蒸肠粉的做法:网络“相关性时代”到来

网络“相关性时代”到来
2011年03月07日09:28腾讯科技[微博]Kathy我要评论(0)
字号：T|T
[导读]“在社交网络之后，下一个大的潮流又是什么呢？...新的潮流已经出现：网络正在从简单的社交共享转向个性化的、具有相关性的内容。”
腾讯科技讯（Kathy）北京时间3月4日消息，据国外媒体报道，新闻聚合服务Techmeme的编辑马亨德拉·帕素雷（Mahendra Palsule）今天在知名科技博客TechCrunch发表署名文章，称网络进入了“相关性时代” ，以下为全文摘要：
在社交网络之后，下一个大的潮流又是什么呢？
科技爱好者们谈论这个话题已经有好几年时间了。我认为新的潮流已经出现：网络正在从简单的社交共享转向个性化的、具有相关性的内容。
这个潮流的关键元素是，配合社交图谱的兴趣图谱（Interest Graph）变得越来越重要。Facebook、Twitter和谷歌正在致力于提供具有相关性的内容，此外还有很多初创公司专注于此。
相关性是解决信息超负荷问题的唯一办法。
转播到腾讯微博

 
上图描述了网上信息发现方式随时间演化的过程。
第一阶段：搜索为王
这就是谷歌如何在二十年的时间里在网上称霸的原因：它通过PageRank把人气最高的网页置于上方；而一个网页的人气是通过链接到它的页面的数量来确定的。
第二阶段： Web 2.0 和社会化书签
在Web 2.0时代，社会化书签服务取得了长足发展，它们将人气高的内容置于页面上方。像Reddit和StumbleUpon这样的网站即使在今天也非常受欢迎。
第三阶段：个性化推荐
像Hunch、GetGlue这样的服务的关键在于为用户建立兴趣图谱，从而通过“口味引擎”来提供个性化的推荐。
第四阶段：个性化的“新奇发现”（Serendipity，指全然无意中发现了新奇东西）
一些新的初创公司侧重在个性化地混合使用兴趣图谱与社交图谱上。个性化的“新奇发现”也就是杰夫·贾维斯（Jeff Jarvis）所说的“出人意料的相关性”。这类服务的例子包括Gravity、 my6sense、Genieo和TrapIt。
究竟什么是相关性？
要针对信息超负荷的战斗中，人们使用的“武器”通常有两种：相关性和人气。这里的“相关性 ”等同于“个性化”，是和“人气”相对的选择。
然而，相关性并不总是意味着个性化。相关性是一种非常动态化的东西，它取决于一个人在某个特定的时间点上的需要。有很多时候，你想了解人气最高的内容，而其他时候，你只想看到个性化的内容。
目前有多种方法来对信息进行相关性过滤。比如谷歌、Paper.li和PostRank是用算法来过滤，而Reddit、Hacker News使用了众包（crowdsourcing）方式。Klout的“影响力排名”可以被用来过滤Twitter消息流，Facebook在新闻流中使用了社交关系这个过滤器，在它新推出的评论插件中使用的过滤器则是社交信号。对于提供具有相关性的内容来说，地理位置是另一种重要的信号，而且它在移动世界中的重要性正在日益增长。
换句话说，相关性横跨了图中的所有象限，在上述的各种相关性过滤方法中，没有哪一种是“最好的办法”，因为对于相关性来说，不存在“杀手级”的方法。TrapIt的首席营销官小亨利·诺斯哈福特（Henry Nothhaft, Jr.）就曾表示，不要迷恋所谓的相关性过滤“杀手级”方法，它只是传说。支持多种发现方法，多种过滤方法，具有灵活性，并支持多种移动平台的服务才会更具竞争优势。
Quora：兴趣图谱的展示
Quora开创了将兴趣图谱作为新闻源的主要信号的做法。 Quora在新用户注册过程中，就要求他们选择关注哪些主题，表明了关注主题的重要性不逊于关注用户。
Quora的新闻源展示了当你把社交图谱和兴趣图谱混合在一起时会怎样：很多人都对它上了瘾，但又很难解释为什么会上瘾。一篇文章出现在你的新闻源中，不是因为你是关注了某个用户，而是因为你关注了相关的主题。
这往往会导致个性化的“新奇发现”，即“出人意料的相关性”，这就是为什么很多人对Quora上瘾的原因。
去年，Twitter和Facebook之间展开了兴趣图谱之争。因此， Quora是如何在这个游戏中占据优势的呢？
Quora是从零开始建立起来的，兴趣图谱就是它的支柱。而 Twitter的“Browse Interests”功能覆盖范围太过广泛，使用效果不佳。虽然Facebook可以让发布商推送新文章到你的新闻源中，但大多数发布商一直都不知道这个功能的存在。
这也是为什么Facebook的“赞”按钮现在会发布完整的新闻源故事的原因。未来显然属于那些最好地利用了兴趣图谱的公司。
由相关性驱动的Web，其影响深远而广泛。一个服务如果可以更好地利用兴趣图谱，它就会获得更好的定向广告效果，而对CPM (每千人浏览页面的费用)式广告的依赖性也可能会降低。而且有可能通过把重心放在交易和订阅上获得更高的营收。网络媒体发布商会更重视相关性指标，比如用户参与度和花费在站点上的时间，而不是像网页浏览和流量这样的原始指标。
社会化媒体可能也不会再痴迷于关注者人数和流量，而且进化到语境驱动的声誉系统和算法上。
兴趣图谱还将被用于构建更好的社交图谱。今天的兴趣图谱将根据不同的相关性需要，进一步细化为品味图谱、财务图谱、本地网络图谱等等。
The Age Of Relevance
7diggs
Tweet
71
View Comments
Mar 3, 2011
Editor’s note: This is a guest post submitted byMahendra Palsule, who has worked as an Editor atTechmeme since 2009. Apart from curating tech news, he likes analyzing trends in startups and the social web. He is based in Pune, India, and you can follow himon Twitter.
What’s the Next Big Thing after social networking?
This has been a favorite topic of much speculation among tech enthusiasts for many years. I think we are already witnessing a paradigm shift – a move away from simple social sharing towards personalized, relevant content.
The key element of the next big thing is the increasing significance of the Interest Graph to complement the Social Graph. While Facebook, Twitter, and Google are already working on delivering relevant content, a slew of startups are focusing exclusively on it.
Relevance is the only solution to the problem of information overload.

