Rethinking Facebook Connect

As startups we need to continuously experiment and question the status quo; and for now we experimented with the Facebook Connect implementation. We started by removing it as default option to sign-in on Wishberg. As expected we got multiple forgot password requests (we built this feature in anticipation of same).  

Many folks questioned about this on Twitter, and I also had conversations with other startups founders who suggested this could be a bad move. So far we are happy with the results. We may / may not revert back (its still not clear) – but since many people asked me why we even thought of experimenting – here are the reasons.

a. Facebook is no longer a powerful distribution platform 

Let me sum up Facebook as a distribution platform for you:
Early days -> 1 + 1 = 11
Later ->  1 + 1 = 2
Now -> 1 + 1 = 1.1
Next -> 1 + 1 = 1.01

Face this, it is true. Facebook is no longer a powerful distribution platform or user acquisition channel for application developers. If the expectation is one user registration through Facebook connect will lead to at least one more., its not happening. 

Zynga achieved its distribution on Facebook through News Feeds; Branchout through notifications and others like Pinterest / Spotify through Open Graph. When more and more applications tried to ‘abuse’ each of these mechanisms Facebook put more restrictions & controls in place (which is correct since Facebook wants to maintain a clean experience for its users). Open Graph is currently the only way to get some effective distribution, Facebook has replaced few custom actions and asked developers to use built-in actions for Like & Follow, they are also merged in Open Graph. It also placed restrictions for applications that abused few actions like ‘read a article’ & ‘viewed a video’ with more controls / validations in place.

Personally I am against spam and to build a clean product we do not aim to spam our users through Facebook (even in name of user acquisition). Also because of the fact that few applications have abused Facebook to acquire users, users are smart and know how to differentiate between a possible spam and genuine link. Good for consumers and bad for developers, Facebook has made it ‘ridiculously easy’ for users to get rid of applications; so if your are spamming – do that at your own risk! 


b. Facebook engagement principles – P2P v/s A2P

You must have read this in news over and again – Facebook is trying hard to appeal to the current youth generation (as the earlier one has grown up!). While Facebook is trying to appeal to younger generation, it is also trying to improve engagement of its current user base. Current reports suggests that a Facebook update reaches approximately (just) 12% of your friends. 

If people stops engaging with other people on Facebook, it will be dead. While Facebook connect is a good way to keep social interactions that happen outside of Facebook discoverable through feeds on Facebook – its natural that Facebook will always be more inclined to have P2P (People to People) interactions featured over A2P (Application to People). 

On a personal note – I don’t think Facebook will have anything great to announce for some time ahead that will excite the developers. For now, FB will focus on improving user-to-user engagement, appeal to youth and its monetization products. So I don’t see the situation improving for developers. 


c. Inconsistent Discovery Experience for feeds

Facebook is not Twitter. Unlike the experience where every tweet is visible to your followers, every feed / status update is not visible to your friends. Its complex and depends on multiple factors – whom you interact with most, which group of friends are you a part of, has the feed gone viral to be showcased to more people outside that network and so on. If P2P feeds are discovered by only 12% of friends, chances for discovery of application feeds will be even lower.

And then there are innumerable pages that a user has liked, there are updates from them which also ask for a mind-share of user in his activity stream. One of Facebook’s monetization product that allows pages (and users) to pay and increase reach of their posts will also work against discovery of application feeds.

Facebook recently announced a new newsfeed which is rolled out to few users but not to all. It has also did a nice little revamp on Timeline view of profile putting all updates on the right side block – and also bringing up user’s likes and interests upfront and pushing open graph updates further down to an blind spot.

Don’t get me wrong here, I am supporting Facebook here as most of these changes are done to improve user experience and engagement for its own users. But in that attempt – the discovery of feeds for applications has got bit inconsistent. There is no science here – and most of the times for application developers it will mean shooting in the dark. 


d. Every user has his/her own identity on every platform

Each user has a different identity on every platform – Facebook, Twitter, Quora, Foursquare, LinkedIn and so on. Its incorrect to assume that the way a user behaves on Facebook will be essentially the way he will on your product or that he wants his friends to know he is using a particular product or service.

There is also a trend that users do not want to register on a product because it only allows only Facebook Sign-up. We did that with Wishberg earlier, but now have opened up email registrations; key here is – ‘Be valuable first, social later.’

The ideal way is to allow users to register and let them connect their Facebook account as a option – which they will if your product is valuable to them. You can prompt users to connect their Facebook account, but not force them to do so!


e. Psychology of Forced Distribution

As a developer once you implement Facebook Connect, unconsciously you get thinking and start relying completely on Facebook for distribution. You want every action that has happened on your product to be ‘forced shared’ on Facebook – even though a user would want it or not. Its time to stop that as the sharing economy has changed.

Since Facebook distribution is not controlled by you, it gets increasingly frustrating when your product does not go as viral as you thought it would. Instead build some sort of distribution / discovery mechanism on your own product which you control completely – we built couple of them on Wishberg and they have worked remarkably better. Remember – a small number of highly engaged users are much better than a large user base with zero or near zero engagement. 


f. Sharing economy has changed

The sharing economy on Facebook as changed over years. It is no more driven by features or applications, its completely user driven. Users have got smart enough to know what has to be shared and with whom.

Don’t build applications / features that will trick users to forcefully share something on their wall without their consent or knowledge. Focus on your product – users will figure out what they have to share and what they don’t have to. Users are now smarter than most developers think! 


g. Breaking changes that break your plans

While doing a startup / building your product – the one thing you don’t want to lose is time. Startups operate with small teams and any deviations from the product roadmap costs them dear.

And while they are on to it – Facebook wants you to constantly be updated with its latest ‘breaking changes‘ and there is no option but to comply. It sucks out time / bandwidth big time and knowing the diminishing returns from Facebook – it gets frustrating here!

 

Concluding Notes:

Most product managers integrate with Facebook Platform for three reasons – 1. One-click sign-in 2. Social Graph. 3. Distribution (Viral acquisition of users).

It is possible to achieve that without Facebook too.

  1. One Click Sign-in: Create a perpetual logged-in experience for users till he explicitly logs out!
  2. Social Graph: Most successful products like Quora, Twitter, Instagram, etc have build their own network / graph. Remember that same user will have a different identity on every different network.
  3. Distribution: 1 + 1 = 2 is no longer true. Think of discovery and distribution on your own product, you have complete control there.

The aim of this post is not to put negative remarks against Facebook, but to make fellow entrepreneurs know of this when they are building on top of Facebook platform and so that they set right expectations for growth. Happy building!

Update:  We got Facebook Connect back on Wishberg after 30 days of experiment. The main reason was not distribution, but authentication – users do not have to remember one more password! As far as distribution is concerned, Facebook adds little value.