Corante

Authors

Clay Shirky
( Archive | Home )

Liz Lawley
( Archive | Home )

Ross Mayfield
( Archive | Home )

Sébastien Paquet
( Archive | Home )

David Weinberger
( Archive | Home )

danah boyd
( Archive | Home )

Guest Authors
Site Search
Monthly Archives
Syndication
RSS 1.0
RSS 2.0
Check out the The AppGap - a group blog on the tools and trends that are changing the way we work.

Many-to-Many

« SlideShare -- the YouTube of Powerpoint | Main | danah profiled »

October 10, 2006

comScore misinterprets data: MySpace is *NOT* gray

Email This Entry

Posted by danah boyd

Read the ComScore press release. Completely. Read the details. They have found that the unique VISITORS have gotten older. This is not the same thing as USERS. A year ago, most adults hadn’t heard about MySpace. The moral panic has made it such that many US adults have now heard of it. This means that they visit the site. Do they all have accounts? Probably not. Furthermore, MySpace has attracted numerous bands in the last year. If you Google most bands, their MySpace page is either first or second; you can visit these without an account. People of all ages look for bands through search.

Why is Xanga far greater in terms of young people? Most adults haven’t heard of it. It’s not something that comes up high in search for other things. Facebook’s bimodal population pre-public launch shows that more professors/teachers are present than i thought (or maybe companies are more popular than i thought? or maybe comScore’s data is somehow counting teens/college students as 35-54…).

Can someone tell me exactly how comScore measures this? Is it based on the known age of the person using a given computer? Remember that many teens are logging in through their parent’s computer in the living room. Is it based on reported age? I kinda doubt it but the fact that there are more 100+ year olds on MySpace than are living should make people think about reported data. Is it based on phone interviews? How do they collect it? This isn’t really parseable into English.

My problem is that all of these teen sites show a heavy usage amongst 35-54. I cannot for the life of me explain how Xanga is 36% 35-54. There’s just no way. I don’t get how the data is formulated but it seems like an odd pattern across these sites to see a drop in 25-34 and a rise in 35-54. Older folks aren’t suddenly blogging on Xanga. So what gives? My hunch is that comScore’s metrics are consistently counting teens as 35-54 across all sites. My hypothesis is that because comScore is measuring per computer and teens are using their parent’s computer, comScore can’t tell the difference between a teen user and a parent user. If so, maybe all this is telling us is that parents have definitely listened to the warnings over the last year and are now making their teens access these sites through their computer?

Finally, when we talk about data, we also need to separate Visitors from Active Users from Accounts. The number of accounts is not the same as the number of users. The number of visitors is not the same as the number of users.

All this said, there is no doubt that more older people are creating accounts. Parents are told that they should check in on their kids. Police officers, teachers, marketers… they are all logging in to look at the youth. Is that the same as meaningful users? Some yes, some no.

From my qualitative experience, the vast majority of actual users are 14-30 with a skew to the lower end. Furthermore, the majority of the accounts are presenting themselves as 14-30. To confirm the latter (which is easier), i did a random sample of 100 profiles with UIDs over 50M (to address the “last year” phenomenon). What i found was:

  • 26 are under 18
  • 45 are 18-30 (with a skew to the lower)
  • 10 are over 30 but under 70
  • 1 is over 70 (but looks less than 18)
  • 6 are bands
  • 11 are invalid or deleted
  • 1 is complete fake characters (explained in descript)
A few more things of note…
  • 18 have private profiles
  • Of those over 30, only 2 has more than 2 friends (one has 3 friends; one has 5)

This account data hints that the general assumption that approximately 25% of users are minors is correct. Of the remaining, the bulk is under 30. Qualitatively, i’m seeing the most active use from those under 21. Given account practices, i don’t think that i’m off in what i’m seeing.

I do suspect that MySpace is holding strong at being primarily for younger people but that older folks have definitely been checking it out a LOT more. Still, i’m still suspicious of the fact that 35-54 are common across all youth sites. I’d really like to see comScore’s data on something that we can check. Maybe LiveJournal?

