开发商谈手游测试发行过程的关键因素

原作者：Guest Author 译者：Vivian Xue

每个开发商都将面临把产品介绍给未来用户的时刻。测试发行（soft launch）是对你的应用软件进行的第一次检验，你可以从中挖掘软件的潜能、了解软件的优缺点。

本文中，Zorka.Mobi的产品经理Alexey Adamovich和Nitro Games的市场营销总监Mikko Kähärä分享了他们关于测试发行的经验以及一些实用的建议。

开发商可能拥有无数的设备和测试仪器，但是真实用户的行为往往是专业人士所无法预料的。要知道用户可能会在屏幕的同一块区域或者某一个按钮上点击100次，此时一个“准备就绪的软件”也可能会崩溃。

此外，测试发行还能够帮助你测试软件的盈利模式、注册途径，改进用户体验等等。

用户v.s.外包

让我们现实一点：并非所有的开发商都能负担得起一个成熟的QA团队，然而外包的价格又太过高昂。简单地说，用户全身心地投入到测试中并且深度沉浸在应用里。因此，技术性的测试发行可以称作是“QA发行”。

另外，外包团队无法像开发商一样清楚了解产品并且掌握着有限的资源比如产品迭代数量、测试周期等等。

产品技术测测试发行的国家中，流量成本较低是印度、巴基斯坦和伊朗。在这些国家，消费价格指数是0.05美金，而1000份安装量已经足够测试软件的稳定性。这意味着你差不多花个50美金就能正式发行了。

然而，对于那些本地化的应用软件来说这有点困难。因为它们属于某个特定市场，在测试发行时不能只考虑低流量成本。否则用户将完全无法理解产品而你所获得反馈就是一堆垃圾。此外，这可以帮助你避免购买一些不必要的测试设备。

当你在安卓平台上测试发行时，最好只选择旗舰设备。一般来说，你需要10款顶级的安卓设备。这将帮助你获得最有价值的反馈并消除硬件问题。你最好立刻把开发者面板中的旧设备解除掉或者减少它们的数量，并且如果你希望在购买流量时减少消极和不真实的反馈，你应该明确列出必要的限制条件。

Candy Crush Soda Saga(from pocketgamer.biz)

所有的这些都适用于起步阶段的非游戏类应用。对于游戏来说，测试发行要更困难一点。由于其中的反馈和软件指标更多更复杂，它需要游戏设计者、分析师和盈利专家共同协作。

Alexey Adamovich：“测试发行”对你来说意味着什么？

Mikko Kähärä：测试发行是我们将游戏介绍给一部分用户，以此校验和测试游戏体验的积极性的一个阶段。

我们紧密跟踪游戏的性能参数、用户反馈、营销效率并进行迭代。目的是在正式发行后带给玩家尽可能完美的体验，并确保我们获得良好的销售成果。

Alexey Adamovich：你们怎么选择测测试发行的国家？

Mikko Kähärä：我们根据游戏的核心玩家和商业模式来选择一系列国家，评估它们的发展空间和营销效率。

自我们专注于制作同步多人游戏以来，我们在一些同时在线人数多、用户获取成本合理的国家进行测试。这帮助我们检验了游戏后端功能的扩展性。

Alexey Adamovich：加拿大和澳大利亚是最热门的发行地，这是真的吗？

Mikko Kähärä：从行业的总体情况来看，确实是这样。这两个国家是西方市场的典型代表。然而由于发行商通常选择在这两个市场测试游戏，情况可能发生了一些变化。作为测试发行的固定试验区，那里的用户的行为表现受到了影响，尤其是那些早期用户。

Alexey Adamovich：在测试发行之前，你们会确定每个渠道的安装成本吗？

Mikko Kähärä：在测试发行期间，我们测试了不同的用户获取策略、营销类型和不同的市场活动优化等，目的是达到一个可接受的CPI水平——虽然是在受性能约束条件下。

Alexey Adamovich：只通过应用商店搜索吸引客户，而不在社交媒体上买流量，这是一个好主意吗？

Mikko Kähärä：如果这样做效果更好，为何不呢？尽管这不太可能发生。搜索流量确实非常有效率的，但是我认为它无法提供100%的流量。搜索缺乏营销活动和基于价值的优化，它能够大大提升效率。

Alexey Adamovich：在买流量的时候，你们会选择一个狭窄的群体还是广泛的群体？

Mikko Kähärä： 其实二者都有。并且我们有效利用了相似的群体。比如在Facebook上，我们利用他们的算法，在表现优秀的玩家的基础上寻找类似的玩家群体。

Alexey Adamovich：你认为在测测试发行中更重要的是什么：测试架构还是检验游戏机制和盈利性能？

Mikko Kähärä：游戏机制和盈利性能。主要集中在估计留存率和终身价值，以及测试最佳用户体验。在进入测试发行阶段之前必须修复完所有大Bug。

Alexey Adamovich：你能列出5个你在测试发行阶段最看重的指标吗？

Mikko Kähärä：留存率，日活跃用户的平均收入、日每用户平均收入，评级以及关键的营销指标（CPI每次安装成本、CPE每次参与收费等）。

Alexey Adamovich：你们曾经在测试发行过后减少游戏内的onboarding吗？

Mikko Kähärä：Onboarding /首次用户体验是一个持续迭代和发展的关键过程。它的可以增加也可以减少。Onboarding的目的是为了获得反馈，而不是技术工作量。（Onboarding这个概念来自人力资源管理领域。Onboarding过程包含三个步骤：适应、吸收、加速。要让玩家适应、熟悉游戏进程，就得先给他们提供一些必要的道具、资源等游戏设置机制，给他们创造一个良好的开端，游戏邦注）
Alexey Adamovich：游戏测试发行阶段提供多少内容合适？

