Wall Street Journal , NY
May 26 2005
Measuring the Impact of Blogs
Requires More Than Counting
May 26, 2005
If you read press coverage about blogs, you might conclude that just
about all Americans are reading a blog. But then you wouldn’t have
time to read the press coverage, because if surveys are to be
believed, you’re probably busy creating your own blog.
The numbers of the blogosphere range widely. Are there 10 million
blogs, or 32 million? Do a quarter of online Americans really read
blogs, as one oft-cited survey found? And why do rankings of the most
popular blogs vary so much?
Adding to the confusion: disagreement over exactly what a blog is. In
our young era of blogging, there’s still no consensus. “Blog” derives
from “Web log,” and everyone agrees that a blog should be regularly
updated, with new entries in reverse chronological order — and that
the entries can be about anything. But millions of people establish
blogging accounts with free software providers like Google Inc.’s
Blogger, Microsoft Corp.’s MSN or Six Apart Ltd.’s LiveJournal — it
takes mere minutes — and then never post to their blogs. Others
password-protect their blogs and use them to share photos and data
with a small group of family members, friends or colleagues. Whether
or not you count all those represents a big chunk of the swing from
10 million (cited recently in the New York Times and USA Today) to
31.6 million blogs (Ottawa Citizen and the Ann Arbor News). Both are
world-wide estimates.
First, let’s step back and consider why we’re counting blogs at all.
You no longer see articles that attempt to demonstrate the legitimacy
of the Web by stating how many Web pages there are. But blogs are
still in the process of entering mainstream consciousness, so
numerical credibility is important; bloggers themselves cite the
statistics a lot.
ABOUT THIS COLUMN
The Numbers Guy examines numbers and statistics in the news,
business, politics and health. Some numbers are flat-out wrong,
misleading or biased. Others are valid and useful, helping us to make
informed decisions. As the Numbers Guy, I will try to sort through
which numbers to trust, question or discard altogether. And I’d like
to hear from you at [email protected]. I’ll post and respond to your
letters. WSJ.com subscribers can sign up to receive e-mail when new
columns are published (nonsubscribers click here to sign up), and you
can read more columns at WSJ.com/NumbersGuy.
It turns out that counting blogs isn’t as hard as counting Web pages.
When writers who use common blogging software want their blogs to be
publicized, they choose to automatically “ping” computer servers for
companies like Technorati Inc. () and Intelliseek’s
BlogPulse (), whose goal is to measure and index
blogs. Then Web users can go to those companies’ Web sites and run
searches to find blogs that have written about topics they’re
interested in. BlogPulse now indexes about 11 million blogs
world-wide; Technorati, about 10 million. Over the past six months,
both have seen a doubling in the number of blogs on the Internet.
“Nobody asks how many Web sites there are out there,” Natalie Glance,
a researcher with BlogPulse, told me. “We’re fortunate that all these
ping servers do exist. But it’s really a hard question to answer.”
That’s because not all blogs ping the search services. In Korea, some
software providers say they have millions of blogs, and neither
BlogPulse or Technorati count them all. “It’s pretty significant
undercounting,” Ms. Glance said. David Sifry, Technorati’s chief
executive, told me at a blogger conference last week that he was
headed to Seoul later that week to try to get Korean blogs in his
index. (Already, about two-third of the blogs indexed by Technorati
are in languages other than English.)
Technorati and BlogPulse both define blogs as being meant for public
consumption. This is an important distinction because Internet
companies seeking to cash in on the surge in blogging have rolled out
products that combine blogging software with other tools like
photo-sharing and social-networking services. When you create an
account with one of these companies, you’re considered to have a
blog, even if you never write a post. The same goes if you restrict
access to a select group of readers. Microsoft’s MSN Spaces says it
has 10 million accounts, but a spokeswoman says more than half of
those accounts are available only to a restricted set of users.
Meanwhile, BlogPulse’s Ms. Glance says that half of MSN Spaces blogs
appear to be blank, based on her research.
Some analysts have tried to count private blogs. Perseus Development
Corp., a Braintree, Mass., market-research company, last month
reported 31.6 million blogs, using an unusual approach: It added
reported numbers of blogs from companies like MSN, with its own
projections for number of blogs for companies like Google that don’t
disclose stats. It arrived at the projection by forming random
strings of letters, and then searching to see if those letters
corresponded to a blog on the service. Services with lots of matches
were assumed to be hosting more blogs than those with fewer matches.
“We tried to extend the random-digit dialing from the telephone world
into the blog world,” Jeffrey Henning, chief operating officer of
Perseus, told me.
(The Blog Herald, a blog about blogs, counted over 60 million blogs
this week, relying on figures from operators world-wide.)
No one has sole control of the definition of blog, but it seems to me
that for the sake of counting, Technorati and BlogPulse are right to
exclude the private blogs. That puts their estimates below those from
some other analysts, but the companies are focusing on what they can
directly count, and relying less on estimates.
