,我们能够提高广告展示频率并增加广告收入,而且没有导致负面评价或客户投诉量增加。”
- Miguel Pastor,Halfbrick 产品经理
[null,null,[],[],[],null,["[*arrow_back*\nExplore other case studies](/case-studies) \n[View more solutions](/solutions) \n\nHalfbrick increases revenue by 16% with Remote Config personalization \n\n*** ** * ** ***\n\nIntroduction With a team of 70 fully-remote developers,\n[Halfbrick](https://www.halfbrick.com/) strives to make games\nthat people remember for a lifetime. While they began by\ndeveloping licensed games back in 2001, they quickly made a name\nfor themselves by developing original titles with a retro,\nnostalgic feel. Halfbrick is responsible for popular titles like\n[Jetpack Joyride](https://play.google.com/store/apps/details?id=com.halfbrick.jetpackjoyride),\n80s-style action platformer\n[Dan the Man](https://play.google.com/store/apps/details?id=com.halfbrick.dantheman),\nand the instant-classic\n[Fruit Ninja](https://play.google.com/store/apps/details?id=com.halfbrick.fruitninjafree),\nwhich has now been downloaded over a billion times.\n\n*** ** * ** ***\n\nChallenge As most of Halfbrick's revenue comes from running ads with\n[AdMob](https://admob.google.com/home/), maximizing ad placement\nwithout damaging user retention is a constant concern.\nPreviously, the team ran a lot of manual experiments and A/B\ntests to find the ideal number of ads, but it took up a lot of\nengineering time and resources. They wanted a way to run more\ngranular experiments with less manual input, allowing them to\nquickly and easily optimize their ad placement.\n\nThey also wanted to maximize user engagement by showing a rating\nprompt at the moment when a player is most enjoying the game,\nlike at the end of a tutorial or checkpoint. This was another\nmanual process that the team hoped to automate and optimize. \n\n*** ** * ** ***\n\nSolution \nMany of Halfbrick's games were already powered by Firebase\nproducts such as [Cloud Functions](/products/functions),\n[Cloud Storage](/products/storage), and\n[Realtime Database](/products/realtime-database) on the\nbackend. And they were also using\n[Remote Config](/products/remote-config) to dynamically\ncontrol and change the behavior and appearance of their\ngames, so they were quick to adopt Firebase's latest feature,\n[Remote Config personalization](/docs/remote-config/personalization).\nThis new feature allows developers to specify objectives and\nthen uses machine learning to find the best parameters for\neach user to maximize those objectives, essentially\nperforming individualized, continuously-improving A/B tests.\n\nHalfbrick used Remote Config personalization to optimize\nthree key metrics: ad frequency, time to unlock a new level,\nand the timing of a ratings prompt.\n\nBy using personalization to decide when and how often to show\ninterstitial ads, Halfbrick's games now automatically show\nplayers the optimal number of ads based on their behavior and\npreferences, delivering a better user experience, and\nincreasing retention and return on ad spend. They were able to\nincrease the average number of ads per user from 3.4 to 4.2 ---\nwhich led to a 16% increase in average revenue per daily\nactive user (ARPDAU) --- without affecting engagement,\nretention, or ratings. Personalization also helped Halfbrick\ndetermine the right amount of time to make each player wait\nto unlock a new level, increasing the overall amount of time\nthey spent watching ads. \nIn addition, Halfbrick is using personalization to determine\nthe best time (i.e. when users are most enjoying the game) to\nask users to rate their app, without any manual setup or\nguesswork from the team. Personalization helped them increase\nengagement with the rating prompt by 4% and increase the\nnumber of positive ratings by 15%.\n\nWhen they ran these experiments manually, it would take a\ndedicated teammate one week to configure, test, analyze, and\nroll out the results. Now, they can set up Remote Config\npersonalization in a day or two, then allow it to keep\nlearning and optimizing without regular human intervention.\nThis has allowed Halfbrick to save money and time, using that\nengineering time to build new features for their games.\n\n*** ** * ** ***\n\nResults\n\n16% - increase in average revenue per daily active user (ARPDAU)\n\n15% - increase in positive app store ratings \n\n\"The granularity achieved with Remote Config's personalization feature is impossible for a human to instrument. Personalization has given us new insight into how we can optimize our ad strategy and even helped us challenge our own assumptions that players don't like too many ads. Incredibly, we've been able to increase the frequency of ads and grow ad revenue without seeing an uptick in negative reviews or customer complaints.\"\n\n\n- Miguel Pastor, Product Manager, Halfbrick \nTry Firebase today\n\n\nIntegrating it into your app is easy.\n[Get started](https://console.firebase.google.com/) \n\nAll Firebase products \n\nBuild\n\n- [App Check](/products/app-check)\n- [App Hosting](/products/app-hosting)\n- [Authentication](/products/auth)\n- [Cloud Functions](/products/functions)\n- [Cloud Storage](/products/storage)\n- [Data Connect](/products/data-connect)\n- [Extensions](/products/extensions)\n- [Firestore](/products/firestore)\n- [Firebase ML](/products/ml)\n- [Genkit](https://genkit.dev/)\n- [Hosting](/products/hosting)\n- [Realtime Database](/products/realtime-database)\n- [Firebase AI Logic client SDKs](/products/firebase-ai-logic)\n\n[Generative AI](/products/generative-ai) \n\nRun\n\n- [A/B Testing](/products/ab-testing)\n- [App Distribution](/products/app-distribution)\n- [Cloud Messaging](/products/cloud-messaging)\n- [Crashlytics](/products/crashlytics)\n- [Google Analytics](/products/analytics)\n- [In-App Messaging](/products/in-app-messaging)\n- [Performance Monitoring](/products/performance)\n- [Remote Config](/products/remote-config)\n- [Test Lab](/products/test-lab)"]]