游戏邦在:
杂志专栏:
gamerboom.com订阅到鲜果订阅到抓虾google reader订阅到有道订阅到QQ邮箱订阅到帮看

通过《Polymer》数据分析苹果推荐的重要性

发布时间:2012-10-30 15:07:13 Tags:,,,,

作者:Whitaker Blackall

因为懒惰以及缺少时间,我已经拖延了一篇关于《Polymer》的稿子很长时间,但是正因为如此我们才可以看到,这几个月来围绕着苹果的推荐以及各种更新内容的出现,游戏排名,销量以及分析均发生了巨大的变化。下文的图表是我在几周前开始写这篇文章时所获取的,尽管可能较旧了,但是因为在那之后也未发生任何特殊的变化,所以还是一样具有功效的。让我们开始从这些图表中寻找端倪吧!

Polymer(from phonearena.com)

Polymer(from phonearena.com)

访问量

我之所以从访问量开始分析是因为我对此印象深刻。《Polymer》的访问量已经达到了20多万次,这一数值让我非常惊讶。这张图像便明确地呈现出本篇文 章的讨论焦点:苹果的推荐非常有帮助!当然了,在Touch Arcade以及148 Apps等平台上取得的积极评价也对这款游戏的发展带来的巨大的帮助。但是我仍然坚信苹果的推荐是推动《Polymer》获得最高销量的最关键因素。

sessions(from gamesbrief)

sessions(from gamesbrief)

这张图表中所突显的两次高峰时期便是游戏得到苹果推荐的时候。第一次高峰开始于4月26日,也就是《Polymer》刚刚发行。那时候游戏得到了iPhone App Store的推荐(被划分到“值得注意的新游戏”这一类别中)。但是真正让人惊讶的是这一类别并不是推荐页面中最显著的区块。为了找到这一推荐,用户必须先找到“游戏”标签,但是通常情况下这一标签也并不明显。所以我能够想象如果游戏出现在App Store推荐首页将获得多么“可怕的”销量。

游戏推荐标签

第二次高峰是在《Polymer》的iPad版本发行之时。这一次我们不仅再次获得了iPhone App Store的推荐(还是在相同的类别中),同时我们还登上了iPad的主推荐页面,即用户只需要在App Store首页点击游戏图标便可以进入下载页面。尽管我知道在这里我们不可能获得与iPhone App Store一样出色的流量,但是结果还是大大出乎我们的意料(特别是这次版本的游戏售价为1.99美元而不是之前的0.99美元)。最后一点让我吃惊的便是,尽管在iPad版本发行时我们只获得一半的销量(与之前的版本相比),但是在发行后又陆陆续续出现了小幅度的攀升时期,也就意味着许多最初的购买者在之后还会不时地打开游戏进行反复体验。

新用户

new users(from gamesbrief)

new users(from gamesbrief)

这一图表与上一张图表没有多大区别,但是比起游戏的访问量,它向我们呈现出了用户共使用多少不同的设备去打开这款游戏。重要的是,既然它是用于衡量设备,所以便不能确切地呈现出所有新用户的数量。实际上,许多已经在手机上安装了这款游戏的玩家也再次选择在iPad上安装了其更新内容。如果7月16日的新用户真的比4月23日多出了3325名,那么游戏定也取得了非常高的销量。

访问时间长度的分布

Session Length Distribution(from gamesbrief)

Session Length Distribution(from gamesbrief)

这张图表非常有趣。首先,我发现玩家打开《Polymer》不到3秒钟的次数为10000次。就我本身而言,如果我打开一款应用并迅速关掉它,那便只是为了摆脱烦人的推送通知,但是我却从不会发送任何推送信息。我猜这些打开游戏不到3秒钟的玩家应该是在打开游戏后收到一封短信,所以先切换到短信页面然后再回到游戏中。

有趣的是访问时间为3至10分钟的的人数最多。尽管有可能更多玩家是倾向于3分钟,但是这一结果仍然让我非常欣喜,因为这至少超越了1至3分钟的范围。更了不起的是,待在游戏中长达30分钟以上的玩家的游戏次数甚至高达10000次了!太不可思议了!我自己便从未玩过这么长时间的游戏,而看到这么多人如此沉浸于我的游戏中倒是让我兴奋不已。