Still, the number of blogs isn’t really that informative, since so
many blogs are abandoned soon after they’re launched. It’s more
useful to look at the volume of blog posts. According to a
presentation by Technorati’s Mr. Sifry at the blog conference, daily
volume is 800,000 to 900,000 posts. But Ms. Glance says BlogPulse,
which says it has more blogs in its index, counts only between
350,000 and 450,000 posts a day — and that number has held steady
for about a year, even as the total number of blogs has accelerated.
Regardless of who’s right, notice that these number are well below
estimates for the total number of blogs, countering the image of
blogging as a multiple-times-a-day activity. Ms. Glance says that
based on her research of activity in January, the typical active
blogger posted an update just once every 10 days.
The total number of active blogs — those with a post in the past 30
days — was 3.5 million on May 1, according to BlogPulse. That was up
just 30% from last September, even as the site found that the total
number of blogs increased nearly 200% over that time. That suggests
there’s a lot of dead air out there.
Pinning Down Readership
The number of blogs doesn’t tell us much about the medium’s
relevance. How many people are reading blogs?
In a telephone survey of U.S. Internet users last fall, the Pew
Internet & American Life Project found that 27% of respondents said
they read blogs. (Users were asked: “Please tell me if you ever do
any of the following when you go online. … Do you ever read someone
else’s web log or blog?”) But in the same survey, Pew asked: “In
general, would you say you have a good idea of what the term Internet
‘blog’ means, or are you not really sure what the term means?” Just
38% of Internet users answered “yes.” Of the 27% who said they read
blogs, about 40% answered “no” to the blog-awareness question.
Some of those people who didn’t really know what blogs are, yet say
they read them, may have been feeling social pressure. At the height
of election season, blogs were being covered heavily in the press and
bloggers were seen by some as celebrities and expert analysts. Survey
respondents who considered themselves sophisticated political
observers may have wanted to be reading blogs, even if they didn’t
know how to find them. As I discussed in a previous column, social
pressure can skew survey results to yield numbers that aren’t
credible.
“Sometimes people don’t want to say, ‘I don’t know,'” Lee Rainie,
director of the Pew Internet & American Life Project, told me.
However, he doesn’t think social pressure played that big a role
here: “It’s not the same as saying, would I spend my tax dollars to
clean up the environment.”
In an update to the survey earlier this year, Pew reported that 25%
of Internet users said they read blogs — a small decrease from the
2004 results. The blog-awareness question wasn’t asked that time
around. Multiplying the results with Pew’s estimate for the total
number of Americans online yields an estimate of about 32 million
American adults who read blogs. That number is frequently cited in
press coverage of blogs.
How Important Are They?
Even if millions of Americans read blogs, there are very few
individual blogs that have a significant number of readers. Several
Web sites attempt to rank blogs for popularity, but it’s not always
clear how they arrive at their numbers.
Little Green Footballs (littlegreenfootballs.com) — Charles
Johnson’s conservative blog that rose to the top echelon of blogs
with its coverage of CBS’s flawed report on President Bush’s National
Guard service — is ranked No. 4 in rankings published on The Truth
Laid Bear, a site often monitored by bloggers. But Little Green
Footballs is 12th on Technorati’s top 100 list. Both rankings
evaluate blogs based on how often they are linked to by other Web
sites, though Truth Laid Bear limits its universe to an “ecosystem”
of about 23,000 blogs, thereby diminishing the number of blogs in
contention and the number of incoming links.
Evaluating a site based on how many other sites link to it has some
validity — it provides a measure of reputation and name recognition,
and Google uses a similar strategy to deliver its search results. But
this measure fails to take into account the prominence of the site
doing the linking. (A link from a high-profile blog would likely
deliver more traffic than dozens of links from unknown blogs.)
There is a more straightforward measure of popularity: The number of
visits to the sites. Truth Laid Bear has a traffic ranking, which
counts average daily visits to each blog. On that ranking, Little
Green Footballs ranks No. 7, with 87,155 visits per day, behind
liberal blog Eschaton. Yet Blogads.com, a blog-advertising network,
shows Little Green Footballs with a weekly average of more than
752,000 visits, ahead of Eschaton.
Advertisers may not be happy with either number, since they count
total visits, and not the “unique visitor” figure that is the
standard currency for many kinds of online advertising (advertisers
don’t want to pay twice to reach the same reader). “That’s a big
issue,” Henry Copeland, founder of Blogads.com, told me at a
conference last week. “We’re very aware that’s a flawed number.”
The anonymous publisher of Truth Laid Bear, who uses the name “N.Z.
Bear,” told me in an e-mail, “My systems are lousy for making
definitive pronouncements about the entire scope of the blogosphere
… But for the blogs I *do* track, they offer pretty good data,
especially in comparing between blogs.”
