According to Alexa, "Dean has peaked...on way down...Edwards moving up"

[AP]

This is VERY unscientif, as the poster admits. Nonetheless, I thought the net-heads here would be interested.

From the Edwards blog:

I love being part of the solution
Posted by Smucci
Wednesday November 26, @12:19AM

I hope this is not too off subject because I love the concept of counting our blessings. It's just that I was playing around on the alexa site this morning and have been excited all day to share it with you all. We are part of the solution and it feels so good.

Alexa is a web site, which has tools which track activity on web sites. Their information is based on individuals who have their toolbar downloaded to their browser. It doesn’t work for AOL accounts or Netscape or Apple but it does represent a large audience. I consider their data as similar to polls. It doesn’t have everyone’s information but represents a large enough sample to be able to uncover trends. The following is data I gleaned from the reports on the Edwards, Kerry, Dean, Gephart and Clark sites. These are the only ones with enough traffic to be interesting.

TRAFFIC RANK - the lower the number the better
                  3 month avr. 1 WK Avr Today (11/25)
Edwards 44,485 15,581 15,059
Dean 3,506 4,973 6,078
Clark 10,599 7,094 8,548
Kerry 23,255 28,722 40,119
Gephart 23,223 51,731 54,306

1. Edwards has moved from an average of 44,485 to 15,059, which is reflected in both the weekly and daily numbers. This is tremendous growth.
2. Dean has gone from 3,506 to nearly 5,000 for the weekly and 6,000 for the daily. While the numbers are great the momentum is backwards.
3. Kerry’s backward momentum is almost equal to Edwards forward motion
           
REACH PER MILLION VIEWERS - larger the number the better
                        3 month avr. 1 WK Avr Today (11/25)
Edwards 30 45.5 55
Dean 277 250 260
Clark 123 185 175
Kerry 65.5 53.5 50
Gephart 56.5 29 25

1. Edwards is moving upward
2. Dean is moving backwards
3. Kerry is moving backwards

REACH RANK - Lower number is better
                        3 month avr. 1 WK Avr Today (11/25)
Edwards 40,178 27,640 22,729
Dean 3,988 4,773 4,829
Clark 9,627 6,639 7,092
Kerry 18,407 23,610 28,073
Gephart 21,456 43,353 49,794

1. Edwards show tremendous momentum forward
2. Dean is either stagnant or backward
3. Kerry is falling behind

Page View per user - Larger number is better
                      3 month avr. 1 WK Avr Today (11/25)
Edwards 3.6 14.3 9.5
Dean 5.2 3.8 2.6
Clark 3.7 3.5 2.8
Kerry 2.4 2.8 1.9
Gephart 4.3 2.9 3.4

        This is the number of pages a visitor to the site views. It is clear that while there are, as yet, fewer visitors once there the Edwards site is generating interest. Everyone else is going backwards.Page Views Rank - Lower number is better
                3 month avr. 1 WK Avr Today (11/25)
Edwards 47,557 10,335 13,236
Dean 3,656 6,199 10,520
Clark 12,755 9,278 14,447
Kerry 35,237 41,928 72,170
Gephart 31,224 73,479 72,922

1. Edwards has gone from 47,557 to 10,335 for the week and today still reflects dramatic improvement
2. Dean is going backwards and when considering the quality (number of pages viewed) with the quantity (number of viewers) is not remarkably ahead of anyone but Kerry & Gephart
3. Kerry is having a bad day.

Well, I’m sure there is someone on the blog who can create the link to this site. There are fun things you can do with graphs and comparisons, which I didn’t go into here. I’ve heard it said that there are liars, damn liars and statistics. Clearly placing importance on any one set of numbers (other than the election return) is inadvisable. However, I believe it’s safe to say that these numbers reflect what I am seeing on the street. That is, Dean has peaked and is on his way down, Kerry is in trouble and Edwards is moving up rapidly.

[Gloria]

1. Interesting stuff!!

eom

[pruner]

2. I'm not even gonna get into the many problems with Alexa…

but you should note that Dean's stats are only for http://www.deanforamerica.com

his blog is run on a separate site: http://www.blogforamerica.com

[DemDogs]

31. Everyone's blog is on separate sites I think

So that is not much of a reason for the decline in hits on the campaign site.

[AP]

32. Alexa measures traffic on mirrors too.

And if you look at their breakdown on stats, you see that they measure where people go on the sites, and the blogs were listed for all the candidates I looked at.

And the best evidence that the data is good is the price they charge for fuller details. Presuming that price results in sales, there's a value to it which must reflect its accuracy.

If anyone were really intrepid, they could check Amazon's SEC filings and see if Alexa generates reasonable revenue. That would be a pretty good measure of the quality of the data.