用户访问信息

User Session Info(from gamesbrief)

User Session Info(from gamesbrief)

共有7180名玩家在打开一次《Polymer》后便不再接触这款游戏了。如此便有可能给其他玩家留下不好的印象!虽然一天玩多次游戏的玩家数量为11048,但是在这天结束后他们也未曾再碰触游戏(因为“一次访问用户”也包含在这数值中,这意味着共有11,048-7,180=3,868名用户在一天内打开多次游戏,但是在此之后便彻底离开了游戏)。而最让我欣慰的还是有8338名安装了游戏后并在之后继续回到游戏中的玩家。真心感谢这些人!除此之外我还需要强调道一点是:就像我之前所说的,“用户”实际上代表着不同的设备。所以如果有个玩家在自己的iPhone上玩了1000次《Polymer》,但是只在iPad上玩过1次,那么他在iPad平台上的游戏体验便可以被当成是“一次访问用户”或者“一天活跃用户”。所以我非常清楚这一数值具有严重的倾斜性。

事件总结

Event Summaries(from gamesbrief)

Event Summaries(from gamesbrief)

首先,我发现玩家们共创造出了300万个polymer!从这个数值中我们也可以看到Two Minutes游戏模式是至今为止最受玩家欢迎的模式,而后便是One Polymer以及Bombs。我们需要注意的是Two Minutes是游戏开始阶段唯一一个解锁模式,所以它能够成为最普及(特别是在休闲玩家间)的模式也没有什么稀奇。

Two Minutes记分

Two Minutes Scores(from gamesbrief)

Two Minutes Scores(from gamesbrief)

One Polyme记分

One Polymer Scores(from gamesbrief)

One Polymer Scores(from gamesbrief)

Bombs记分

Bombs Scores(from gamesbrief)

Bombs Scores(from gamesbrief)

这些分数图表都非常有趣,因为从中我们可以看出休闲玩家的人数大大超过了严肃的玩家。也许我用词不当了,因为我并不敢确定《Polymer》能否称得上是一款严肃的游戏。我的意思是,大多数情况下,当玩家在玩《Polymer》时都不需要使用过多策略去获得更高的分数。似乎比起获得更高分数,大多数玩家对于移动,配对,Polymer的弹出等游戏体验更感兴趣。影响着One Polymer结果的一大元素是:游戏的持续时间。出于休闲目的而选择One Polymer模式的玩家可能只会花费1分钟的时间去提交分数。最终便导致他们会在一个设置中提交10次不同的低分polymer。另一方面,那些为了争取更高分数的玩家往往会为了一个polymer而投入10分钟以上的时间,所以他们有可能在一个设置中只进行一次挑战,而最终也导致他们只能获得与休闲玩家相近的分数。

Polymer的组合

Polymer Size(from gamesbrief)

Polymer Size(from gamesbrief)

可以说这是所有图表中最让我满意的一张,整个图表中的曲线走向近乎完美。这张图表清楚地呈现出了polymer是如何随着数量的增加而提升难度,几乎没有任何例外。唯一的例外也很有趣:组合4块polymer反而比2块polymer来得简单。当我最初看到这一场景时我也感到非常疑惑,但是深入思考后也觉得这不无道理。我们可以采取许多不同的方式去组成4块polymer(基于各种组合),但是组成2块polymer的方法却只有一种。所以,尽管创造出2块polymer是件简单的任务,但是创造4块polymer比之更加轻松。

有趣的数值

polymer(from gamesbrief)

polymer(from gamesbrief)

我吃惊地发现竟然有人创造了69快polymer。在游戏桌面上共有70块polymer,这便意味着他们只遗漏了其中的一块。除此之外还有一个让我惊讶的结果便是有位玩家只创造了一块polymer。我甚至想象不出来他是怎么了(游戏邦注:最后作者发现,有个用户创造了69块polymer,而他仅有2岁的儿子创造了一块polymer)。我想这其中可能存在一个漏洞,但任何一个玩家都有可能创造出让我们意想不到的结果!

