« The Politics of Denial - Fiscal Responsibility vs Un-Checked Power | Sacred Cow Dung Home | Let George Washington Spread Your Message | A Clever Analog Promotional Technique »

October 18, 2005

Measuring The ACTUAL Blogosphere Part 1 - Technorati View

Technorati 100 or Feedster 500?

10 million or 70 million blogs?

Confused? … So was I.

… Time For A Blogosphere Meta-Analysis.

Introducing "The Sacred Cow Dung 500 Index" (or the "Dung 500 Meta-Analysis")

This is the first of a series of posts that are the result of my efforts over the summer to try to get a better feel for the ACTUAL blogosphere.

There were a number of things happening in the blogosphere that I found puzzling and contradictory last summer. Not only were estimates of the total number of blogs wildly divergent (from 10m to 70m) but a rash of new weblog ranking schemes were being introduced which seemed to focus only on "Who's at the Top". And I couldn't reconcile my own blog's data (which seemed insignificant to me) with it's apparently high-ranking relative to the rest of the blogosphere.

The Rule of Five

As both a VC and an entrepreneur, I am often focused on new, emerging, or potentially emerging markets. While it may seem intriguing to look "up-market," most people who analyse markets regularly, with an eye towards market entry strategies, find it much more productive to get a sense of the "down-markets" as well.

When discussing market entry strategies, I often use my "inverted pyramids slide" which implies there are only three layers to uncover in any market.

Slide81invertedtwins

However, in practice, you generally must tease out five layers.

Market distributions are described mathematically as a subset of Binomial Distributions known as "Poisson Distributions" because the French mathematicians noted how often they looked like a "fish" -- with a small head, fat body, and long thin tail -- hence the French word "Poisson" for "Fish." (Oddly enough, the French mathematician was also named "Simeon Denis Poisson" -- How's that for co-incidence? Another example of Destiny in a name?)

But this tripartite anatomy also means there are two additional "transition zones" -- from body to head and from tail to body. Perhaps, it would be better to think less in terms of fish and more in terms of human anatomy -- head, neck, torso, legs, and feet. However, such is convention and tradition -- as misleading as it is useful.

"Tops and Bottoms" Are Generally Misleading

The trouble with tops and bottoms of markets is that they really don't tell you much about market dynamics. The tops and bottoms of markets tend to be rather static. Often the tops and bottoms exist only because of the definitional limitations of how you initially define a market. Also, what keeps something at the top of a market, or the bottom of the market, may not be relevant to understanding what is driving movements and layering within the "bulk" of the market.

Step 1 -- What Does the Curve Actually Look Like?

This first step is so fundamental to start to understand any market that I found the lack of any apparent effort to do so mind-blowing. All of the new indexes seem to be focused, not only on the top of the market, but on the top of the top of the market. Who the hell cares!?!? What about the rest of the curve -- where the action is might really lay?

Blind Men and the Elephant -- "Precision without Accuracy"

Not only were people asking the wrong questions (in my mind) but their "data-rich" analyses were coming to contradictory conclusions. This is the classic "Blind Men and the Elephant". If you have a hammer, you tend to use your hammer -- even if a screw-driver would be a better tool. No one is lying. In fact, they are precisely reporting their measurements of "Ears", "Trunks", "Feet", "Tails" etc. -- yet we are left you "no sense of an accurate depiction" of the whole animal.

Enter Meta-Analysis -- "The First Glimpse of the Elephant" -- "Accuracy without Precision"

I was first introduced to Meta-Analysis during my medical training in the 1980's. Physicians are no strangers to massive quantities of "non-definitive" "highly-precise" studies with contradictory results. And, like other professions, in the face of conflicting literature, physicians must make definitive decisions with real-world consequences to their patients lives -- long before the so-called "definitive studies" come out.

This is where meta-analysis comes in. It is powerful because it achieves "accuracy without precision" by enabling people to get a first glimpse of the elephant by pooling all the data of all the "blind men" and drawing new conclusions based of the aggregate of conflicting studies.