(I’d really really really love to be proven wrong on this. If anyone has data that can provide an alternate explanation to the comScore numbers, please let me know!)

Update: Fred Stutzman and i just jockeyed back and forth to find something we could agree on wrt the comScore numbers. Here are some ways of making sense of the data of VISITORS:

  • Xanga is more of a teen-flavored site than MySpace, Facebook or Friendster
  • Facebook is more of a college-flavored site than MySpace, Friendster or Xanga
  • Friendster is more of a 20/30-something flavored site than MySpace, Facebook or Xanga
  • Of users going to these four sites, MySpace does not swing to any one group; it draws people of all ages to visit the site.
  • A greater percentage of adults (most likely parents) visit MySpace than any of the other social sites

This is all fine and well and confirms most intuition. The problem is that what we CANNOT confirm via this data is that more adults visit any of these sites than minors. Again, intuitive but the comScore data seems to indicate that adults visit each of sites more than their key population. This is really visible in their “total internet” users which seems to suggest that the vast majority of visitors to all of these social sites are adults. I cannot find a single person who works for one of these companies that believes this.

I’ve spoked to numerous folks since i posted last nite. Most believe that comScore gets this data by running a program on people’s computers. Young people are supposed to use a separate account than their parents. This data seems to indicate that comScore is wrong in assuming that people will do so. Most minors probably use their parent’s account to check these social sites. So, if we assume that, Xanga is obscenely a teen site, Facebook probably has nearly as many high school users as college users and MySpace swings young but is used by a wider variety of age groups than most social sites.

Finally, it’s all nice and well that Fox Interactive spokespeople confirm this data but i’ve watched over and over as FIM has confirmed or said things that were patently untrue in public. I don’t know if this is because FIM (the parent of MySpace) doesn’t know what’s going on on MySpace or if it’s because they don’t care whether or not they are accurate publicly. I don’t honestly believe that FIM has any clue about the age of its unique visitors. They know the purported age of people who have accounts and it would be patently false to say that 35-54 dominates account holders.

Frankly, i’m uber disappointed with comScore but even more disappointed with all of the press and bloggers who ran with the story that MySpace is gray without really looking at the data. This encourages inaccurate data and affects the entire tech industry as well as policy makers, advertisers, and users. I’m horrified that AP, Slashdot, Wall Street Journal, and numerous respectable bloggers are just reporting this as truth and speaking about it as though this is about users instead of visitors. C’mon now. If we’re going to fetishize quantitative data, let’s at least use a properly critical eye.

Comments (7) + TrackBacks (0) | Category: social software


COMMENTS

1. Visitor on October 13, 2006 9:53 PM writes...

How can you knock comscore when FIM agreed to the metrics? You are throwing a lot of bogus accusations around.

'I don’t honestly believe that FIM has any clue about the age of its unique visitors. '

The standard practice in web measurment is "unique" visitors. Did I just not visit your site? Don't your logs count me as a visitor? I may become a repeat visitor (doubt it) which may entail me to be a user down the road.

Perhaps myspace is now a mainstream site now representing the average profile of the internet user. You shouldn't make negative accusations.

Permalink to Comment

2. Anthony on October 16, 2006 3:07 PM writes...

There's some interesting data about LiveJournal at http://www.livejournal.com/stats.bml

Social networking sites tend to draw groups of friends into them - most of my real-life friends are on livejournal, because at some point the number of people in my social circles on LJ reached a critical mass, and creating an account became worthwhile for communication. Similarly, Tribe.net has become *the* place for BurningMan folks, as well as several other "alternative" communities. I can sell Tribe to almost anyone who is a part of those communities, but to someone who is not in any of those communities, Tribe isn't worth the time.

Permalink to Comment

3. Brian O' Hanlon on October 17, 2006 12:39 PM writes...

I wrote some generic stuff here, on my thoughts at the moment, which I hope may provoke some thoughts.