[pruner]

34. Alexa is the bottom of the barrel (Re: commercial traffic measurements)

their software only measures traffic from people using the Alexa broswer, which represents a very small segment of the internet population.

their results are just about as scientific as a Wolf Blitzer poll.

[AP]

35. Alexa's opt-in format, measuring actual live users probably makes it MORE

accurate.

If you believe that a poll of 400 NH voters tells you something about the relative strength of Dean and Kerry, why do you think Alexa is not a good measurement of web activity.

At the very least, you know it's not measuring spider programs or DNS attacks, and things like that.

And, again, if it were so worthless, why can Amazon charge decent prices for the information it collects?

[pruner]

38. whatever…

you don't know wtf you're talking about.

I almost feel bad for you… having to grasp at straws like this to try & bring Dean down.

[bearfartinthewoods]

50. denial........

juxt xan't standto see any indication that dean isn't beloved.

like de-nile...his support mightjust prove to be a mile wide and an inch deep. there is a linit to how many red faaced, spit spewing rants one can handle.

[AP]

51. that was a comment about Alexa, mostly. Not really about Dean.

"grasping at straws"? Is that what you call trying to make an argument based on facts and logic?

[AP]

39. Pruner, maybe I missed something, but...

...why's a poll of 400 voters less reliable than Alexa?

And do you think the Amazon customers buying the data are having the wooled pulled over their eyes?

Don't you think Amazon has applied the same principles of sampling and polling to Alexa that, say Zogby applies to his polling data?

It doesn't take much effort. All the tools and logic and forumlas are out there, and a little effort makes the data much more reliable.

And I could just as easily say that I'm sad that you're so nervous about your candidate's viability that you feel the need to say any old silly thing to discredit something that's obvioulsy meeting a decent threshold of reliability.

[pruner]

40. a better analogy would be…

a poll of 400 people who use rotary dial phones… which would represent a very small and unrepresentative percentage of the overall population.

the day after the Clark campaign announced that their one-day traffic had surpassed Dean's traffic on that day (about a month ago) Mathew Gross (Dean's webmaster) e-mailed reporters traffic reports from some of the widely-respected traffic monitoring companies they use. those results flew in the face of what Alexa was reporting. he also included a detailed explanation as to the deficiencies in Alexa's methodology (which are many… in addition to the fact that they only monitor Alexa users).

unfortunately I don't have a link handy to any of the articles that discussed this.

I would suspect that most of the companies using Alexa's commercial traffic measurement services are relatively small and can't afford any of the bigger and more repectable companies.

I can practically assure you that Amazon doesn't rely on Alexa for their traffic reports.

[AP]

41. Well, that's your best argument. Why didn't you say so earlier?

But I would like to see more info before I draw my conclusions.

[pruner]

42. I hear ya…

I wouldn't take my word for it either if I were you (you specifically… not you in the general sense).

;)

[AP]

36. Alexa's opt-in format, measuring actual live users probably makes it MORE

accurate.

If you believe that a poll of 400 NH voters tells you something about the relative strength of Dean and Kerry, why do you think Alexa is not a good measurement of web activity.

At the very least, you know it's not measuring spider programs or DNS attacks, and things like that.

And, again, if it were so worthless, why can Amazon charge decent prices for the information it collects?

[AP]

37. Alexa's opt-in format, measuring actual live users probably makes it MORE

accurate.

If you believe that a poll of 400 NH voters tells you something about the relative strength of Dean and Kerry, why do you think Alexa is not a good measurement of web activity.

At the very least, you know it's not measuring spider programs or DNS attacks, and things like that.

And, again, if it were so worthless, why can Amazon charge decent prices for the information it collects?

[pruner]

33. Clark's blog just recently switched to a unique url…

but Kerry's blog is located on a sub-domain, and so traffic would be counted among the main site's traffic.

[jobycom]

3. At least amongst voters who visit web pages

It's very hard to get a feel for much off the Internet alone, just as it is in an office or a factory, because the medium itself tends towards homogeneity. Certain types of people work in offices, or factories, or visit the web. There could be a strong growth for Dean or any candidate amongst segments of the population that either don't use the Internet much, or wouldn't visit political web pages.

Great research, though. I wouldn't have thought to look at numbers like that. It means something, but I'm not sure what, exactly.

[dsc]

4. It might simply be a function

of the fact that he is doing such a good job of emailing us and keeping us in the loop that we aren't visiting his site that often. I go to my local Dean site and my email more often than the official site.

[AP]

5. The other candidates also send out email. And if Dean's sending out so

much email that he's discouraging visits to the web site, there's another interpretation: people don't like being spammed and they're intentionally avoiding the site.

[pruner]

6. John Kerry's grass roots are the REAL DEAL

[dsc]

8. actually it is possible to keep a person informed

without spamming. That distinction may elude you but it does exist.