Purists hate meta-analysis because it is "invalid" to pool data from disparate studies, each with precise experimental designs design to "prove" their hypotheses. And they are right. However, meta-analysis has assumed an central role in the evolution of "evidence-based medicine" -- which I would generalize as --

The Three Stages in the Evolution of Consensual Truth

  • Stage 1 -- Diversity and Divergence -- Precision without Accuracy
  • Stage 2 -- Convergence -- Accuracy without Precision
  • Stage 3 -- Confirmation and Consensus -- Accuracy with Precision

Basically, meta-analysis is a way to facilitate Stage 2 Convergence. By depicting the vague outlines of the elephant, meta-analysis facilitates the transition to Stage 3 -- where the definitive elephant studies are created which confirm the existence, and generates the consensus view, of that elephant in detail.

[… more on Meta-Analysis in Later Posts and in the related links below.]

The Dung 500 -- A Blogsophere Meta-Analysis

So I spent a good part of August doing a series of ad hoc "good enough" analyses and wrote a long article which I've decided was just too long to make a good blog post.

While I was able to answer a number of questions that interested me at the time, I realized I could now go back and create a true index which would allow me to do an easy blogosphere meta-analysis.

So in September, I built a list of 500 weblogs. These weblogs were chosen to be a representative sample of different regions of the blogosphere. (I knew where to look from my August "first-cut" analysis.) Then, over a four-day marathon in mid-September, I collected data from multiple public sources on each of the weblogs.

Because this is about accurate depiction and not precision, I have restricted myself to public data sources and simple excel spreadsheet analyses and graphs. I will avoid using any "statistical mathemagic." If you can't plainly see it on a graph, then I don't care about it.

[ … More on methods in later posts.]

"Measuring the ACTUAL Blogosphere" Series Overview

I'm organising the "Measuring the Actual Blogosphere" posts into four groups --

  • The Blind Men Individually
  • The Blind Men Comparatively
  • The Blind Men in Aggregate
  • Questions We Can Answer About the Elephant

So, the next few posts, we will be looking at the Sacred Cow Dung 500 strictly from each Blind Man's View. Then we will start to compare the views and later begin to answer some basic questions about the blogosphere.

Today's Blind Man -- Technorati

For the sake of this discussion, Technorati is --

a blogosphere search engine which relies on "intra-blogosphere" link analysis.
So Technorati must do two things well for their engine to work --

  • Precisely distinguish a weblog from a "non-blog" web-site
  • Collect data on "blog-to-blog" linking

Data which anyone can easily collect from Technorati (with a "little" effort) includes --

  • Unique blog identification
  • Common blog "ownership" via "Technorati Profiles"
  • The number of unique blogs which link back to a given blog
  • The total number of blog links which link back to a given blog
  • And a Technorati-derived Ranking of blogs

The Technorati Component of the Dung 500

It's always good just to plot the damn data, so here goes --

1dung500technoraticrankvblogsandbloglinks
Figure 1. Technorati Rank vs Unique Blogs Linking and Total Blog Links

In Figure 1, we are doing a simple plot of Technorati rank vs the number of unique blogs linking back and the total number of blog link-backs.

Notice how the data tends to "hug" both axes. For those of you who don't routinely look at raw market data in general -- and internet data in particular -- you may find this a bit of a shock. However, this is fairly typical. While people love to quote market sizes in general because it sounds "big" -- most traditional "market metrics" tend to be dominated by a market "oligopoly". Significant emerging trend data tends to be overwhelmed, or washed out, by the magnitude of the top few players.

Few people realize that over 70% of internet traffic is funnelled through a handful of "mega-portals" (Yahoo, Google, etc.) and a straight plot would show a similar relative insignificance of the rest of the web.

In the blogosphere, the situation is worse since there is a much lower barrier of entry and maintenance cost for blogs than traditional web-site -- both in time and in money. Traditional web-sites cost money to develop and to host. This keeps both supply and growth relatively controlled. Not so in the blogosphere. Anyone can spend an hour on blogger.com and create a dozen blogs which are never used again. People often test out the various platforms and generate dozens of "test blogs" before starting their one "real blog" (or, wisely, abandoning blogging as a total waste of their valuable time).