ComScore Media Metrix and Neilsen//NetRatings are the sources most
often used by online advertisers to track unique visitors. Neither
tracks blogs as a matter of course, though comScore did look up
traffic for 13 prominent blogs in April, upon my request (I picked
ones from the top of the various rankings). Just five met the
company’s minimum threshold for statistical significance of about
150,000 monthly visitors. Media and gossip site Gawker had the most,
with 304,000 unique visitors. The others that cleared the cut:
Defamer (287,000), Boing Boing (250,000), Daily Kos (212,000) and
Gizmodo (209,000). Among those that didn’t were prominent political
blogs Instapundit, Power Line and Eschaton. (I asked NetRatings about
the same 13 blogs, and it had reportable data only for Defamer, Daily
Kos, Boing Boing and Gizmodo — and the sample sizes didn’t meet
standards for statistical significance.)
ComScore and NetRatings both recruit panels of online users who agree
to install software that monitors their behavior. The companies use
sampling techniques similar to those of political pollsters.
By point of comparison, comScore says the New York Times’s Web site
had 29.8 million unique visitors in April.
* * *
My column last week on the number of Armenians who died in mass
killings and deportations in the Ottoman empire 90 years ago sparked
a lot of mail, including several letters criticizing the column for
minimizing the deaths and for seeming to set a numerical threshold
for determining if genocide was committed. Neither was intended by
the column, which looked behind a historical death toll that is often
repeated without explanation, to examine how it was calculated.
Armenian groups and advocates who are pushing Turkey to call the
killings a genocide often cite death tolls in their accounts.
Here are some letters, edited for space and clarity:
Dr. Papazian made the most salient point in your article — getting
lost in the numbers shouldn’t divert attention from what is most
important, which is that the Armenians were subjected to a methodical
and diabolical genocide perpetrated by the Turkish government and its
constituents. … Using numbers to play politics with the lives of
those who died for their cultural identity and their religion is a
shameful game that only the guilty and the conscienceless would play.
— Peter Abaci
A point you seemed to have completely ignored in your article were
the number of Muslims killed. This is a scary issue and the
psychology behind the bloodshed in the Middle East. A dead Christian
gets counted but no one cares about a dead Muslim. The McCarthy study
touches on that issue very well. … Also why is McCarthy the only
academic you mention from a Turkish-friendly point of view?
— Omer Koker
I am an Armenian descendent living in Brazil. … You can ask any
Armenian you find in Diaspora, in any country of the world, and
you’ll find that they had lost relatives during the genocide. In my
case, I lost three of my grandparents, burnt inside a church, among
other relatives. … It’s not a game of numbers, it’s a question of
conscience and justice.
— Andrew R. Apovian
Your first mistake is grossly undercounting the number of experts and
countries that affirm the Armenian genocide as a fact. At least 100
of the most renowned genocide and Holocaust experts in the world have
affirmed the genocide of 1915. Moreover, scores of countries, U.S.
states, municipalities, and international organizations have done so.
See here and here for lists of experts, countries, international
organizations, and others that have formally affirmed the Armenian
genocide. … The Armenian genocide has been studied to death, as
have the fatality figures. The genocide is a fact that has been
proven time and time again. You should not have presented it as
debatable, just so that you would appear to be “fair-minded.”
Another example of Number Guy’s undercounting is the absolutely
amazing statement that Armenians who died of “disease and starvation”
should probably not be counted as part of the killing. Numbers Guy,
the “deportations” are regarded as death marches by all experts on
those events.
— David B. Boyajian
Several readers cited the recognition of the genocide by scholars and
governments. But these recognitions don’t always cite specific
numbers of Armenians killed. And the politicians and scholars
generally didn’t directly study the death toll. What interested me
was the actual source of the numbers — the scholarship upon which
the numbers were based.
As for David’s second point: I didn’t say such deaths probably should
not be counted; I said it was debatable. The estimates of Armenian
deaths are derived indirectly from reduced population counts, from
all causes. There are no reliable counts of how many died directly
from Ottoman actions, hence the uncertainty.
Increased study into what you called the harbingers of genocide (such
as undercounting the targeted population) might even help improve
early detection and help save lives. Continuing to focus on counting
the dead in order to define genocide will only perpetuate debate
after-the-fact, which will always be too late.
— Basil Valentine
I appreciated your thoughtful article about how many Armenians died
in 1915. However, I am more optimistic about the chances of resolving
these disputes. Specifically, I have been campaigning for the setting
up of an independent international historical commission under the
umbrella of UNESCO [the United Nations Educational, Scientific and
Cultural Organization]. This is quite distinct from the recent
proposal of the Turkish government for a bilateral Turkish-Armenian
historical commission. The Turkish proposal is a significant step
forward and has been welcomed by many governments but contains
serious flaws — not least that the Armenians will not be willing
participants. The Armenians argue that no more research needs to be
done but this cannot be accepted when so many facts and documents are
disputed, and archives still closed. … A bilateral commission
composed of Turks and Armenians will inevitably be dominated by
political not scientific issues, filled with partisan historians
bickering over the validity of documents and definitions.
— Patrick Byrne, editor, Turkey In Europe online magazine