B.

http://www.cooperationcommons.com/cooperation-commons/remember-lateral-thinking

Permalink to Comment

4. Doug Neal on October 22, 2006 8:31 PM writes...

Interesting data, it would be great to see more statistics like these on a larger data sample.

I'd be curious what your thoughts are on a new website we are launching which aims to provide professional and social networking for a specific vertical market segment (legal professionals). Let me know what you think:

http://www.lawyer-link.com

-Doug

Permalink to Comment

5. fp on October 26, 2006 9:10 PM writes...

Most of my Internet friedns and colleagues are over forty. Most of them have a little corner of MySpace where they dabble in MySpace-ishness. Some of them are musicians and tehy do more than dabble. The demographics really don't interest me, but MySpace has collected large numbers of people of all ages.

Your anecdotal sample shows that most people who use it are your age. My anecdotal sample would confirm the opposite, that most people who use it are... well, not as old as me certainly, but pretty damn gray.

Permalink to Comment

6. Sean Howard on November 2, 2006 11:30 AM writes...

Danah,

Brilliant post. I am amazed at how people view Internet Usage data as if it was a Holy Book.

You certainly seem to have struck on a nerve. Wow. I did not expect this article to create the negative comments it did. It's like you attacked the older generation or something. Oh wait. I'm one of them. ;)

It's a brilliant look at some assumptions that are driving media and accepted as truth by everyone. That the youth are leaving in droves. Here's a great example. A well written article in the Washing Post that is based almost entirely on some comments from a sample of one class of teens.

http://www.washingtonpost.com/wp-dyn/content/article/2006/10/28/AR2006102800803.html

I've been trying to put my finger on why this article in the post bothered me. Sure, Teens are fickle. But why this quick jump to them "abandoning" all the current social networking sites... Your article helped to paint light on how this is a potential myth fed from many sides.

I don't expect an answer as I know you are busy as all get out. Great article. Looking forward to your next one.

Sean

Permalink to Comment

7. Jack on January 10, 2007 12:07 AM writes...

comScore tracks which users in a household are accessing the computer at any given time by analyzing each user's type signature.

There is no "account login" at all as you suggest in your post.

The technology is built by collecting, recording and analyzing each user in every households' typing patterns. That information is used to produce a type signature which can identify that user whenever he or she is using the computer.

This is probably best illustrated with an example:

John's (age 13) type signature is that it consistently takes him .5 seconds to hit an i after typing an h, .743 seconds to hit a w after hitting an a, .134 seconds to hit an n after an a, etc etc for all consistent paterns in the way John types. He also types at 64 wpm on average.

John's dad's (age 45) type signature is that it consistently takes him .924 seconds to hit an o after typing an f, 1.34 seconds to hit a t after hitting an e, .543 seconds to hit an r after an e, etc etc for all consistent paterns in the way John's dad types. He also types at 22 wpm on average.
etc. etc.

This is done for an entire family.

comScore then simply compares how a user is typing to the type signatures of the different users in the household to determine which household member is using the computer at the time of use.

That is how comScore distinguishes between family members using a single computer. They may also be using click signatures to distinguish between users as well (see this study done by Wharton Prof. Balaji Padmanabhan & UCDavis Prof. Yinghui Yang - http://knowledge.wharton.upenn.edu/papers/1323.pdf?CFID=720523&CFTOKEN=57530247). Not sure though.

Jack

Permalink to Comment

TrackBack URL:
http://www.corante.com/cgi-bin/mt/teriore.fcgi/53145.

POST A COMMENT




Remember Me?



EMAIL THIS ENTRY TO A FRIEND

Email this entry to:

Your email address:

Message (optional):




RELATED ENTRIES
It's Live! New JCMC on Social Network Sites
Race/ethnicity and parent education differences in usage of Facebook and MySpace
User-generated neologism: "Indigenous content"
history of social network sites (a work-in-progress)
New Freedom Destroys Old Culture: A response to Nick Carr
responding to critiques of my essay on class
Tagmashes from LibraryThing
Spolsky on Blog Comments: Scale matters