The numbers of blogs are artificially inflated vs the rest of the web, partly, because of ease of creation and, partly, because there is no motivation to "clean up." Very few people ever delete their test sites since they don't pay for their hosting anyway. Hence the large percentage of "dead blogs" that are counted as part of the blogosphere.

If people were charged even $1 per month per blog, it would be interesting to see how many blogs would remain.

[… more on "blog inflation" in later posts..]

The Technorati 100

The first step to "get data off an axis" is to switch to LOG SCALE.

2dung500technoraticrankvblogsandbloglinkslog
Figure 2. Technorati Rank vs Unique Blogs Linking and Total Blog Links (LOG SCALE)

Figure 2 is the same as Figure 1 except we've changed the x-axis (Technorati Rank) to a LOG SCALE. This allow us to get a better view of what's happening in that crowded area around the y-axis by systematically distorting the view to over-emphasis the right over the left.

The first thing you see is "Why the Technorati 100." This is why most of the data shown by Technorati is from this region. The link analysis appears to work well for the first 100 weblogs.

The second thing you see is that while both total link-back and unique blogs linking correlate well, the number of unique blogs correlates the best with Technorati rank. This should not be surprising since that's how the Technorati ranking works -- each blog carries one vote and any additional votes don't count.

This "one blog one vote" keeps Technorati link-analysis-based ranking internally consistent but quickly breaks down as we will see below. Unlike Technorati's democratic approach, Google follows a "not all links are created equal" approach to link analysis.

[… more on this later.]

Technorati Rank - Link vs Site Correlation

I also plotted a LOG-LOG version of Figure 1 as Figure 3.

3dung500technoraticrankvblogsandbloglinksloglog
Figure 3. Technorati Rank vs Unique Blogs Linking and Total Blog Links (LOG-LOG SCALE)

This is clearly invalid as you hit the region beyond 100,000 when you see an increasing portion of blogs with no blogs or links linking to them (logs don't work for numbers less than 1). So it over-represents the links for those blogs ranked above 100,000 by eliminating all of the zeros and ones.

Nevertheless, it does suggest the presence of different region of the blogosphere behaving differently. The Top 10 vs The Next 10,000 vs The Next 100,000 vs the Next 1,000,000.

Also, you notice that while unique blogs still correlate well throughout the data range, the Total Blog Link-backs correlation begins to fall apart for blogs ranked 10,000 or higher.

Technorati Rank 1 to 100

I plotted the top 100 Technorati Ranked weblogs - first with LOG SCALE and then Linear Scale.

4dung500technorati1to100vblogsandbloglinkslog
Figure 4. Technorati Rank 1 to 100 (LOG SCALE).
5dung500technorati1to100vblogsandbloglinks
Figure 5. Technorati Rank 1 to 100.

Both Figure 4 and 5 point to a significant difference between the first 10 and the remaining 90 blogs of the Technorati 100.

Also notice the changing slope in the various views. Changing slope tells you the changing relative weight each incremental blog carries in determining your rank. For the top 10 blogs, it takes thousands of blogs to move them up by only one rank.

Technorati Rank 100 to 1000

6dung500technorati100to1000vblogsandbloglinks
Figure 6. Technorati Rank 100 to 1000.

The Technorati rank value of each incremental blog is a lot more for this group. It generally takes around 1 additional blog to move you up a rank in this range -- this is the "one-for-one" range.

Technorati Rank 1000 to 10,000

7dung500technorati1000to10000vblogsandbloglinks
Figure 7. Technorati Rank 1000 to 10,000.

The Technorati Rank region from 1,000 to 10,000, is were you begin to see the first signs of trouble with Blogosphere Link Analysis.

Basically, you only need about 100 blogs to link to your blog to get its rank into the top 10,000. As few as 500 incremental blogs will move you into the top 1,000 blogs. But that means each blog linking is worth about 20 in rank.

But wait! … Doesn't that mean that if I have a blog ranked at 5,000 because of 200 unique blogs link to it, there are about 10 blogs ranked above me, and 10 blogs ranked below me, who also have 200 links. So how can Technorati tell the relative rank of blogs 4990 to 5010 if they all have the same number of blogs linking to them?