Mikko Kähärä：游戏的核心循环是必须的，再加上一些metagame元素和盈利机制就够了

本文由游戏邦编译，转载请注明来源，或咨询微信zhengjintiao

Every developer comes to a point when it’s time to introduce the product to future users. A soft launch is a first trial for your app when you can test its potential and reveal its strong and weak points.

In this article, Zorka.Mobi product manager Alexey Adamovich and Nitro Games chief marketing officer Mikko Kähärä share their experience and practical tips about soft-launches.

Developers may have limitless amounts of devices and testers, but real users often do things that professionals won’t even think of. It’s true that people can click a hundred times on the same place in the interface or on a button, and at this point, a “completely-ready application” may crash.

Moreover, a soft launch helps to test several in-app solutions such monetisation models, registration type, UX tweaks, etcetera.

Users versus outsourcing

Let’s be realistic: not all developers can afford to have a fully-fledged QA team, while outsourcing this job may cost a fortune. To put it simply, users put their heart into testing and immerse themselves deeper into the app. Thus, a technical soft launch can also be called a “QA launch”.

Additionally, outsourcing providers don’t know the product as well as the developer and have limited resources like number of iterations, testing period, etcetera.

The best countries for a technical soft launch with cheap traffic are India, Pakistan and Iran. In this case, you get $0.05 average CPI rate, while 1,000 installs are enough to check the stability of the application. It means you can run this campaign for as much as $50.

However, it becomes a bit more difficult with localised applications. They shouldn’t be soft-launched with cheap traffic, because they belong to a specific market. Users won’t understand a thing and you’ll get rubbish instead of feedback. Besides, this will help to avoid unnecessary costs for purchasing test devices.

When conducting a soft launch on Android, it’s a good idea to go with flagship devices only. Normally, you will need around 10 top models. This approach will bring the most relevant feedback and eliminate hardware problems. You’d better immediately disable or limit older models in the developer panel, and if you want to reduce the share of negative and incorrect feedback when buying traffic, you should specify all the necessary restrictions.

All this applies to non-game applications in the first place. Soft-launching games is a bit harder. It requires game designers, analysts and monetisation experts, because in this scenario you get much more feedback and app metrics.

Alexey Adamovich: What does the term ‘soft launch’ mean for you?

Mikko Kähärä: Soft launch is a period when we introduce a title to a limited audience for verifying and testing the positive gameplay experience.

We follow game performance data, customer feedback and marketing efficiency closely and iterate development. The goal is to offer best possible experience to the players once the title is fully released and also make sure that we have a good business case with it.

-How do you choose a country for soft launch?

We choose a number of countries based on the title’s core audience and business model to gauge scalability and marketing efficiency.

The goal is to offer best possible experience to the players once the title is fully released and also make sure that we have a good business case with it.MIKKO KÄHÄRÄ
Since we focus on synchronous multiplayer titles, we also run tests in countries where we can acquire a high number of simultaneous players affordably. This also helps us test the scalability of our backend functionalities.

-Is it true that Canada and Australia are the most popular options?

In the industry in general, this has been the case. They represent the wider Western markets quite well. Since it is very common for publishers to test their games in these markets, it might however tweak the big picture. Being a constant guinea pig with soft launch releases might affect the overall behaviour of especially early adopters.

-Do you fix the install cost for each source before the soft launch?

We test with different kinds of UA tactics, campaign types, varied event optimisations, etcetera during soft launch with variable rates, although aiming for an acceptable level of CPI in the constraints of performance.

-Is it a good idea to stop buying traffic from social networks and attract users only through search traffic from the app stores?

If it gets the job done better, why not. This would be quite unlikely, though. Search traffic can be very efficient, but I don’t see a scenario where it would provide 100 per cent of traffic. Search lacks event and value-based optimisation, which tends to be quite efficient.

-When buying traffic, do you focus on a targeting a narrow or wider audience?

It’s a mix of these. Also we do utilise lookalike audiences quite effectively. On Facebook, this means that we can create similar audiences based on well performing player cohorts, utilising their algorithms.

-What do you think is more important in soft launch: to test the build or to check game mechanics and monetisation performance?

Game mechanics and monetisation performance. Mainly focus on retention and life-time value estimations, along with testing the optimal user experience. Big bugs need to fixed before going into soft launch.

-Can you name five key metrics that you outline during the soft launch?

Retention, ARPDAU, Day-x ARPU, ratings and key marketing metrics (CPIs, CPEs etcetera).

-Have you ever had to cut an in-game onboarding after the soft launch?

Onboarding / FTUE is a key process that is continuously iterated and developed. It can be increased or reduced. Onboarding is about getting feedback and not about the amount of technical work.

-How much game content is enough to soft launch a title?

The core loop needs to be clearly on board, along with sufficient metagame elements and monetisation.（source：Pocketgamer  ）