来自venturebeat网站，翻译版权所有。

——为何？如何？谁？数据分析告诉你真相。我们会帮助你了解比赛场地。我们的免费营销分析圆桌会议今天注册。

       如果你使用facebook，也许曾经玩过，或者被邀请玩FarmVille，CityVille，或黑手党战争—三Zynga等最流行的游戏。众所周知，Zynga公司是社交游戏的先锋，像糖果粉碎和神庙逃亡那样会为玩家推荐设置喜好。但Zynga更是一种休闲游戏公司，通过数据与分析中心，彻底改变了数据驱动的产品开发和优化。

       因为我相信我的事业和公司可以通过使用的数据来了解用户的行为，所以我长年怀着浓厚兴趣来跟随Zynga。Zynga的产品和分析团队几年前就做了一个任何主流公司需要做的移动应用程序或网站，以保持竞争优势和推动增长。

       Zynga产品开发者从一开始就了解的用户分析的重要性。当其他的游戏公司只看基本数字时，Zynga已意识到为了产生火花般病毒扩散，用户行为需要真实地去了解本质。这种非常规的分析方法使Zynga一夜成名。

       在Zynga推出跟踪粒度级别的用户行为时，这在当时是一个相当新的概念。当时，没有任何的分析平台能够满足他们的需求，所以Zynga的工程师搭建自己的分析基础。Zynga的原分析方法启发了一些最聪明的开发人员将他们从Zynga的经验，以同样的方式，超越了社交游戏行业，建立技术工艺的分析模型。

       在这里，你可以从Zynga学到与农场动物或怪物无关的经验。

1。发布前创建分析模型，以衡量预期

       发行任何Zynga游戏前，团队建立游戏的性能分析模型。该模型研究的因素包括如在产品内置病毒钩和用户信息获取渠道。从获取数据中，模型尝试预测关键指标，包括每天新安装的数量，随着时间的推移数量是怎么衰减的，及病毒式传播的K值，保留一天数量，每日活跃用户的收入。

       团队在使用这些模型获取数据的基础上，了解如何设计以使游戏最大程序的增长，用户参与和收入增加。这些模型还提供了一个比较基准，以便尽快进行游戏启动，该小组可以快速判断是否符合预期目标。

2。不要追逐短期收益，以致牺牲持久增长
       太局限于短期收益将使你的长远收益受损，以Zynga为例。一个典型的例子是它的“闪购”活动。在Zynga的游戏中，有一款限期特价虚拟商品。Zynga第一次销售时，非常成功。收入在那一天达到巅峰。

       在短期内，数据显示，这种销售方式是收入的一个动力—更多的用户转换为买家。但最终，事与愿违。用户期望特价活动，而Zynga发现自身越来越依赖更大更频繁的销售活动以让用户购买更多的虚拟商品。长期来看，这种销售策略是一个纯负面的，因为用户拥有的货币比他们所能消费的多，因而销售变得越来越少。

3。数据不应是你的一切

       Zynga从不通过其游戏内容来登上社交游戏王位。事实上，Zynga前副总裁Ken Rudin曾分析说过：“Zynga是乔装成游戏公司的分析公司。”

       虽然最初Zynga对指标度量的关注引领了极为成功的增长和收入，一些Zynga校友也认为公司是极数据驱动的。闪光的销售就是一个是如何看待纯粹数字导致错误决定的例子。
       前Zynga产品经理和现火箭游戏联合创始人，Niko Vuori，最近讲述了我，Zynga的激光聚焦度量可能已经是一些错失的原因，它导致了错过一些正确的，但更难以衡量，但同样重要的东西，例如提高游戏的整体可用性。“最终的结果是人们非常关注的数据但并没有花足够的时间来关注提高游戏质量”，Vuori说。“最主要的事情，我想很多人已经从Zynga认为数据是有它的位置，它可以帮助你做出正确的决定，但你应该仍然是开放地做不同的驱动性的东西。数据不应该统治你的世界。”
      Roy Sehgal，是一位Zynga早期的副总裁和总经理，现在是一个在我的公司的投资者，告诉我在进入数据前，有好的假设是多么的重要。“数据驱动是一个负载的术语，”他说。“我相信你需要假设驱动和使用数据来验证你的假设有效（或无效）。这些数据确定了你的假设是正确的还是错误的，并强调了潜在的改进用户体验的领域。”
      虽然他们犯了一些错误，但Zynga的数据驱动文化可以规范化社交游戏，同时引入全球数据驱动的产品开发和优化。如今，通过互联网和手机获取的如此巨量的数据，各行各业的公司都比以往任何时候更能把用户数据转换成他们最强大的商业工具。
      Spenser Skates和联合创始人Curtis Liu在2012建立了基于移动事件分析的公司Amplitude。在此之前，他创立的Android应用程序的文本和语音sonalight担任算法交易者DRW交易集团。

Lessons from Zynga: Data is essential, but it shouldn’t rule your world

——Why? How? Who? Analytics tell the story. We’ll help you understand the playing field.Register today for our free marketing analytics roundtable.

