作者：Roy Hwang

四个月以前，因为无知而胆大，我为自己树立了一个目标——成为世界上最优秀的游戏分析师之一。还笑？如果我告诉你我认为游戏机制是游戏开发者讽刺自己的术语呢？或者DAU（日活跃用户）听起来更像新的中国哲学学说而不是基本指标？可以说，我距离目标还有很长的路要走。所以我像大部分人那样做了——我Google了很多资料。当我学习完Google来的东西后，我就买了书。我读完书后，我又去向行业专家讨教。正当我开始变得骄傲自大时，我受到了封闭测试和软发布的考验，正是那期间，我发现我的40%的学习时间本可以更好地利用。所以我决定写本文，为此我想出以下几个标题：

简单9步成为游戏分析师！

9件事助你成为游戏分析大师！

一月速成王牌游戏分析师！

如果以上标题让你觉得热血沸腾，那么做好被泼冷水的准备吧！本文不是教程。你不会在下文中读到关于游戏分析的方方面面的建议。我之所以写本文，是因为我希望在我决定成为游戏分析师的时候，我能首先读到这样的文章。借本文，我想简要地介绍一下成为一个合格的游戏分析师的过程。

game analytics(from gameanalytics.com)

1、从单一的游戏类型入手

游戏类型有不同的定义，比如从平台的角度说，有PC、游戏机、手机等；从机制的角度说，有匹配消除类、MMO、模拟等；从种类来说，有动作类、冒险类和赛车类等。这些类型是可以互相结合的。当选择游戏类型时，应确保它在市场上有足够的占有率。至少，这种游戏类型有排行榜的前100名中有比较体面的名次。挑选最初的游戏类型并不意味着你这一辈子都要研究那种类型。它只是作为一个起点。你应该知道收益递减法则，如果你一开始就研究多种游戏类型，你积累下来的知识会越来越少。与其艰难地把游戏机FPS和休闲免费手机游戏的共性特性归纳清楚，不如一开始就好好研究一种游戏。

2、阅读El-Nasr、Drachen和Canossa写的《游戏分析学》

每个游戏设计师都必须有一个起点，这本书就是一个好开头。本文知识丰富、角度全面、由浅入深。虽然本书的长度过长（超过800页）、语调不一致（包含50多种不同的视角）和技术参考过时，但它包含相当实用的信息。当你还没有任何游戏分析的基础，这些海量信息是非常值得学习的。

3、认识行业专家

如果你完成了上面两步，那么接下来你可以使用搜索引擎了。在搜索栏中键入指标和行业热词，你应该会看到大量文章和网站。如果你觉得自己运气很好，那就在搜索栏中键入某款游戏名称和“分析”两个字。在你过滤掉无用信息后，结果应该主要关于行业研究或发现，你会从个人和游戏服务商发布的研究中发现有价值的东西。

与书籍推荐一样，专家的分析或观点通常（必然地）与商业利益有关。总是注意这些信息的语境，尽量搜索反对意见。在游戏分析经济中，知识仍然是金，所以在专家的博客和网站中表达感谢。一旦你订阅了专家的网页内容，就可以跟进专家的Twitter了，社交媒体改变了信息传播的方式，所以请确保你使用了所有可用的工具。

4、正确认识指标

到这时，你应该已经阅读了非常多关于游戏指标的东西，它们甚至已经成为你的日常交谈话题。事实上，我每天都花五分钟时间讨论把大额玩家叫作“鲸鱼玩家”是否更可爱，如果我们不考虑它就是ARPU的话。这五分钟我永远也拿不回来。

在大部分的分析会议上，指标都被作为谈话的起点和终点。对于所有关于乐趣因素和创新设计的演讲，游戏最终是由它们的底线数字判断和衡量的。你是否开发了一款好游戏？你怎么肯定你的游戏很好？你妈的认可不算数。通常来说，指标是微妙或细微的问题的指示器。指标告诉你方向，然后由分析师深入细节并找出问题。这就是为什么存在游戏分析师这种职业。知道指标不算什么，它只是一个最基本的要求。

但注意，并不是靠指标就能产生《Candy Crush》。有些人认为使用指标就能做出操纵玩家花钱的山寨游戏而不是给玩家带来乐趣的好游戏。过分使用数据当然会产生糟糕的结果，但我和其他许多人认为，不使用数据同样是很危险的。最终，大部分业内人士认为，数据分析型设计（与数据导向型设计相反）能够帮助开发者做出既有趣又赚钱的游戏。

5、基本技能要求

从SQL和Excel开始。SQL是主要的数据库语言，尽管不同平台上有很多变体，但基本面是一样的。SQL帮助你检索、格式化、组织和操作来自最传统的数据储存区的数据。Excel当然是世界上使用最广泛的数据分析工具，它有非常强大的表格功能。记住，这些只是基本技能。分析数据的技术和工具是令人惊喜的，且总是在变化，所以一两年后，你会发现自己只会操作表格或编写SQL查询，你会面临极大的竞争劣势。