The above matrix is a representation of how the process of online information discovery has evolved over time.
Phase I: The Search Dominated Web
This is how Google began its dominance over the web two decades ago, using PageRank to surface the most popular web pages as identified by other web pages that linked to them.
Phase II: Web 2.0 With Social Bookmarking
In the Web 2.0 era, social bookmarking services gained significant traction, surfacing popular content. Sites likeReddit andStumbleUpon are hugely popular even today, driving millions of page views.
Phase III: Personalized Recommendations
Services like Hunch, GetGlue, etc. have focused on building an Interest Graph for users, to deliver personalized recommendations using a ‘taste engine’.
Phase IV: Personalized Serendipity
The latest crop of startups is focusing on personalization using a combination of Interest and Social Graphs. Personalized Serendipity is what Jeff Jarvis calls‘Unexpected Relevance’. Examples includeGravity,my6sense,Genieo, andTrapIt.
What Exactly Is Relevance?
The battle against information overload is sometimes presented as a choicebetween Relevance and Popularity, where ‘relevant’ is equated to ‘personalized’ as against popular.
However, Relevance does not always mean Personalized. Relevance is very dynamic – it depends on the needs of a person at a specific point in time. There are times when users want to know about the most popular stories, and other times when they seek personalized content.
There are multiple approaches to filtering information for Relevant Content. Google, Paper.li, and PostRank are examples of algorithmic filtering, while Reddit, Hacker News use a crowdsourcing approach. Klout can be used to filter Twitter streams by influence, while Facebook usessocial affinity as a filter for its newsfeed and social signals for itsnew Comments Plugin. Location is another high-impact signal for delivering relevant content, gaining importance in a mobile world.
In other words, Relevance spans across all the quadrants of the Discovery Matrix above, and none of the above approaches to filtering for relevance is the ‘best approach’. There is no killer approach to Relevance. Henry Nothhaft, Jr., CMO of TrapIt, described it as“the myth of the sweet spot”. The competitive edge will be with services that support multiple discovery methods, multiple filtering approaches, have flexibility, and support multiple mobile platforms.
Quora: A Showcase Of The Interest Graph
Quora has pioneered the use of the Interest Graph as a dominant signal for its newsfeed. Quora asks new users to select Topics to follow, as part of its onboarding process, which is the first revelation that Topics are as important as Users to follow.
Quora’s newsfeed is an interesting showcase of what happens when you mix an Interest Graph with a Social Graph – and the result is the mysterious addictiveness so many have experienced, but found difficult to explain. An item pops up in your newsfeed not because you were following a user, but because you were following a related topic.
This often leads to Personalized Serendipity – or Unexpected Relevance – which is why Quora gets many people hooked.
The war over the Interest Graph began between Twitter and Facebook last year,as Erick described so eloquently. So how did Quora beat them to this game?
For starters, Quora is built from the ground-up with the Interest Graph being a backbone of the framework. Twitter’s‘Browse Interests’ is too broad and primitive to be of use, even at present. And while Facebook has a mechanism for allowing publishers to push new items to your feed, most publishershave been unaware of this functionality.
This is also the reason why Facebook’s Like Button now publishes afull news feed story. The future clearly belongs to who best captures the Interest Graph asMax Levchin and Bill Gurley put it.
The implications of a Relevance-driven web are wide-ranging and broad in scope. Better utilization of the Interest Graph by services will lead to better ad targeting, and a potential decrease in reliance on CPM/CPC-based advertising. Monetization focus will be on higher yields through transactions and subscriptions as Dave McClureonce described. Online media publishers will focus on Relevance Metrics revealing engagement and time-spent on site, than primitive metrics like page views and traffic.
Social media may lose its obsession with follower numbers and traffic, evolving to context-driven reputation systems and algorithms.
Interest Graphs will be used to buildBetter Social Graphs. Today’s monolithic Interest Graph will getfurther specialized into Taste Graphs, Financial Graphs, Local Network Graphs, etc., yielding higher relevance for different needs.
The Age of Relevance beckons!