    If you’re on Facebook, chances are you’ve either played or been invited to FarmVille, CityVille, or Mafia Wars — three of Zynga’s most popular games. Zynga pioneered social gaming as we know it today, setting the tone for successors like Candy Crush and Temple Run. Zynga is much more than a casual gaming company though – it’s a data and analytics powerhouse that has revolutionized data-driven product development and optimization.

    Because I’ve bet my career and my company on using data to understand user behavior, I’ve followed Zynga with great interest for years. What its product and analytics team did years ago is what any mainstream company with a mobile app or web site needs to do today to maintain a competitive edge and drive growth.

    Zynga product developers understood the importance of robust user analytics from the beginning. While other gaming companies were only looking at basic counters, Zynga recognized that it was user behavior that really needed to be understood in order to spark viral engagement. This unconventional approach to analytics helped catapult Zynga from zero to hero almost overnight.

    Tracking user behavior at a granular level was a fairly new concept when Zynga launched. At the time, there weren’t any analytics platforms that met their needs, so Zynga’s engineers built their own analytics infrastructure. Zynga’s original approach to analytics has inspired some of the brightest minds in technology to craft their analytics models in the same way, applying lessons they took from Zynga to industries beyond social gaming.

    Here’s what you can learn from Zynga that has nothing to do with farm animals or the mob.

1. Create analytical models before launch to measure expectations

    Before launching any Zynga game, the team built an analytical model for the game’s performance. The model examined factors like the viral hooks built into the product and the user acquisition channels. From this data, the model attempted to predict key metrics, including the number of new installs per day and how that would decay over time, virality K-factor over time, Day 1 retention over time, and revenue per daily active user.

    Teams used these models as a basis for understanding how to engineer the game for maximum growth, engagement, and revenue. These models also provided a baseline for comparison, so that as soon as the game launched, the team could quickly gauge whether or not it was on track to meet expectations.

2. Don’t chase short-term gains at the expense of durable growth

    Having too narrow a focus on short-term gains can cost you long term revenue, as was the case with Zynga. A prime example of this was its ‘flash sales’ campaign. In Zynga games, there are virtual goods for a set price. The first time Zynga ran a sale on these, it was immensely successful. Revenue on that day shot through the roof.

    In the short term, the data suggested that sales were a great driver of revenue — more users were converting to buyers. But eventually, the strategy backfired. Users came to expect the sales, and Zynga found itself running bigger sales more frequently to get users to buy more virtual currency. Over the long term, the sales strategy was a net negative, as users had more currency than they could spend, and sales became less and less effective.

3. Data shouldn’t rule your world

    Zynga didn’t ascend the social gaming throne through the content of its games. In fact, Zynga’s former VP of Analytics, Ken Rudin, is famously quoted as saying that Zynga was “an analytics company masquerading as a games company.”

    Although Zynga’s focus on metrics led to extremely successful growth and revenue initially, some Zynga alumni think the company may have been too data-driven. The flash sales backfire is one example of how looking purely at the numbers led to the wrong decision.

    Former Zynga product manager and current cofounder of Rocket Games, Niko Vuori, recently recounted to me that Zynga’s laser focus on metrics may have been one of the main reasons it missed out on things that are more difficult to measure, though just as crucial, like improving the overall usability of a game. “What ended up happening is people were exceptionally focused on the data and didn’t spend enough time looking at the qualitative gameplay,” said Vuori. “The main thing I think a lot of us have taken from Zynga is that data has its place, it helps you make decisions, but you should still be open to doing things that are different, that are gut-driven. Data should not rule your world.”

    Roy Sehgal, an early vice president and general manager at Zynga who is now an investor in my company, told me how important it is to have good hypotheses before diving into the data. “Data-driven is a loaded term,” he said. “I believe you need to be hypothesis-driven and use data to validate (or invalidate) your hypotheses. The data identifies where your hypotheses were right or wrong and highlights areas of potential improvement to the user experience.”

    Although they made some mistakes along the way, Zynga’s metrics-driven culture allowed it to standardize social gaming while simultaneously introducing the world to data-driven product development and optimization. Today, with so much data being captured through the web and on mobile, companies in every industry are more capable than ever to turn user data into their most powerful business tool.

    Spenser Skates founded mobile event-based analytics startup Amplitude in 2012 with cofounder Curtis Liu. Prior to that he founded text-by-voice Android app Sonalight and worked as a Algorithmic Trader for DRW Trading Group.