数据更新

updated stats(from gamesbrief)

updated stats(from gamesbrief)

(虽然其它图片都较早,但是这张图片是我刚刚获得的,所以其中包含了iOS 6.0的更新数据)从中我们也清楚不需要再包含iOS 5.0之前的游戏内容。让我惊讶的是,比起iOS 5,如今的iOS 6用户还非常少,我想主要原因应该是我们未在iOS 6发行后及时推出《Polymer》的更新内容。而我希望随着今天的游戏更新,iOS 6用户将随之增加。

设备

device(from gamesbrief)

device(from gamesbrief)

(这也是我刚刚获得的图表,所以从中可以看到iPhone 5的相关数据)这也是一张很有趣的图表,因为这里关于iPhone 5的数值还是非常少。但是就像我在更新数据中所提到的,主要原因还是在于那时候玩家还不能在新系统上玩《Polymer》。所以希望我们的更新能够改变这种情况(为了适应更高的屏幕,我们还特别增加了两行polymer)。

总下载次数

All Time Downloads(from gamesbrief)

All Time Downloads(from gamesbrief)

《Polymer》的下载量已经超过了11000次!

总单位变化

All Time Unitsfrom gamesbrief)

All Time Unitsfrom gamesbrief)

到目前为止,发行当天的销量仍然属于最高销量。苹果的推荐,短暂的0.99美元促销活动,各种媒体的报道,在Twitter上好友们和粉丝们的口口相传推广,以及这篇发布在Reddit(游戏邦注:新闻网站)的文章都是推动着游戏不断发展的主要原因。如今它仍然被放置在侧边导航中!尽管在iPad上的更新一开始未能带来较大的销量增长,但是其滑落速度也相对缓慢,最终促使我们能够长时间地保持每日下载量的稳步发展。

单位累积

Cumulative Units(from gamesbrief)

Cumulative Units(from gamesbrief)

从这张图表中我们可以看出,尽管在2.0更新内容的初始阶段未能获得较好的成绩,但是因为滑落速度也相对缓慢,所以我们在不断发展中累积了更多销量。

总利润

All Time Profitfrom gamesbrief)

All Time Profitfrom gamesbrief)

这张图表代表的是发行以来的所有利润。最高利润值是在2.0更新内容发行之后,因为那时候iPhone和iPad的App Store都推荐了这款游戏,并且那时候的游戏执行的是促销价,而各大媒体的宣传也具有很大的功效。并且因为后来游戏的滑落速度比它在最初发行时来得缓慢,所以我们在更新后所获得的收益也远远多于最初发行时的收益。

iPad排行

iPad Rankings(from gamesbrief)

iPad Rankings(from gamesbrief)

iPhone排行

iPhone Rankings(from gamesbrief)

iPhone Rankings(from gamesbrief)

排行单位

Units with Rankings(from gamesbrief)

Units with Rankings(from gamesbrief)

苹果的推荐非常有帮助!可以说苹果的推荐是帮助我们提升排行名次的最佳方法之一。但是如何才能获得苹果的推荐?我想只有苹果才知道方法,但是我们也掌握了一些诀窍。一方面,苹果希望开发者能够使用他们的iOS SDK功能。所以我便尽所能地使用了iOS功能,包括GameCenter排行榜,多人游戏模式,成就机制,应用内部购买以及Universal更新。其次,获得苹果推荐的一大关键要素便是优秀的设计。如果你的应用难以让苹果感受到推荐的价值,他们便不会愿意去推荐它!最后,努力得到媒体的关注!我便是通过在Twitter上接触更多媒体记者而做到这一点。当一款游戏获得媒体的广泛关注时,苹果便会对其留下深刻的印象。

国家

Countries(from gamesbrief)

Countries(from gamesbrief)

很明显,对于我来说美国便是最大且最重要的市场。但是我发现进一步分析国际市场也是一件很有趣的事。我希望搞清楚为什么有6名美国玩家愿意回到游戏中。

应用内部购买

In-App Purchases(from gamesbrief)

In-App Purchases(from gamesbrief)

应用内部购买占据着我的总利润中的8.5%。正如你所看到的,大多数玩家更愿意接收昂贵的应用内部购买(即2.99美元),因为其中包含了较多的附带内容。我们可以发现,到目前为止,最高的应用内部购买频率是发生在屏幕上弹出某一内容并询问玩家是否愿意购买时。如果游戏隐藏了购买内容,玩家便很难为此主动掏腰包。

第1个月

The First Month(from gamesbrief)

The First Month(from gamesbrief)

