评论
打印
When you're surfing through Netflix on the bus, the train, work, or in the comfort of your couch at home—
当你在公交车上，火车上，工作的时候，或者舒舒服服地躺在家里的沙发上刷Netflix的时候，
how do you pick what to watch?
你是怎么决定看什么视频的?
Is it a catchy title?
是看标题吸引人的?
Maybe an interesting synopsis?
还是简介有意思的?
Perhaps it's the unwillingness to watch another episode of The Office.
还是受不愿意再看一集《办公室》的念头驱使。
Save Bandit!
“救救强盗!”
Or just maybe, a particular piece of cover art speaks to your spirit?
又或者是特别的封面艺术击中了你的内心？
But go on your friend's Netflix account—if they aren't still using your subscription—
但如果你进入你朋友的Netflix账号——如果他们没有在用你的账号的话——
and you may find the cover art there doesn't interest you.
你可能会发现，TA账号里的视频封面你根本就不感兴趣。
In fact, what you see might be completely different.
事实上，你看到的内容可能跟你自己账号里的完全不同。
It's no accident—Netflix's thumbnails are all tailor made for you.
这并非偶然——因为Netflix的封面都是为用户量身定做的。
At its heart, Netflix is all about creating personalized experiences—or, rather, calculating them.
就其本质而言，Netflix的核心是创造——更确切地说是计算——出个性化的体验。
They'll draw you in with curated trailers of upcoming releases,
他们会用即将发布的预告片来吸引你，
point out new episodes from previously watched videos,
会提示你以前你看过的视频的更新内容，
and even gauge your interest in content via match scores.
甚至还会通过匹配率来衡量你对一个内容可能会有多大兴趣。
And while streaming services are notoriously tight-lipped about sharing their viewership data.
尽管流媒体服务商们对他们的用户数据是出了名的三缄其口，
over the years, Netflix has shared glimpses into how their technology works.
但这些年来，Netflix还是给我们提供了那么一丢丢窥见他们的技术的运营模式的机会。
According to internal studies, a typical viewer spends 1.8 seconds considering each title,
Netflix的内部研究显示，一个典型的观众会用1.8秒的时间来审视每一个标题，
and Netflix believes it only has 90 seconds to get your attention before you move onto another activity.
而Netflix认为，在用户改变主意去做别的事情之前，它只有90秒的时间来吸引我们的注意力。
And among all the things that could catch your attention and make you watch a show– or several—
在所有能吸引你注意力、让你看一个——或多个——视频的东西中，
Netflix found that the biggest influence were the thumbnails.
Netflix发现，能对用户产生最大影响的是封面。
Humans are intensely visual creatures.
人类是高度视觉化的生物。
Our eyes move three to four times a second to process new information.
我们的眼睛每秒会动三到四次来处理新信息。
And because Netflix's goal is to get your attention and hold it,
由于Netlix的目标就是吸引并牢牢抓住你的注意力，
the company puts a lot of work into choosing every thumbnail you'll see.
所以，他们在挑选用户能看到的每个封面上都投入了大量的工作量。
But before they can decide what image will show up on your account, they have to sift through a ton of data.
不过，在决定在你的账户上显示什么图像之前，他们得对海量的数据进行筛选。
An hour-long Stranger Things episode has almost 86,000 video frames.
一小时的《怪奇物语》有近8.6万帧视频。
To figure out which ones will make the best thumbnails,
为了弄清楚哪些能当作最好的封面，
Netflix uses a pretty scientific selection process called Aesthetic Visual Analysis, or AVA.
Netflix用了一种非常科学的选择过程，叫做美学视觉分，简称AVA。
AVA is a set of tools and algorithms that search Netflix videos for the best images and pull them out to create thumbnails.
AVA是一套搜索Netflix的视频，找出最适宜于制作封面的图片的工具和算法。
The process can be broken down into two basic steps.
这个过程可以分为两个基本步骤。
First is frame annotation.
第一步是注释框架。
A program analyzes every static video frame of a video,
用程序分析一个视频的每帧静态的图像，
and image recognition algorithms use information gleaned from each shot to create metadata.
接着，图像识别算法使用从每个镜头收集的信息来创建元数据。
The metadata is like an electronic fingerprint—identifying characteristics unique to each video frame.
元数据就像每个视频帧特有的电子指纹识别特征。
All of this info helps build a database of information that makes it easier to pick out the best images for thumbnails later.
所有这些信息都有助于建立一个信息数据库，以便以后更容易地为封面选择最佳图像。
And to sort all this data, the company groups it into roughly three categories that are key to identifying good images:
为了对所有这些数据进行分类，该公司将这个数据库的数据大致分成了对识别好的图像非常关键的三个类型：
Visual—focusing on brightness, color, contrast, and motion blur.
