Follow up on top social networks by engagement July 23, 2007
Posted by jeremyliew in social media, social networks, user generated content, web 2.0.trackback
Last night’s post on the top social networks by engagement provoked some good response. Some readers questioned whether Pageviews per user per month was the best measure of engagement given that some social networks make use of AJAX and other RIA (Rich Internet Application) tools that would reduce pageview count.
As an alternative I also looked at the top social networks ordered by time spent per user per month (again excluding sites with less than 500k unique users per month). Here are the 14 such social networks with over 40 minutes spent per month in June according to Comscore:
The top five sites by PVs are still the top five sites by time spent and there is still quite a gap between them and the other social networks. Notably, Gaia jumps from being number 5 by PVs/user/month to #1 by minutes/visitor/month, reflecting the fact that a large portion of its use is a casual immersive world delivered in Flash (which isn’t counted though pageviews).
Other than that the rest of the list is largely the same. It seems like at a high level both measures of engagement come to the same conclusions.
Lightspeed Venture Partners Blog 
[…] Link to Article ajax Follow up on top social networks by engagement » Posted at Lightspeed […]
hey jeremy, if you have the data available for fubar.com could you send it to me? my internal stats max out at 1 hour session times, and 25-30% of all our visitors have been in that 1 hour+ bracket since it launched… i’m curious if comscore thinks it’s under 40 minutes or if my stats are busted. cheers!
-mike
good followup jeremy… thanks 🙂
however, seems like according to these stats there isn’t much separating the top 5… they’re all withing 10-20% of each other. that’s a lot less substantial than the previous rank ordering due to PV’s, where the variance was up to 50% (and in fact, placed Gaia last in the top 5 instead of 1st).
so i would say that while PV’s isn’t massively uncorrelated with time-on-site, the comparative conclusions one might draw using PV #’s instead of time-on-site #’s is dramatically different… were i to compare Gaia & Orkut by those 2 measures, i’d have a vastly different conclusions.
Q: also would be interesting to see if there was any change in Facebook time-on-site due to introduction of Platform + Apps, or whether it didn’t make much of a difference.
regards,
– dave mcclure
http://500hats.typepad.com/
Interesting. Ezboard is a traditional message board, though, no? Is that still technically a social network? Not that it’s relevant to this particular post, just curious on a side note.
In the end, really, social networks are kind of just blown out message boards, in my opinion. 🙂
You might also want to think a bit about the inverse correlation between pageviews per user and advertiser response (ad clicks) per user — to the extent this does hold up or in some cases, does not hold up. The “quality” of the site for advertisers is a metric I want to see more people start thinking about and talking about — raw impressions numbers don’t always tell the full story, it’s also about potential engagement with people who can pay the bills for these sites. I have yet to see a lot of compelling data or research on this come out from any of the comscore/netratings types but then again it could be that it did and I just missed it while being in the trenches…
Jeremy tabulated above the following14 top social networks for June:
• Average Minutes per Visitor (Column 4)
• Average Pages per Visitor per month (Column 5)
• Average Visits per user per month. (Column 6)
Based on the analysis, the following derivatives may further be obtained:
a) From Columns 4 & 5, the average “page exposure” ( C4/C5 time spent per PV in seconds);
b) From Columns 4 & 6, the average “stay” (C4/C6 average number of minutes per visit) and
c) From Columns 5 & 6, the average pageview per visit (C5/C6).
Thus:
(where M=Minutes , V=Visitors, Pv=Pageviews, Vi=Visits)
Col 4/Col 5 M/Pv: Average Page Exposure in seconds
Col 4/Col 6 M/Vi: Average Stay ,Minutes per Visit
Col 5/Col 6 Pv/Vi:Avergae Pageview per Visit
Sec per Pv Min per Visit Pv per Visit
C4/C5 C4/C6 C5/C6
1 GAIAONLINE 27 19.5 42.8
2 BEBO 24 14.9 36.7
3 Orkut 16 10.0 37.0
4 MYSPACE 18 8.4 27.6
5 FACEBOOK 22 9.6 26.7
6 Comm Connect 27 14.0 31.5
7 FRIENDSTER 27 9.8 22.2
8 TAGGED 24 13.4 33.5
9 VAMPIREFREAKS 28 14.2 30.2
10 HI5 24 10.8 27.0
11 HABBOHOTEL 42 8.7 12.2
12 EZBOARD 34 6.7 11.8
13 HOVERSPOT 21 6.2 17.5
14 IMEEM 38 8.6 13.6
Top 14 Average 27 11.1 26.5
The last row is the average for the top 14 sites. Thus the following “benchmark” may be established:
For the top 14 social networks:
During a visit, a visitor, on an average, spent 11 minutes and viewed 26.5 pages at the rate of 27 seconds per page .
[…] search result pages sitting open – idle – I may not even be in front the computer! Jeremy Liew talked about PV vs. time spent the other […]
Doesn’t this also mean that AJAX/Flash UIs don’t seem to increase (or maybe even hurt) user engagement? With how young most of these sites are, the legacy issues of page construction shouldn’t be driving the issue.
Is it possible for me to know how many MySpace users are form the United States?
Do you have that number?
Im writing my honor program thesis at the University of Puerto Rico, and I need the aproximate number of US MySpace users.
Thanks.
[…] result pages sitting open – idle – I may not even be in front the computer! Update: Jeremy Liew talked about PV vs. time spent the other day and noted some change in how sites would rank. Unfortunately, interaction stats […]
Yes, I think using the actual time a visitors spends makes a lot of sense. I am using the “minutes” for quiet some time to give accurate and quality data back to my clients.
From the looks of it, Myspace still leads the heap in Total Unique visitors
[…] e quantitativa do que pode rolar de qualitativo nas redes. Foi o que encontrei na postagem do Jeremy Liew, da qual reproduzo o quadro […]