这是游戏自4月26日发行以来的首月销量。我们可以发现《Polymer》在获得最高排行后便开始向下滑落。有趣的是,排行变化主要取决于销售的频率,而不只是销售质量。换句话说,比起1周时间内获得1000次的下载量,1天1000次的下载量更能够帮助你提升排行名次。这一点非常明显,但同时我还想说的是:重视你在发行当天所获得的一切。你肯定能够在发行当天获得尽可能多的下载量。但是切忌将目标扩展到1周或1个月,也别想着在发行前去吸引玩家的注意力。如果你能在发行当天吸引更多玩家的注意,你便能够攀上排行榜的更高位置。

体重变化

Weight(from gamesbrief)

Weight(from gamesbrief)

最后一张图表是受到Mike Meade的启发。在最近与他的谈话中,我提到了自己在制作《Polymer》时的体重变化,他便建议我制作一张有关“程序员在开发阶段的体重”的图表。所以我便捏造了一些数据而制作了这张图表(因为我并未记录下这一过程中的体重变化),但是,从本质看来这些数据都是合理的。我是在去年11月结的婚,而为了在婚礼上保持好身材,我便进行了减肥计划,并最终减到了170磅。但是从2012年1月1日开始(游戏邦注:即开始制作《Polymer》),我每周平均增重了2磅。因为每一天我都坐在椅子上埋头进行游戏编程。而当停下工作时,我又开始疯狂大吃(即在我觉得自己应该补充食物的任何时候)!并且那时候的我根本就不在意自己的身材。在《Polymer》发行后的2个月内,我仍然致力于游戏的完善中,所以最终导致我在整个开发阶段(6个月左右)共增重了40多磅!多可怕的数字啊!从8月份以来我又回归到正常的生活轨道上,并减去了15磅左右的赘肉,真是谢天谢地。

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

The power of getting featured by Apple

By Whitaker Blackall

I’ve been promising a Polymer numbers post for a long long time. But I’m lazy and haven’t gotten around to it. And for that, I’m sorry. However, my laziness has paid off in a way, because now you can see how the rankings, sales, and analytics have varied over many months, various Apple features, and updates. I took some of these screenshots a couple weeks ago when I started writing this post so some of the data is a bit old but nothing out of the ordinary has happened since then, so you’re not missing anything. Also, if some of the pics are too small to read, open them in a new tab and view them full size. One last thing: I don’t list any dollar amounts, because I’ve been told that it’s not smart. But you can get a pretty good idea from unit numbers, etc. Okay, let’s get started!

Sessions

I’m starting with sessions because it just blows me away. Polymer has been played over 200,000 times. That’s just crazy to me. This graph clearly shows what will be completely obvious by the end of this post: Apple features are awesome. Yes, having great reviews on Touch Arcade, 148 Apps, and many other publications helped a lot. But I’m convinced that the Apple features were by far the number one contributing factor to the massive sales spikes.

The two main spikes in this graph represent both times I was featured by Apple. The first was beginning April 26th, when Polymer was released. It was featured in the iPhone app store, in the “New and Noteworthy” section of the Games category. The amazing thing about this is that it isn’t nearly the most visible spot for a feature. To get into this section of features, the user has to first find the Games banner, which isn’t always entirely obvious. I can only imagine how many sales you’d get with a front page iPhone app store feature.

The Games feature banner

The second spike is when the iPad version of Polymer was released. This time, I was re-featured in the same spot in the iPhone store, but was also featured in the iPad’s main feature section, reachable by a single swipe on the front page of the app store. While I’m sure it doesn’t get as much traffic as the iPhone app store, it was still quite incredible (especially since this time the app was at $1.99 instead of $0.99). One last really cool thing is that even though I got only about half the sales on the iPad update’s launch (compared with original launch), there were actually slightly more sessions, which I’m assuming means that a lot of the original buyers opened it up again after having it dormant on their phones for awhile.

New Users

This graph isn’t that different than the last graph, but instead of showing how many times the game was played, it shows how many unique devices opened it up. The important thing to note here is that since it is measuring devices, these aren’t necessarily all new users. In fact it’s pretty obvious that a lot of people who already had it on their phones installed it for the first time on their iPads upon the major update. If there really were 3,325 more actual new users during the week of July 16th than there were during the week of April 23rd, than sales would have been much higher.