一类是视觉型图像——侧重亮度、颜色、对比度和运动模糊。
Contextual, which documents face and object detection, motion, and shot angles.
一类是情境型图像——即记录人脸和对象检测、运动和拍摄角度的图像。
And compositional—which focuses on visual principles in cinematography, photography, and design.
还有一类是构图型图像——侧重于电影摄影，摄影和设计对视觉原则的运用。
The second step is a process called image ranking.
第二步是给图像排等级。
An algorithm uses the metadata to pick out specific shots that Netflix has determined are the most attractive and clickable:
算法利用已有的元数据挑选出Netflix认为最吸引人、最容易被点击的特定镜头：
ones that aren't blurry, have varied imagery, feature major characters, and don't contain sensitive or unauthorized branded content.
就是那些不模糊、图像多样、主角突出、不包含敏感或未经授权的品牌内容的镜头。
Then, finally, a creative team steps in to use the best images to design the thumbnail artwork.
最后，创意团队开始加入进来，将最好的图像设计成封面。
But the process doesn't stop there—Netflix still has to figure out which ones work best for each user.
到这里还不算完——Netflix仍然需要找出最适合每个用户的封面。
A/B testing is executed—again, again, and again.
此时，他们需要反复执行A/B测试。
You'll regularly see changes in your thumbnails—based on your engagement with previous titles.
我们定期会看到封面发生变化——根据我们对之前的标题的反应。
On the most basic level, let's say you're a fan of comedy, and you've watched a bunch of stand-up specials.
拿最基本的来说，假设你是一个喜剧迷，你看了很多单口相声特别节目。
When you search for Good Will Hunting,
那么，当你搜索《心灵捕手》的时候
you may get a thumbnail with Robin Williams, a famous comedian, and one of the movie's main characters.
你就可能看到封面是罗宾·威廉姆斯，他是一位著名的喜剧演员，同时也是这部电影的主角之一。
But people more into romance titles could be shown the cover art with the two leads kissing.
但更喜欢浪漫题材的用户得到的封面可能就是两位主角接吻的镜头。
There are also regional differences.
Netflix的封面还会因地域而异。
From glancing at a show's thumbnails across different countries,
从不同国家的用户观看节目的封面来看，
you could infer Germany is more into abstract images, and US viewers may prefer clearly defined characters and story plots.
我们可以推断德国更喜欢抽象的缩略像，而美国观众则可能更喜欢角色和故事情节定义清晰的封面。
Though there are a considerable number of thumbnails to choose from, and lots of science that goes into each decision,
虽然可供选择的封面很多，而且每一个决定都有很多科学依据，
it doesn't mean Netflix gets it right all the time.
但这并不意味着Netflix每次都能做出正确的判断。
One Twitter user found the cover art for, "Like Father" a movie starring Kristen Bell, Kelsey Grammer, and Seth Rogen,
一位推特用户发现《虎父无犬女》，由克里斯汀·贝尔、凯尔西·格兰莫和塞斯·罗根主演的电影，的封面，
had cover art that didn't exactly match who the lead actors in the film were.
与电影中的主角并不完全匹配。
It was also kinda odd to see on my own account,
而且，在我自己的账号里看到
that the Catwoman thumb has an image of a barely recognizable supporting actress Sharon Stone,
《猫女》的封面是一个几乎认都认不出来的配角莎朗·斯通，
and Blade II had someone other than its star, Wesley Snipes, in the thumbnail.
《刀锋战士》的封面竟然不是它的主演韦斯利·斯尼普斯，也是很诡异。
Netflix is obsessed with A/B testing new features like video promos, intro skipping, and auto-playing trailers—
Netflix对用A/B测试视频宣传、插播和自动播放预告片之类的新功能可以说是十分着迷——
just as obsessively as they are with testing thumbnails.
就像它痴迷于用A/B测试来测试封面一样。
Netflix wants everyone to watch more, so it's unlikely they'll stop doing these tests.
Netflix希望每个用户都能多看一些内容，所以他们不太可能停止这类测试。
And if you don't like being a guinea pig, there is the option to opt out.
如果你不喜欢当一只小白鼠，那你可以选择不打开Netflix。
But with binge-watching quickly becoming a national pastime,
不过，随着刷剧迅速成为一项全国性的消遣活动，
it's likely that, when Netflix asks you, "Are you still watching?"
当NEtflix问你，“你是不是还在看我们的视频？”的时候，
You will be.
你的答案还会是“在看”。