[madfloridian]

7. That Dean, he just keeps on peaking....on and on.

:evilgrin:

[caledesi]

9. Yeah, not matter what they say Dean has the $, the "mo" and

the IT factor.

[Scott Lee]

48. He's like the dang Energizer Bunny

peaking....and peaking....and peaking....and peaking.....

[LuminousX]

10. FYI on Methodology

Dean traffic is divided mainly between two domains: deanforamerica.com and blogforamerica.com. Many other candidates have multiple core sites and several ancillary sites. The only way to truly rate a candidate in this way is to identify those sites and aggregate them.

[AP]

11. What's difference between blog.deanforamerica.com and blogforamerica.com?

http://www.alexa.com/data/details/traffic_details?q=&url=www.deanforamerica.com

Where do people go on deanforamerica.com?   (what's this)





deanforamerica.com ~ 52%


photos.deanforamerica.com ~ 13%


deanlink.deanforamerica.com ~ 9%


blog.deanforamerica.com ~ 6%


action.deanforamerica.com ~ 4%


linux11331.deanforamerica.com ~ 3%


blogcgi.deanforamerica.com ~ 3%


ballot.deanforamerica.com ~ 1%


letters.deanforamerica.com ~ 1%


pledge.deanforamerica.com ~ 1%


lists.deanforamerica.com ~ 1%


images.deanforamerica.com ~ 1%


school.deanforamerica.com ~ 1%


Other websites ~

[LuminousX]

12. blogforamerica.com is a seperate domain

that mirrors blog.deanforamerica.com but that traffic isn't registered at Alexa in the deanforamerica.com total.

Compariason 1

Now, unfortunately, I can't create concrete proof because we don't have the actual traffic numbers to do a percentage on the deanforamerica.com traffic and see if it is the same as blogforamerica.com traffic. I can only hope I can get corraboration from other web people that Alexa tracks the domain traffic and it doesn't matter if it is mirroring a subdomain of another domain.

[AP]

14. Presumably, when the mirror updates the site it's mirroring, the traffic

gets registered to bog.deanforamerica.com.

That's activity on the site. blogdeanforamerica.com is generating traffic for blog.deanforamerica.com.

No?

[LuminousX]

19. No.

The mirroring doesn't occur using http, it is an internal process on the server so isn't registered in the logs. At least that is the way it works on my servers. I doubt the Dean campaign, a campaign that was looking for LISPers, would use a system less efficient than the one I employ, but they might.

[AP]

22. According to Alexa:

http://pages.alexa.com/prod_serv/traffic_learn_more.html#traffic_rank

What are sites and Web hosts?
Traffic is computed for sites, which are typically defined at the domain level. For example, the Web hosts www.msn.com, carpoint.msn.com and slate.msn.com are all treated as part of the same site, because they all reside on the same domain, msn.com. An exception is personal home pages, which are treated separately if they can be automatically identified as such from the URLs in question. Also, sites which are found to be serving the "same" content are generally counted together as the same site.

Of course, there's no way to know if this means they've grouped together the mirrors in this particular case, but, presumably, mirroring is what they're talking about.

[LuminousX]

24. That would indicate mirroring

But since we don't know if it has been done in this case or not, it highlights the caveat of relying upon Alexa for political analysis.

[AP]

26. There are a thousand reasons not to rely on Alexa for political analysis

but that one wouldn't be ranked at the top, from my perspective.

I think, all things considered, it's a probably pretty good measure for internet traffic. And judgig from the prices they seem to be able to get for their more detailed information, it seems like people find it reliable, or, at least, valuable.

[CMT]

13. the only response I can say

is that for months people have said Dean has peaked and will start to fade and it hasn't happened yet.

[AP]

15. You can only vote for one person.

If some people are going up, others have to go down. The issue now is who's taking the votes of whom, and how far will it go.

[Sean Reynolds]

17. And polls show it's Kerry and Lieberman that are going down.

Dean still keeps growing his national lead as well as key state leads. Kerry, Lieberman continue their fall. IMO Edwards it will be a Dean/Clark/Edwards race.

[Nazgul35]

21. ignoring the fact that in most polls....

there is about 20% undecided....you belief that it has to come from someone else is not entirely correct....the undecideds could be making up their minds....if you look at some of the polls from the earlier primary states, there is a definate trend....one that favors Dean....in most states Dean started out in single digets, but as the primary gets closer, Lieberman drops, Clark Edwards and Dean go up with Dean being competitive in all primary states......

That's why national polls are silly....even primary polls are unreliable and those later states are not going to be making choices based on the full 9 candidates because by the time they get to vote, the race will be down to two (hopefully) or three candidates....

If Edwards doesn't hit in the early primary states....he's toast....and this time around...roughly 85% of the delegates will be selected by mid February...giving even greater importance to Iowa and NH....

[helleborient]

16. It would be better if the thread was accurate...

Instead of steeped in hyperbole for Edwards and borderline falsehood about Dean.

