Facebook App versus FB Connect?

Giff Constable social media


One of my projects posed the question: should we create a native Facebook application (example: Playfish’s Country Story) or a separate website using Facebook Connect (example: Thread.com).  FB Connect, after all, still gives you access to the social graph, feeds, etc.  Here are some of my preliminary thoughts after conversations with other folks, and I would welcome your input.

A Business Issue more than Technical

Several developers are telling me that the issue is probably more business-related than technical.  From what I am hearing, while FBML sounds like a headache, the iframe approach makes embedding your web app or flash application within Facebook a lot simpler.  That implies that it is pretty straightforward to do both if you want, save for the extra work of designing and rendering a different front end for the 760 pixel Facebook app width.

Reasons to Create a Facebook App

  1. Reduce friction: This is a biggie. I would expect greater conversion and retention if you “stay where people live”.  This has been the approach of social game companies, and no doubt why Popcap moved Bejeweled Blitz into Facebook itself.  Trying to create a second destination site, even if you leverage FB’s registration system and social graph, might hurt growth.  However, with that said, Thread.com seems to have taken a FBConnect-only approach and has gained solid traction from what I hear.
  2. Discovery: a FBConnect site can leverage many marketing channels such as the news feed, fan pages, etc, but the current app directory makes it easier to discover Facebook apps, not FBConnect websites.  Facebook might change this given their interest in promoting FBConnect, and furthermore I’m not sure how many people really discover apps this way rather than through friends and ads.
  3. More open namespace: on the Web, it can be tremendously difficult to get a good consumer friendly URL.  On Facebook that is much less of an issue.
  4. Leverage native FB Functionality: if your application is built within Facebook, users can continue using features like Facebook chat.

Reasons to Create a FBConnect Website

  1. Control over advertising: a Facebook app lives within their framework and you cannot control the advertisements that show up in the sidebar or get a piece of that monetization.  If you want to monetize primarily through advertising, you probably want to maximize control over what ads are showing up and how.fbapp-advertising
  2. Control over design: With an FB app, you are limited to a column of 760 pixels and Facebook has control all around you.  If you want to maximize usability and simplicity and have total control over UI optimization, then you are better off building your own website.
  3. Control over metrics: while I believe that transparency around metrics can be quite useful from a marketing perspective, sometimes you want to control when and how they get released.  FB apps force you to share monthly active users.  However, if your numbers are awesome, then you really do want to appear in the top Facebook apps list.
  4. Control over your destiny: having your own website de-risks your dependence on Facebook.  You still need to adhere to Facebook’s policies around their APIs, but should they decide to compete against you, shut you down, or shut down the entire API, as long as you have been collecting email addresses yourself and training users to come to your URL, then you will be badly hurt, but have at least a chance to survive.

My friend John Swords posed the question as to whether Facebook advertisements had different conversion rates if destination was a Facebook app versus a Facebook Connect site, and I’m curious if anyone out there has experimented with this.

Facebook’s policy changes, which affect viral methods, will probably keep the inputs to this decision in flux.  Another factor to watch is Facebook’s payment system, which should be attractive for developers by reducing payment friction, but I assume that will be available to both FB apps and FB Connect sites.

I suspect that more startups will take the “do both” approach.

I would love to hear your thoughts on this, whether you think these thoughts are on or totally off target.

Thank you Joe, Becky and John for your thoughts on this topic.