Session Length Distribution

This graph is pretty cool. First of all, I find it amusing that people have opened Polymer for less than three seconds about 10,000 times. That’s just funny to me. Usually the only time I’ll open an app then close it again immediately is when I’m getting rid of a notification badge, but I’ve never sent any notifications. My best guess is that these less-than-three-second times are from when someone opens it, gets a text, immediately switches to text back, then switches back to the game (for example).

The far end of this graph is definitely the most interesting though. It’s pretty cool that most sessions are 3-10 minutes long. I’m assuming that most of those are more towards the 3-minute side, but it still makes me happy that it beat out the 1-3 minute bar. Even more awesome is that people have played Polymer for longer than 30 minutes in one sitting about 10,000 times. That’s insane! I rarely play iPhone games for that long, and it brings joy to my little heart that some people are that engrossed in my game.

User Session Info

There were 7,180 people who opened Polymer once then never opened it again. Damn, guess it didn’t make a good impression! There were also 11,048 people who may have played Polymer more than once in a single day, but after that day was over, never played it again (although the “one-session users” are included in this, so I think that means that there are therefore 11,048-7,180=3,868 people who opened it more than once in a single day, but never after that). Now, the number I really like is the 8,338. All of these guys installed it then came back and played again at a later date! Thanks guys. I like you. There is one more thing to note with these numbers: as I said before, the “users” actually represent separated devices. So, for example, if someone plays Polymer 1,000 times on his iPhone, but only once on his iPad, that one iPad play would count towards the “one-session users” as well as the “users active on only one day.” Therefore, I’m sure these numbers are pretty skewed.

Event Summaries

First of all, I find it very cool that over 3,000,000 polymers have been made! Crazy. These numbers also make it clear that the Two Minutes game mode is by far the most commonly played, followed by One Polymer, then finally, Bombs. Keep in mind that Two Minutes is the only mode unlocked at first, so it makes sense that it would be the most common, especially for casual gamers.

Two Minutes Scores

One Polymer Scores

Bombs Scores

These score charts are interesting because they show that casual gamers greatly outnumber more serious gamers. Well, I should probably rephrase that because I’m not sure that Polymer could ever be considered a serious game. What I mean is, most of the time, people play Polymer without much strategy for getting high scores. It seems that most people are far more interested in just the sliding, matching, Polymer popping, and overall experience of the gameplay than actually getting a high score. There is one thing, however, that could skew the One Polymer results: the length of time the game lasts. People who play One Polymer casually may only take about a minute to submit their score. As a result, they may end up submitting ten different low-scoring polymers in one sitting. On the other hand, someone who is going for a high score may take ten minutes or more on a single polymer, resulting in only one play per sitting, thus greatly skewing the results towards the more casual gamers’ scores.

Polymer Size

This is probably the most satisfying graph in the whole post. I find it very very cool how nearly perfect the curve is. This graph makes it clear that polymers get increasingly difficult with every added piece, almost without exception. The only exception is an interesting one: apparently 4-piece polymers are quite a lot easier to make than 2-piece polymers. When I first saw this, I was really confused, but upon further thought, it makes a lot of sense. There are a number of ways you can form 4-piece polymers, with various combinations of pieces, but only one way to form a 2-piece polymer. Therefore, even though 2-piece polymers are extremely easy to make, 4-piece polymers are still pretty easy and much more common.

This is Interesting…

I find it crazy that someone made a 69-piece polymer. There are 70 on the board. They used all but one piece. Wow. I think he tweeted me when he did it but unfortunately I can’t remember who it was! Even more amazing than this is that someone made a ONE PIECE POLYMER. Excuse me, but HOW THE HELL?! I have no idea how that’s possible. It must have been a bug, but I just find it amusing. I’m sure whoever it was was pretty surprised too.

Firmware

(Even though some of the other pics are a bit old, I took this one just now so you get updated stats with iOS 6.0) I think this makes it pretty clear that it’s becoming pretty unnecessary to cover anything less than iOS 5.0. I’m a bit surprised that so few people have used iOS 6 compared with iOS 5, but I think the main reason for that is just that there hasn’t been a Polymer update or anything noteworthy since iOS 6 was released. I’m hoping that with today’s update, the iOS 6 numbers will increase!