Although, it hard to show graphically, if you look at some of the data in table form, you see that Technorati may not really be "one blog one vote". While this approach works when you have many unique blogs per rank, when you have many ranks with the same number of blogs, you have to rely on something else.

This seems to be the role for total blog-links. Total links seems to enter as a way to distinguish ranks once blogs alone fail to discern any difference -- which begins to occur in the 1,000 to 10,000 range. This maybe a secondary algorthim used to extend the "useful" range of blogosphere link analysis.

Technorati Rank 10,000 to 100,000

Since Sacred Cow Dung lives in this region, it was here that I began to see some red flags early on in my "blogging life.".

When Sacred Cow Dung was ranked at 50,000, it only had 37 unique blogs linking to it with a total of 52 blog-links. But then I noticed that another blog with 38 blogs linking to it was ranked around 48,000 and another with 36 blogs was ranked around 52,000.

In other words, the incremental value of one blog in this region is about 2000 in Technorati ranking -- which figures 8 through 10 confirm.

But how can Technorati relatively rank 2000 blogs with the exact same number of blogs linking? Sure, an additional 50 total links might help a bit but it seemed clear to me that the granularity of the Technorati intra-blogosphere link analysis is just not capable of such a feat. How could it? Answer: It can't and it doesn't.

8dung500technorati10000to100000vblogsandbloglinks
Figure 8. Technorati Rank 10,000 to 100,000 | Blogs and Blog-links.

9dung500technorati10000to100000vblogs
Figure 9. Technorati Rank 10,000 to 100,000 | BLOGS ONLY.

In mid-September I was still ranked about 50,000 with 37 blogs when I collected the Sacred Cow Dung 500 Data. Shortly afterward, Technorati suddenly "found" (or "doubled") the total number of blogs that linked to me. I was "catapulted" from 50,000 to 20,000 in rank in 24 hours. All with no change in traffic. Hmmmmm.

10dung500technorati10000to100000vbloglinks
Figure 10. Technorati Rank 10,000 to 100,000 | BLOG-LINKS ONLY.

Also note that in this region, there are starting to be major deviations between blogs with "lots of blog-links" and Technorati rank.

Technorati Rank > 100,000

The problems you see with Technorati Link Analysis are compounded when to go beyond 100,000 - or when the total number of blogs linking drops below 20.

11dung500technorati100000plusvblogsandbloglinks
Figure 11. Technorati Rank > 100,000.

12dung500technorati100000plusvblog
Figure 12. Technorati Rank > 100,000.

While the number of unique blogs linking remains internally consistent with the ranking -- the value of each blog linking has sky-rocketed. Depending on where you are in this range, one blog-linking to you blog is worth well over 50,000 in rank. In fact, I was unable to find any blogs ranked higher than 800,000 in September. And the blogs I did sample share the same identical rankings. In other words, the data was beginning to "bunch up" or "clump" into groups with the same identical rank. Although it's hard to tell from these charts, these data points represent "many" blogs with identical ranks superimposed on each other -- with no data in between.

13dung500technorati100000plusvbloglinks
Figure 12. Technorati Rank > 100,000.

Your First Blog-Link is Worth More Than 18 million in Technorati Rank

2 links puts you into the top 500,000 blogs and, somewhere between 500,000 and 1 million, the links go to zero.

So --

If there are No Links, There's No Link Analysis.

In other words, if we use Technorati's current estimate of over 18 million blogs, 95% of the total number of blogs out there -- HAVE NO LINKS LINKING BACK TO THEM.

Therefore, Technorati Link Analysis only works for the top 5% that actually have links and is rendered useless to measure 95% of the blogosphere.

Or, if you are just starting out and want your blog to move into the top 5% of the blogosphere (according to Technorati Rankings) -- just go to blogger.com and create one "dummy" blog with one link to your blog and that single link is worth over 18 million in rank and drives you into the top 5% of all blogs.

Pretty cool … but also pretty stupid.

