Online social network participation inequality and the impact on the User Engagement KPI
In my recent post about what and how to measure Social Networking websites I suggested the following online social network user engagement KPI’s:
- User Engagement
- Anonymous visitors to members conversion rate
- Active member length
- Time since last login
- Total time spent on site
With a comment added in regards to the User Engagement KPI saying that: “The User Engagement KPI is a custom session metric designed and calculated from a set of basic metrics such as for example; pages viewed, time spent on site, time since last login, comments posted or other content submitted, subscribed to a feed or alert and so on – the calculation of this KPI is highly dependent on the structure of the Social Network in question).”
I think it is of the utmost importance that I add a note to that, as you cannot presently look at an average or median User Engagement KPI without segmenting it, due to the fact that there is an online social network participation inequality. The fact is that most users do not participate very much!
Your UGC-users (User Generated Content-users) are split into three groups:
- Observers - those who do not contribute
- Contributors - those who contribute now and then
- Participators - those who account for most of your contributions
And as you probably guessed - the distribution of the contributions follows Zipf’s law – should you plot the User Engagement KPI for each user (If you run a blog, you can for fun play with the idea that User Engagement equals “comments” – and as such plot the number of comments per unique user, assuming you have access to full visitor segmentation in your analytics tool).
I also suggested in the above mentioned online social networking measurement post that you had to create a KPI Index. Here you will have to extend the User Engagement KPI Index with an attitude on the distribution.
Online Social Network User Engagement KPI Index (Distribution attitude)
- 90% - Observers
- 9% - Contributors
- 1% - Participators
The general attitude and rule-of-thumb is not surprisingly that you have about 0.1% participators on a blog (which differs from online social networks in general, by having a very low participation) - this is of course something you have to determine for your blog or online social network.
My conclusion is that:
When using a User Engagement KPI (as suggested by myself) it is highly unrepresentative using either average or median numbers (or voices for that matter, as in specific comments) to conclude on the overall attitude of your online social network, due to the fact that most online social networks have participation inequality! Segmentation must be applied.
August 10th, 2007 at 15:54
Thanks for the article.
How about measuring:
1. Stickiness
2. Velocities (RFM+visit/purchase)
3. First purchase momentum
4. Repeat purchase momentum
5. Freshness factor (how often content is refreshed versus how
frequently users visit the site)
August 11th, 2007 at 5:47
Hi there…
Thank you very much for your input. And as you might have noticed, I am a GREAT fan of spending an immense amount of time in picking the right KPI’s – however; I try really hard, as in advocating, to ensure that one do not mix up traditional metrics (that certainly affect the KPI’s) and the KPI’s themselves.
That said and having in mind that we are talking about Online Social Networks KPI’s, I think the following KPI’s that you mentioned are better for particular e-commerce properties. (or this at least with my interpretation of Online Social Networks) :-)
2. Velocities (RFM+visit/purchase)
3. First purchase momentum
4. Repeat purchase momentum
BUT that the following two metrics very much relate to Online Social Network.
1. Stickiness
5. Freshness factor (how often content is refreshed versus how frequently users visit the site)
I would (and that is not necessarily the truth) probably use “freshness” as a metric in my quest for optimization on the “Visits per week” KPI.
And I think that we are most likely on the same page when you talk about “Stickiness” and I talk about “User engagement” – As I see this KPI being calculated from multiple basic metrics and that it is supposed to be VERY specific for the web property in question.
Thank you again. This is great input on a really (for me) fascinating KPI topic.
N.B.
I am very likely to get inspired by your first three metrics when writing up my “What and how to measure e-commerce websites” post. :-)
Cheers
Dennis