6、使用清单列表

使用清单的前提很简单：你越少依靠记忆力越好。当判断游戏表现时，许多分析师会使用他们已经掌握的知识开始分析数据。如何积累已经学会的知识？通过重复背诵指标、维度和技术分析数据。但无论你的记忆力多好，要记下所有东西仍然是困难的，何况可能会发生记忆错误。

清单可以减少记忆压力。它们以简明的方式锁在分析思考中，使过程更可重复，并体现错误—-允许你专注于思考新挑战和新问题。如果你想学习本文，就使用清单列表吧。它们都是很有价值的。

7、循序渐进地学习困难的新技能

如颗你想成为分析师，你要学习的困难的技能非常多（游戏邦注：例如Python、emcache、R、Qlikview/Tableau,、nformatic等等）。学习困难的新技能的问题是，非常费神，通常需要你付出巨大的努力。但不要着急，多给自己一些时间和空间去磨练你的技能。

8、拒绝简单化

在媒体的号召下和苹果的极简主义设计风格的影响下，人们对简单化或流线化的渴望达到了前所未有的程度。游戏分析学也随之越来越追求在文章中简单地解释自身，以至于单一的数字成为衡量成功的主要标准。但说到底，各个数字取决于大量其他因素的相互作用，达到一两个基准点并不能保证游戏成功。甚至看似关系密切的第2天和第30天留存率也并不总是相关的。相反地，你应该衡量尽可能多的方面（在不特征游戏性能或超出预算的情况下）。设计数据结构使之提供最佳表现和最大的灵活性——不只是最容易读懂。从多个角度评估游戏表现——不只是人人都使用的指标。UI和可视化设计应该追求简单化，但其他方面如数据库和统计算法未定是抽象的。所以，拥抱复杂吧。

9、找到解决方案的能力

作为分析师，你的工作不只是观察数字和每天、每周、每月做报告。这些工作有电脑负责。你的工作是找出导致这些数字出现的原因，以及推论出改进的可能方法。优秀的分析师是游戏工作室的积极鉴别者。设想一下数据和分析学能有什么作用，然后去实现这些作用。

最后，享受成为游戏分析师的过程。毕竟，我们的工作是几乎能够实时得到数百万玩家验证的。偶尔停下来问问自己：“还能比这个更好一点吗？”（本文为游戏邦/gamerboom.com编译，拒绝任何不保留版权的转载，如需转载请联系：游戏邦）

Becoming a Games Analyst

by Roy Hwang

Four months ago, emboldened by a complete lack of knowledge, I set a goal to become one of the world’s best game analysts.  Laughing yet?  How about if I told you I thought game mechanic was an ironic title that game developers had given themselves?  Or that DAU sounded more like a new Chinese philosophy than a baseline metric obsession?  Suffice it to say, I had a long way to go.  So I did what most people do — I Googled.  A lot.  And when I ran out of stuff to Google, I bought books.  And after I read those books, I asked questions to industry experts.  And just as I was starting to get cocky, I got baptized by the fire of closed betas and soft launches during which I figured out that 40 percent of my learning time could have been better spent.  So I decided to write this post.  Consider the following titles:

From Zero to (Analytics) Hero in Nine Easy Steps!

Nine things you can do today to start Dominating Games Analysis!

How to be a killer Games Analyst in One Month!

If any of those sounds good to you, prepare to be disappointed!  This post is not a tutorial.  You will not read the next nine recommendations and know everything about games analysis.  This post is the post I wish I had first read when jumping into games analysis.  It’s a starting bundle that outlines the process of becoming a viable games analyst.

Pick a game type

Game type definitions vary from distribution platforms (e.g. PC, Console, Mobile) to mechanics (e.g. Match-3, MMOs, Slow Sims) to genres (e.g. Action, Adventure, Racing) to some combination of the three.  When selecting a game type, make sure it has enough market share to justify the selection.  At a minimum, the game type should grace a well-known top 100 grossing list (or two).  Picking an initial type doesn’t mean you’ll have to spend your career analyzing that game type.  Picking a type means you’ll be setting a clear starting point.  Be aware of the law of steeply diminishing returns when studying outside one area focus.  There is value in studying multiple game types – just less so at the onset.  You can spend your initial time better than trying to stitch together insights gleaned from console first person shooters and casual freemium mobile games.

Read Game Analytics by El-Nasr, Drachen, and Canossa

Every game analyst has to start somewhere, and this book is as good a place as any to begin.  The book is informative, broad in perspectives, and accessible on introductory and complex topics alike.  While critics of the book would probably comment on the book’s length (800+ pages), tonal inconsistency (it contains over 50 different point of views), and outdated vendor and technology references, the book contains really useful and practical information.  When you are starting from zero, this amount of information isn’t a detraction but an opportunity to learn.