In Summary, The Blogosphere According to Technorati

As of 10/7/05,the total number of blogs "measured" by Technorati = 18,900,000

  • Less than 450 blogs have over 1000 blogs linking to them = 0.002 %
  • Less than 15,000 blogs have over 100 blogs linking to them = 0.08 %
  • Less than 180,000 blogs have over 10 blogs linking to them = 1 %
  • Less than 800,000 blogs have at least 1 blog linking to them = 4 %
  • Over 18,000,000 blogs have 0 blogs linking to them = 95%

In other words, using link analysis, Technorati can only "rank" among the top 5% of all blogs and -- judging from their data -- their methodology starts to fall apart pretty quickly after the top 0.1%. This makes sense since the relative value of a link depends on where you are in the curve.

  • For the Technorati 10, it takes thousands of links to move one place in rank.
  • For the Technorati 100, it take hundreds of links to move one place in rank
  • At 50,000, one link is worth over 2,000
  • and your first link is worth over 18 million.

Other Interesting Technorati Data points.

Technorati has something it calls "Technorati Profiles". If one has a Technorati Profile, one can "claim" your blogs in one place and see their relative ranks and their link information in one place. Links to these profiles can be posted on their respective blogs.

Oddly enough, if you search "Technorati Profile" in Google but restrict it to the Technorati Domain, you get less than 700 and this hasn't changed much recently. It suggests that less than 1000 bloggers have actually ever bothered to claim their blogs. However, most of these bloggers have more than one blog claimed so they may represent up to 2 or 3 thousand blogs. And that's just the one's they have claimed, it uncommon to find profiles which include the dozens of "now dead" test-blogs these bloggers authored at one time (another couple thousand blogs by my best "guesstimate".

I'm not sure what this means -- yet.

BTW, I'm not picking on Technorati. They have clearly delivered what they have set out to deliver. The issue is what it all means.

It is the "inflationary biases" of the gungho blogosphere pundits and conference promoters that needs to be held in check by a little taste of reality from time to time.

Preview of "Measuring The ACTUAL Blogosphere Part 2 - Google View"

In the next post of this series, I'll be cuing up the next blind man -- Google.


"Spinning a Tall Tale out of a Long Tail"

As an aside:

For those of you whose livelihoods depend on "Spinning a Tall Tale out of a Long Tail" (you know who you are), I've included the following slide which reinforces everyone's preconceived notion of the blogosphere -- without actually lying.

I've merely eliminated the inconvenient "log scale" labelling from figure 3 to preserve your audience's currently distorted impressions of the ACTUAL blogosphere.

You won't get lynched at a Weblog, or Web 2.0, conference, if you use this slide. Proceed at your own risk if you use Figure 1.

3bconsensusview ofblogosphere1
Figure 3b. The "Politically Correct" Version -- Safe to Show at Conferences of "True Believers".

Related Links

Technorati

Feedster

The Long Tail

Blog Ranking / Popularity

Sites Dedicated to Blog Measurements

Poisson Distributions

Meta-Analysis

In statistics, a meta-analysis combines the results of several studies that address a set of related research hypotheses. The first meta-analysis was performed by Karl Pearson in 1904, in an attempt to overcome the problem of reduced statistical power in studies with small sample sizes; analyzing the results from a group of studies can allow more accurate estimation of effects.

Meta-analysis is a collection of systematic techniques for resolving apparent contradictions in research findings. Meta-analysts translate results from different studies to a common metric and statistically explore relations between study characteristics and findings.

Although meta-analysis is widely used in evidence-based medicine today, a meta-analysis of a medical treatment was not published till 1955. In the 1970s more sophisticated analytical techniques were introduced in educational research, starting with the work of Gene V Glass. The online Oxford English Dictionary lists the first usage of the term in the statistical sense as 1976 by Glass. The statistical theory surrounding meta-analysis was greatly advanced by the work of Larry V. Hedges, Ingram Olkin, John E. Hunter, and Frank L. Schmidt.

A weakness of the method is that sources of bias are not controlled by the method. A good meta-analysis of badly designed studies will still result in bad statistics. Robert Slavin has argued that only methodologically sound studies should be included in a meta-analysis, a practice he calls 'best evidence meta-analysis'. Other meta-analysts would include weaker studies, and add a study-level predictor variable that reflects the methodological quality of the studies to examine the effect of study quality on the effect size.

