Horizontal vs. Vertical Feature Stack
Today Google announced its Buzz integration with Gmail service. In case you haven’t heard what the “buzz” is really about, it is yet another social network service that is very much similar to Facebook, in the sense that it allows Gmail users to share each other’s status, posts, links, pictures and videos. The interesting part though is that Google Buzz attempts to combine/integrate all the existing social network services such as Facebook and Twitter into its own interface. In other words, Buzz allows you to import and eventually export your social posts from and to the other social networks, like Facebook. You can see your friends posts on Facebook, receive tweets from Twitter, and eventually being able to publish to those services directly within Buzz. And, this is what I call a vertical feature stack.
By vertical feature stack, I mean a software that attempts to integrate several existing services into a single user interface. Kinda like the SSO (Single Sign On) idea, where a client/user can access all his/her services via a single UI/ID. On the other hand, a horizontal feature stack implies that a specialized UI or access point is used for each of the services. This is our current way of using these social network services, each one of them has its own access interface.
I remember first talking about the terms of horizontal and vertical feature stack with a friend last summer. He described the Asian market as more of a vertical feature stack oriented market, where users prefer a single UI being able to provide or access multiple services. He gave me the example of the IM client “QQ”, which not only provides the basic IM service but also a whole array of other social services such as emails, customized avatars, video/audio chat, blogs, social games, screen capture, customized emotions, social groups and more. Interestingly enough, this vertical feature stacked IM client eventually pretty much took over the entire market of all the services it provides. Internet users in China now use QQ as their main service provider for all these services. However, it’s a completely different story in the western market, where horizontal feature stack oriented services rule the Internet. I use Gmail as my personal email service, AIM as my IM service, Facebook as my “blog” and of course WordPress. I don’t use twitter but so many of my friends do, as whatever service you want to call it. I play social games on random specialized game sites such as EA’s social game site. I don’t have a customized avatar service but I’ve certainly seen ads of such.
Just think about how many application programs I have to open to get to all these things, iChat, Safari, Mail and a bunch of browser tabs or windows. This sure can quickly get out of control and become very hard to manage and switch between these services. A vertical feature stack oriented client, such as Google’s Buzz, can certainly help getting rid of opening a bunch of windows or tabs to access all the social network services. And that, can save me a lot of hassle. Not to mention being to able to see new “posts” coming in from all the sources must be a pleasant experience.
On the other hand, horizontal feature stacked applications do have their obvious advantages, namely specialization. A specialized interface design can significantly increase the human computer interaction performance. I’ve always hated the idea of accessing web-services using a general purpose web-browser just because a specialized standalone application program can do things in a much more efficient way. This is especially true on embedded devices such as a smart phone. Just think about how we access Facebook on an iPhone. The Facebook app of course! Who would want to go through all the hassle of accessing the general purpose webpage interface via a web browser?! No one!
Also, the horizontal vs. vertical feature stack comparison can also be applied to microprocessor architectures. When you only have a single CPU/core, you push all your processing tasks onto the same core. This obviously makes things very simple as all tasks are simply executed sequentially (assuming no hyper-threading). However, as history has proven, this strategy does not scale very well as tasks must wait for prior scheduled ones to complete before being executed. Simple to manage, but hard to be efficient. On the other hand, a multi-CPU/core architecture allows each task to be executed independently on its own dedicated processing unit in parallel, which greatly increases the efficiency but of course, coordinating and synchronizing between the tasks become exponentially more difficult. More efficient, but hard to manage.
At the end of the day, this really is a trade-off between efficiency and simplicity. Horizontal feature stack favors efficiency, where as its vertical cousin favors simplicity. Let’s see how this will apply to the social network services as time will tell the whole story.
