I’ve been pondering the right algorithm for measuring the success of a social media app – what may be called the apps ability to incite peerfluence. I first came across this term courtesy of Stowe Boyd via Twitter and also note that Social Media‘s Seth Goldstein used it in his company’s ad:tech SF press release as well.
UPDATE: Turns out that peerfluence is a registered trademark of a company called peerFluence – see comment below from their CMO, Bill Franchey. I’ve also had a call from their CEO, Darren Johnston and we are playing phone tag – I’m keen to hear more about what they do 🙂 . To clear up any misconceptions, real or otherwise, my point above re Social Media should not be construed as inferring any relationship between these two companies. In addition, I’ll change the word I use from peerfluence to peerinfluence for the rest of this post.
Peerinfluence can be defined as the ability for one person’s actions to influence their peers. Some folks have a high level of peerinfluence, like the inimitable Robert Scoble – I witnessed this first hand when he blogged about Yoick. Other people have become increasingly peerinfluential, such as Chris Saad, the founder of the Dataportability Group.
I believe that we can call a social media app’s ability to incite peerinfluence as its level of peerfluentiality. This is on the one hand a metric of the app’s ability to attract people who already have a high level of peerinfluence and get them to spread the word about it. On the other hand it is also a metric of how many times and for how long the average person spreads the word.
In determining the peerfluentiality ranking of an app there are a number of factors that need to be taken into account:
* Installs – how many installs in total across all networks and platforms
* Networks – how many networks the app is running on
* Platforms – is it running across social networks, the wider web and mobile
* Network effect – basic level (is there simple cross referral “house ads” between disparate apps,)
* Network effect – advanced level (does the network consist of strategically interconnected apps that apply the same game metaphor, storyline or tap the same demographic; are the cross referrals more subtle, eg the Zynga toolbar)
* DAU – of all the installs what percentage are accessing the app on a daily basis
* Engagement – how often are users accessing the app, how many page clicks are they making inside the app, do you have groups, forums or other means for users to talk, share, interact with others around/about the app.
A mix of these factors will determine the peerfluentiality of an app and some of these factors will be bigger drivers than others.
For example, we have seen that simply having a high install base does not mean sustainably high DAU. In fact, many apps that initially attracted lots of users are now flatlining.
Similarly, purely because an app is highly engaging does not necessarily mean it is initially sticky – if the user onramp is too complicated the chances are an installer is less likely to refer it to his or her friends and those who do have it referred to them may not get to the point where it becomes engaging for them.
As we formulate more data around social media apps I expect to be able to tweak this algorithm and seek out optimal peerfluentiality.