The overall goal of this paper was to acquaint the reader with the procedures and assumptions involved with a Hunter and Schmidt meta-analysis. Meta-Analysis provides a strong alternative to the more traditional review methods, and allow for quantitative conclusions to be reached.

Over the last 15 to 20 years there has been an increased criticism of social sciences research because of the confusing state of the research literature. While one reviewer could find a set of studies which supported his viewpoint, a second reviewer commonly found several which did not support said conclusions. A common conclusion in reviews was "Conflicting Results In The Literature, More Research Is Needed To Resolve This Issue." Which typically resulted in more studies which did nothing to clarify the issue.

Meta-analysis offers a way out of this quagmire. By using carefully constructed and comprehensive coding and accumulation procedures questions which cannot be easily answered with a single study and be resolved using meta-analysis. There are two examples within this site, one a reprint from Lyons & Woods, (1991), and the other an as yet unpublished meta-analysis by Lyons (2003). Other efforts by the author include a quantitative review of the effectiveness of behavioral approaches in encouraging recycling. Other approaches include integrating quantitative review methods with such theory building/confirmatory methods as causal modelling and path analysis.

Posted by cmayaud at 10:19 AM | Permalink| Comments (8)
Del.icio.us Tagging | Digg This | Posted to Blogging | CONTROVERSY OF THE WEEK | THE DUNG 500 INDEX | TOPIC OF THE WEEK | TRENDS | The ACTUAL Blogosphere | Web 2.0

Comments

Whoaa! Good job Christian.

Gotta sit down and get through all this.

Posted by: DanielNerezov at October 18, 2005 11:05 AM

This is a thoroughly interesting analysis. A couple of points. My sense is the 'blogosphere' doesn't exist as such. More like a series of blogocircles. If you look at the likes of Scoble, Rubel, Israel and maybe a handful of others, they're cleverly marketing to each other in a sort of link whoring way. They've built traffic but who cares?

Gapingvoid is a good example where Hugh McLeod has settled on three (at the moment) areas of potential economic interest and used the small number of high profile blog linkers to pimp product. It's a great idea for the bespoke tailor and jeweller that benefits from a global audience or which needs to increase business by a factor of less then x1. The wine pimp blog effort? Not so sure though it makes for interesting analysis.

When I speak with people who are relatively new to this stuff, they really don't care about these guys. Instead, they're much more concerned with getting attention from an incumbent client group who will hopefully perceive them in non-traditional ways. They (and I) believe that spreading the word among a relatively small group is what matters.

Scoble agrees. He says it doesn't matter how many people link or how many hits you generate as long as those hits and links are to people that are of interest to you. He cites BusinessWeek picking up on what he says. I 100% agree. In my case, I've deliberately targeted an existing community where there are 800K page views per month. I want a sub-set of those - around 5K and guess what? With a little finagling from a content perspective, I can get between 30-50% to at least look at a landing page. From there, it is a relatively simple matter of hooking them in with the content I generate because I've already got a fair idea what 'floats their proverbial boat.'

The link analysis you've produced is therefore highly valuable to me because now I can get a sense of what I have to do. I want to be the channel master for that segment. So I absolutely want to show I'm at that left hand side - for my segment. And because I understand that segment, I should be able to differentiate the content. If I don't then I've failed.

Posted by: Dennis Howlett at October 19, 2005 12:51 PM

Very thorough analysis, Christian. Thank you.

Anita

Posted by: Anita Campbell at October 24, 2005 11:50 PM

Very interesting. This underscores my argument that the Long Tail image is misleading and inappropriate - and that it only got started in the first place because of a tendency to see the blogosphere in terms of ranking ("he* gets the most links, and he* gets the next-most...").

Although it's hard to tell from these charts, these data points represent many blogs with identical ranks superimposed on each other

That's the trouble with tied ranks - 5,000 blogs ranking equal 9,000th, followed by 10,000 ranking equal 14,000th, do not a Long Tail make. I'd go further and say that it's the problem with ranking: it'd be more sensible to plot some data on the X axis, surely. Number of links (X) vs number of blogs (Y), perhaps - but that would put the Technorati 100 way over on the right, no longer a commanding clump of skyscrapers but a scattering of outlying fenceposts. And who's going to put their money into fenceposts?