Identify and follow the experts

If you have completed the two steps above, it’s time to hit the search engines.  A combination of metrics and industry buzzwords should return a host of articles and sites.  If you’re feeling lucky, try typing a specific hit game title and “analysis” into the search bar.  The information will consist mainly of re-treads of the same industry studies or findings but once you sort through the noise, you’ll discover valuable information from individual contributors and gaming service vendors.

Like book recommendations, expert analysis or opinion is usually tied (by necessity) to commercial interests.  Always be mindful of the context in which the information is given and search for dissenting opinions whenever possible.  Still, knowledge is gold in the game analyst economy so express virtual thanks and throw any expert blogs and sites into a RSS reader such as Feedly.  Once you have subscribed to the experts’ web content, take your cyber-stalking to the next level.Follow the experts on Twitter.  Social media has changed the way information is delivered so be sure to use all the tools you can in your research.

Know your Metrics

By this point, you’ll have read so much about game metrics that they’ll color your daily interactions.  I actually spent five minutes of my life debating whether or not Shamu the whale would have been perceived as cuter if SeaWorld had named him ARPU instead.  I’ll never get those five minutes back.

Metrics serve as the conversational starting and ending point for most analysis meetings.  For all the talk of fun factor and innovative design, games are ultimately measured and judged by their bottom-line numbers.  Did you develop an amazing game?  How do you know for sure?  Your mom’s approval doesn’t count. Often, a metric is a top line indicator of more subtle or nuanced problems.  The metric points you in a direction and then it is up to the analysts to dive into the detail and figure stuff out.  It’s why we get to have jobs. Knowing the metrics is not a differentiator but it’s a baseline requirement.

But be warned, not all is Candy Crush dust in metrics-land.  Some people believe the use of metrics spawns derivative, unoriginal games geared more towards manipulating users into spending money than making the game fun.  The extreme use of data can certainly produce terrible results but I and many others would argue that the non-use of data is equally if not more hazardous.  Ultimately, most industry people believe data-informed design (as opposed to data-driven design) is something with the capacity to make games both more enjoyable and more profitable.

Acquire the base skills

Start with SQL and Excel.  SQL is the predominant database query language and although there are variations across platforms, the fundamentals are fairly consistent.  SQL allows you to retrieve, format, organize, and manipulate data from most traditional data stores.  Excel, of course, is the most widely used data analysis tool in the world and provides a spreadsheet-centric way of viewing, presenting, and acting on data.  Remember that these are just the starting base skills.  The technology and tools used to derive data insights are amazing and always changing so if after a year or two you find yourself only manipulating spreadsheets or just writing SQL queries, you will be working at a significant competitive disadvantage.

Use checklists

The premise of checklists is simple: The less you rely on your memory the better.  When diagnosing game performance, many analysts rely on their learned knowledge to start exploring the data.  How did it become learned knowledge?  Through repeated memorization of the metrics, dimensions, and techniques required to analyze the data.  But no matter how good your memory is, trying to remember everything is hard and people make mistakes.

Checklists will save you the stress of trying to remember everything.  They lock in analytical thought in a concise manner, make process more repeatable, and prevent mistakes — allowing you to focus brainpower on new challenges and new questions.  If you take away nothing else from this post, please use checklists.  They are really wonderful things.

Slow down on learning new hard skills

If you want to do your job well as an analyst, there will be an overwhelming amount of hard skills you want to learn (e.g. Python, Memcache, R, Qlikview/Tableau, Informatica).  The problem with learning new hard skills is that it is mentally taxing and often requires mental energy you cannot afford to spend while in a high information consumption period.  Don’t be afraid to give yourself ample time and space to grow true hard skill competency and lengthen your learning curve.

Be the enemy of simplicity

Between our sound bite driven media and Apple’s minimalist design ethos, the cultural desire to simplify or streamline has never been stronger.  In games analysis, the growing desire for simplicity manifests itself in articles that purport a singular number to be the principal component of success.  Ultimately, each number is dependent on a whole host of other factors and hitting a goal benchmark or two does not constitute game success.  Even seemingly obvious connections such as 2nd and 30th day retention do not always correlate.  Instead, measure as many aspects as you possibly can (without sacrificing game performance or blowing out the analytics budget).  Design data structures to provide the best performance and most flexibility – not just to be the simplest to read.  Approach game performance from multiple angles – not just the top-down metrics approach that everyone else uses.  Simplicity is great for user interfaces and visualizations but some areas such as data warehouses and statistical algorithms are necessarily abstruse.  Embrace the complexity.

Have a vision

Your job as an analyst is not just to monitor numbers and provide daily, weekly, and monthly reports.  We have computers for that.  Your job is to figure out the reasons behind the numbers and deduce potential avenues for improvement.  Good analysts are competitive differentiators to their respective game studios.  Have a vision on what you think data and analytics can do and then go after it.

Last but not least, try to have a little fun as you embrace the process.  After all, we get to be in an industry where millions of users can validate a hypothesis in near real-time.  At some point, stop and ask yourself, “How could it get any cooler than that?”(source:gamasutra)