For example, look at "Reach per million viewers" It is analyzed as Dean moving downward when the actual numbers reported are 277-250-260 - looks pretty stable to me. There is a 10% change from the 3 mos. ago to 1 wk. average, but the most recent day points a possible jump regaining 3% of that amount.

In the rush to hyperbole for Edwards...it is rarely pointed out that even with gains, his numbers as reported here with the claims for what they mean imply that he is still much worse than Dean.

The claims that page views per user imply quality of a website are questionable at best...particularly in an age where solid website design is accepted to be design in which the structure is relatively flat - meaning users get a great deal of data upfront without having to navigate a number of pages to find what they want. Technologies such as CSS2 and JavaScript help enable more elegant menuing structures that get people where they need to go in fewer pages. The incredibly large 14 pages per viewer for Edwards, a total outlier compared to the other candidates, would imply to me there are difficult to navigate Edwards sites out there...not some claim about quality.

If a user can get to a site and find what they want quickly in a few pages - that's what I call site quality!

[beyurslf]

18. As for the number of pages visited

This is a direct reflection of what's on the first page.

On Dean's website, I can read several full articles on the deanforamerica.com first page, including letters frmo Joe Trippi and Howard Dean. I know I only go occassionally to be updated on the number of volunteers (over 510,000 now). I go to see the bat when it is up. Otherwise, I get updated by emails, which also include the full article in them.

When I went to Edwards site, I had to click hyperlinks to other pages to read every single article. While this creates more "page touches," it doesn't mean someone takes a more in-dpeth look at his site--it is simply the only way to read anything fully.

[AP]

20. A higher percentage of people go to the blog on Edwards's site than

on Dean's (curiously). When you're in the blog, you click through lots of pages, just because that's how blogs work. I think that's where Edwards's page view numbers are getting the bump.

I think the bat is the reason dean's are low. I think lots of people just check the bat, and don't do anything else.

[helleborient]

23. What are these numbers...and where do they come from?

The click-through numbers to the blogs?

[helleborient]

25. That's why most analyses of real web traffic use unique visitors...

As the true measure of a web site's success.

This is the key figure for most website stat packages, and I don't see that published anywhere on this thread.

40,000 visits to a site means something quite different if 4 visitors went there 10,000 times instead of 20,000 people twice

[eissa]

27. Dean peaked again?

I guess people keep saying this hoping they'll be able to say "I told you so!" IF this ever occurs.

[helleborient]

28. Yep, you notice how Dean's people don't engage in...

The all-too-common downgrading expectations right before something significant. We're a campaign of hope...always looking for the next milestone.

[douginmarshall]

29. What about Clark

I find it interesting that there was no mention of Clark. As I read the numbers His site is the one with the highest numbers that is growing. Clark has only been is the race for 2 months and he has more visits to his site and has raised more (I think the money is true) money than everyone but Dean, and he is gaining on him. It is clear that the General is the one on the move.

[AP]

30. From memory...

if you look at the stats, Clarks site visits spiked after he announced by have settled int a range that has a tiny downward slope, but is still high (about 8,000). It still seems like the big stories here are the huge trailing off at the Kerry site (I think I read that it now requires registration for the blog, which killed traffic), and the huge increase for Edwards (from 40,000 to 10,000) in the last three months.

[cryofan]

43. Kucinich is ALSO moving up

Kucinich is right there with Edwards:
http://www.alexa.com/data/details/traffic_details?
&range=6m&size=medium&url=www.kucinich.us#top

[AP]

44. This graph overlaps Edwards and Kucinich

They've moved in tandem, with Edwards a bit higher. That's fascinating.

http://traffic.alexa.com/graph?w=379&h=216&r=6m&u=kucinich.us/&u=johnedwards2004.com

[Freddie Stubbs]

46. Despite an increase in traffic to DK's website, his popularity is stagnant

Perhaps people visit his website and are turned off by what they see there.

[Freddie Stubbs]

45. Dean has been peaking for quite a while

There have been numerous post her on DU for the past few months about how Dean is peaking too early. I'm sure the other candidate would love to be peaking the way he has.

[Scott Lee]

52. If I peaked as much as this guy, I'd be the neighborhood gigolo

[Scott Lee]

47. "Don't fret, folks...just a little bit of ice" - Capt., SS Titanic

I wonder what early death they'll be predicting after Dean becomes the nominee?

[Tinoire]

49. Kucinich is surging past them all except Edwards- both steady even

http://traffic.alexa.com/graph?w=379&h=216&r=6m&u=kucinich.us/&u=


You can run your own comparison charts here :evil grin:

http://www.alexa.com/data/details/traffic_details?&range=6m&size=medium&url=www.kucinich.us#top


This is the moment! http://www.kucinich.us/thisisthemoment.htm

Thank you Cryofan!