*Usually.

Posted by: Phil at November 9, 2005 06:32 AM

Great post...a main point I take away here is that rankings become impossible when you get way into the "long tail," because assuming that the data fits a power law, the integer values will have to repeat themselves, and there's no way to rank blogs that all have the same value (in this case incoming links).



Just as a side note, I think it's important to separate a bunch of different steps that sometimes get confused:



- the choice as to what data you're going to turn into a histogram

- the fit of the histogram data to some curve, e.g. a power law

- the conclusions that can be made based upon this curve and the axes defined in the first step

I point out some common misunderstandings in the first and third steps here, here, and here. A discussion with Phil regarding the second step is here.

Posted by: Adam Marsh at November 9, 2005 08:55 PM

I'd like to see the rest of this series, did you ever write them (I cannot locate any others)?


  • The Blind Men Individually (Technorati, Google, ...)

  • The Blind Men Comparatively

  • The Blind Men in Aggregate

  • Questions We Can Answer About the Elephant

Posted by: Lance at January 23, 2006 01:58 PM

I've always been bemused by Technorati, which seems to have turned blogging (for some, not all bloggers) into a kind of surreal game.


Blogging is not a contest, in my opinion. It is about writing and a passion for a particular topic.


Technorati rankings try to turn it into something where there is less interest in writing about something with passion and more about just collecting links--like being the high scorer on a video game. I've always thought the content should come first. If people want to read what you write, so much the better.


Technorati is a perfect example of how data corrupts. The fact that you can measure something turns the analysis into that which you can measure, rather than trying to understand the real factors (which may be highly qualitative and hard to measure, but much more relevant).

Posted by: Andrew at February 21, 2006 12:39 PM

Christian,

I never thought about blog sphere from this angle. If I understood you correctly there is blog inflation dead blogs, incorrect statistics (or data for it) and as usual a rule of majority as an indicator of truth.

I think its because managerial science doesnt consider a message content as an element of managerial structure (doesnt matter of which kind linear, functional, matrix, network). Consider any network. What we can see? - Some objects and connections between them. But we cant see which connections distribute messages and in what time. From this perspective any network has different structure in a particular time, because of message content, which defines (or form) the managerial structure.

If to go further and consider meaning of message content, different meanings can differently and sometimes surprisingly form the managerial structures, including hidden ones, that difficult or impossible to manage. It is a text virus effect (in my terminology) that causes fast or slow damage of any sorts of managerial structures. Ill not bother you with these specific things, which is usually a subject of my consulting services and researches. But I think the licensing model Im trying to describe might fix the blog sphere inflation to a certain extent.

For example, according to this model, connections arise between authors if a fact of derivative work creation takes place (living connections). When someone decides to buy the right to use the work for making money, a cash (in order to be distributed) will go through the connections between authors which works led to the creation the work with sold rights (living connections again). A free distribution of the works under this licensing model will also create living connections. They are not only the links, but links with concrete human actions to study the work, to test it, to derive money from it, etc. So this can be considered not only as a licensing model (a description of actions), but as a author-user-entrepreneur sphere news, networking and payment project. This might add value for global public, which GPL, Creative Commons and others cant add. Consider what Richard Stallman said about GPL and Creative Commons. I can provide additional explanation if needed.

What do you think? Does it make sense from your point of view?

Since Im not software guy I have a dilemma with this issue should I find business partners to prepare and launch a web application for this project (centralized model) or to think about business relationships to allow people to create their own web applications compatible with this licensing project for running their own business?

Nikolai

n_k at au.ru

Posted by: Nikolai Krjachkov at February 26, 2006 12:06 PM

Post a comment

Thanks for signing in, . Now you can comment. (sign out)

(If you haven't left a comment here before, you may need to be approved by the site owner before your comment will appear. Until then, it won't appear on the entry. Thanks for waiting.)


Remember me?