Devices

(I just took this screenshot right now as well so you could see iPhone 5 stats) I find this pretty interesting, mainly because of how few iPhone 5’s there are. But like I said in the firmware section, I think it’s mainly due to the fact that most people just haven’t been playing Polymer lately. Hopefully the update will change that (especially because of the special two additional rows of pieces added to accommodate the taller screen!)

All Time Downloads

Polymer has been bought over 11,000 times!

All Time Units

By far, the initial launch day was the highest day of sales. This was due to being featured by Apple, having a short $0.99 sale, getting lots of good press coverage, having a ton of really awesome friends and followers on Twitter who helped spread the word, and this Reddit post, which is the fifth most up voted submission of all time in r/gamedev. It’s also even in the sidebar now! Even though the iPad update didn’t result in as many sales at first, it had a slower decline, resulting in more downloads per day for longer.

Cumulative Units

This clarifies some things. This graph makes it obvious that even though the initial spike for the 2.0 update wasn’t as high, it actually resulted in a good amount more sales than the initial release did, due to its slower decline.

All Time Profit

This shows the profit from release to now. My highest day of profit was actually after the 2.0 update release, due to the fact that it was featured in the iPhone and iPad stores, it got some more press, and it was now at $1.99. In fact, since the decline was slower than during the initial release, I actually made a decent amountmore profit from the update than I did from the initial launch!

iPad Rankings

iPhone Rankings

Units with Rankings

Apple features are awesome. Apple features are one of the best ways to get ranked. But how do you get an Apple feature? No one but Apple really knows, but here are some tips. For one thing, Apple loves when you use their iOS SDK features. I tried to use as many iOS features as I could, including GameCenter leaderboards, multiplayer, and achievements, In-App Purchases, and of course the Universal update. Another thing that is extremely important in getting featured is design. If your app doesn’t look Apple-worthy, they won’t feature it (most of the time). Finally, try to get press! How do you do that, you ask? Well, the way I did it was by getting to know a lot of press people on Twitter. And in case you’re not very familiar with Twitter, I wrote a guide to promoting your game through Twitter. Apple is impressed when apps get good press.

Countries

Clearly, the US has been the biggest and most important market for me. But I find it interesting seeing the international breakdown. I wish I could find out why 6 Americans returned it though…

In-App Purchases

In-App Purchases have accounted for about 8.5% of my total profits. As you can see, the most expensive IAP ($2.99) is actually the most commonly bought, but that’s because it unlocks a ton of stuff. You can also see that by far, people more often buy IAP when some sort of popup shows up on screen and basically asks them to buy it. If it’s hidden away in a store, it’ll be won’t be bought nearly as many times.

The First Month

This is the first month of sales, beginning on April 26th, launch day. You can see that Polymer got pretty high in rank but then kept declining until it fell off a bunch of charts. What I find most interesting about this is that rank is primarily determined by frequency of sales, not just by quantity of sales. In other words, 1,000 downloads in one day will make you rank a lot higher than 1,000 downloads spread out through an entire week. I suppose this is rather obvious, but here’s my point: focus everything you have on launch day. Do absolutely everything you can to get as many downloads as possible on launch day. Don’t shoot for launch week or launch month, and don’t prematurely get too much attention before launch. The more buzz on launch day, the higher your rank will go.

Weight!

This last graph was inspired by Mike Meade (of Mikey Shorts fame). I was talking to him recently about how much weight I gained while making Polymer and he suggested I make a graph of “Programmer’s Body Weight Over Development.” So I made a graph with fabricated numbers (because I don’t have any body weight analytics set up) but . . . it’s essentially correct. I got married in November of last year, and for the wedding, I ended up losing weight until I was about 170 lbs. I was really proud of myself! But, beginning January 1st of 2012 (around when I started Polymer), I started gaining about 2 lbs. every week. I was sitting on my ass all day every day programming. I stopped working out. I ate like shit (literally whatever I felt like whenever I wanted). And I generally just didn’t care about my health at all. I definitely don’t recommend this. I continued my binging once Polymer was released for a couple more months, resulting in a net weight gain of about 40 pounds in all over a period of about six months, which is insane and stupid. Finally I’m back on track, and have lost about 15 lbs. since August, so that’s good!(source:gamesbrief)


上一篇:

下一篇: