PR vs. Alexa rankings

[BigFish]

In a nutshell, our sites' homepage has a PR 6, about 6000+ on Alexa. But our two main competitors have PR5 homepages, but on alexa, they are damn near the top 1000th websites in the world.

I know there isn't THAT much difference between PR6 and 5, but this is contradictory in nature and seems odd to me.

[IITian]

A big concern, common to most sampling methodologies, is the range of estimate. Alexa gives a point estimate, of say 73,425 but it would be more responsible on its part to add the range of estimate too by stating something like: Estimated rank of 73,425 with 95% chance that the actual rank is between 9,123 and 2,345,675.

Personally I think Alexa is a useful tool, albeit currently its usefulness is limited to top 1000 or so (more likely 100) sites.

[ogletree]

IT is good for some areas. Gameing, Korean, SEO.

[europeforvisitors]

I have Alexa installed, as does many webmasters. If this theory were true than _every_ site we visit on a daily basis would already be 66% closer to being ranked within the top 20,000? (Our hit, plus the webmasters own hit)

I didn't see any major jump in my Alexa traffic ranking after I installed the Alexa toolbar. Nor have I witnessed any dramatic changes on the occasions when I've been out of town for a couple of weeks and haven't used the Alexa toolbar. (I'm usually somewhere in the 30,000 range, having climbed there slowly but steadily from the 40,000s as my traffic has grown.)

In any case, even if someone could skew the Alexa rankings with a macro or some other trick, so what? Most publishers aren't going to bother. Also, the real strength of Alexa isn't its ability to show who's a player; it's the ability to show who definitely isn't a player.

Here's an example: Let's assume that a PR person is trying to decide whether to invite the editor of antarctica-travel-articles.com on a press trip to the Ross Ice Shelf. If antarctica-travel-articles.com ranks 100 in Alexa, that may or may not be proof that the site can deliver a significant audience for the PR agency's client. The PR person is going to have to use some judgment, just as he or she would have to do if Alexa didn't exist. But, if antarctica-travel-articles.com has an Alexa traffic ranking of 289,324, the PR person can be fairly certain that the site has relatively few readers--and that it shouldn't be included on the PR agency's short list for further vetting.

Bottom line: Alexa is just a tool. It's not perfect, but it doesn't have to be perfect--it merely needs to be useful, and it is useful for some purposes (such as the example I've just given).

[imgonnaseo]

I didn't see anyone commenting on the relativity of the alexa toolbar. if you are looking at the alexa rankings as absolute numbers then you are not understanding what the rank numbers mean. each rank is simply relative to the sites above and below. you cannot tell the difference in actual traffic between ranks. for instance, the #100 ranked site might have 3 million pageviews per day while the #101 site could have 1 million pageviews per day or it could have 2.9 million pv/d. you cannot tell. all you know is one site gets more than the other. because of this, the alexa rankings are only useful to see trends against competitors. then you can sort of see if you are gaining or loosing ground, although you cannot tell by how much. its value is completely subjective and its information is absolutely relative. certainly, any assumptions gleaned from the alexa info should be corroborated with other indicators.

[Brett_Tabke]

> 1 ip once per day for a given site.

No - 1 ip per page view - not per site.

> agent: alexa

No, alexa does not id itself. Only certain partners that distribute the toolbar with their id have the agent=alexa moniker. You must specifically check for it with an activex construct.

Until google has the guts to release their own toolbar data, Alexa is the best measure of traffic on the web today.

[Dave_Hawley]

It could be the best "elephant spray" in the universe and it would still be useless.

Now you are getting close to understanding my point! Yes it is useless and yes it's the best there is. Just like Alexa!

John, Alexa can only rank sites by taking the info from those *using* their toolbar. You tell me what that % is of all people in the world that use the WWW. I bet it's such a low % the figures mean zip.

You can also explain to me how a site that gets 100 hits per day could be ranked around 10,000. If that does't ring any alarm bells nothing will.

To think their system is useful in analysing WWW makes you your own worst enemy.

Dave

[JasonHamilton]

Brett_Tabke,

I don't know where you're getting your information from, but I've installed alexa many times on many computers, over several different versions of windows, and EVERY SINGLE ONE identifies itself as part of the browser string sent to the web server. I'm talking about the toolbar downloaded directly from Alexa.com without changing any properties or settings in the toolbar. Far as I can tell, it does this by default from the only place I know of to download the toolbar.

Example of my own browser hitting a website:

x.x.x.x - - [12/Nov/2003:21:52:24 -0600] "GET /img/d_s.gif HTTP/1.1" 304 - "http://x.com/" "Mozilla/4.0 (compatible; MSIE 6.0; Windows NT 5.1; Alexa Toolbar; .NET CLR 1.1.4322)"

No doubt what Alexa does is far from being completely worthless, however any system that allows for a single user to change its ranking so greatly, you must always take what Alexa says with a grain of salt.

[JasonHamilton]

Please forgive the text formatting, but here's a quick grep of my apache logs for the last few days showing other Alexa users on one of my websites:

4 "Mozilla/4.0(compatible; MSIE6.0; WindowsNT5.1; Q312461; YComp5.0.0.0; Alexa Toolbar)"
30 "Mozilla/4.0(compatible; MSIE6.0; WindowsNT5.0; AvantBrowser[avantbrowser.com]; Alexa Toolbar)"
33 "Mozilla/4.0(compatible; MSIE6.0; WindowsNT5.1; .NETCLR1.1.4322; Alexa Toolbar)"
<snip>

[edited by: Brett_Tabke at 12:19 pm (utc) on Nov. 13, 2003]

[lgn1]

Alexa is only usefull if you are comparing Oranges to Oranges. I use Alexa to check my direct competitors. At least that way, you are baselining for geographic biases and other factors for your industy.

Also I generally look at long term trends. People that inflate Alexa results by having all their friends install the Alexa Toolbar, ussually get bored after a few weeks, as your Alexa Ranking has no impact other than boosting your ego. I rather boost my bottom line.

[europeforvisitors]

Also I generally look at long term trends. People that inflate Alexa results by having all their friends install the Alexa Toolbar, ussually get bored after a few weeks, as your Alexa Ranking has no impact other than boosting your ego.

Exactly, and that's obviously one reason why Alexa says the 3-month average is the number to watch.

[IITian]

Brett_Tabke

Alexa beats every ranking system on the net hands down. It is without-a-doubt the best available public measure of website quality there is out there today. no doubt.

Checked Alexa's top-ranked English language sites. In the top 75 could not find a single .edu or a .gov or even most brand-name companies, but look what I found. A partial list of sites in the top 75 - sites, at least, I wouldn't have expected to be there. ;)

15. www.Premium-offers.com
28. www.trafficsyndicate.com
32. www.trafficmp.com
36. www.revenue.net
38. www.kinghost.com
46. www.internetfuel.com
56. www.exitfuel.com
67. www.internet-optimizer.com
70. www.smni.com
73. www.paypopup.com
74. www.adserver.com

[Brett_Tabke]

Yep, those services do massive traffic numbers don't they. A popup is as much a pageview as any other page on the net. That is a pretty good indication that the system is working as it should.

> and EVERY SINGLE ONE identifies itself

That is just the partner or Alexa itself you are getting the toolbar from.

Alexa on and running right now:

HTTP_USER_AGENT=Mozilla/4.0 (compatible; MSIE 6.0; Windows 98; TeomaBar 2.01)

[google.com...]

[europeforvisitors]

Yep, those services do massive traffic numbers don't they. A popup is as much a pageview as any other page on the net. That is a pretty good indication that the system is working as it should.

I'm not at all sure that counting pop-ups is a good thing, but such counting certainly isn't exclusive to Alexa. Remember when MediaMetrix counted X.10 as one of the top Web properties because X.10 ads were popping up everywhere?

On a related note, FORBES had an interesting article in 2000 about how About.com was boosting its traffic ratings via popunder browser windows at Hitbox.com. The FORBES article, titled "Lyin' Eyeballs," is still online at:

[forbes.com...]

Moral: If there's a weakness in a ranking method, you can be sure that some person or company will exploit it!

[Namaste]

the differance between Alexa & PR is that Alexa updates it's rank regularly...PR was last updated by Google in 1899.

[BigFish]

Two very important points not addressed by anyone (or that nobody by Alexa themselves know)?

1. Since the feeling is that Alexa exclusively uses its toolbar users to gauge website traffic, then HOW MANY actual users of this toolbar are there?

I don't think these arguements will stop until that is answered.

2. HTTP_USER_AGENT=Mozilla/4.0 (compatible; MSIE 6.0; Windows 98; TeomaBar 2.01)

Isn't your TeomaBar just "overriding" Alexa's toolbar here?

[ronin]

Whoever said that Alexa's ranking service is unique... there is definitely at least one other service out there (which I also use): ranking.com. [Mods please remove if against TOS].

While I use both services to give me rough benchmarks, I find that neither give enough info about the size of their data samples, nor methods of acquiring data (does Alexa only use the toolbar?), nor estimated accuracy of their estimated rankings.

I think any company that attempts to provide this kind of service, should display a very large, prominent public health warning about the relative accuracy of their rankings so that people who don't know a great deal about the web will not be duped into believing claims of phenomenal popularity by dodgy companies who then seek to use their own manipulated rankings to 'independently verify' those same claims.

[JasonHamilton]

At the same time, you must keep in mind that if they provide their methods to the public, that makes it infinitely more abusable.

So they don't publish everything. While publishing their +/- error rates would not likely increase abuse of their system, it would likely negatively impact public opinion of their rankings.

[europeforvisitors]

Whoever said that Alexa's ranking service is unique...

I don't know. Did anyone? :-)

there is definitely at least one other service out there (which I also use): ranking.com.

There are other rating services, too. Each uses a different technique, and not all measure the same thing. (For example, some measure the total traffic for a company's "Web properties," which is like measuring the total magazine circulation of Condé Nast Corporation when somebody wants to know the circulation of GOURMET.)

Unfortunately, the rating services that get the most respect (whether rightly or wrongly) are available only by very expensive subscriptions, making them impractical for most Web sites and for all users.

I think any company that attempts to provide this kind of service, should display a very large, prominent public health warning about the relative accuracy of their rankings...

Every Alexa traffic report includes a prominent link to just that:

[pages.alexa.com...]

[Brett_Tabke]

>users of this toolbar are there?

7mil daily, 20 mil installs.

Compare that to a reported 9,000 boxes for competing "Ratings" services.

>bar

No, mine has _never_ reported itself as alexa (I would notice such things).

> don't know. Did anyone

No, I don't think they did. There is nothing comparable to Alexa other than Google on the web today.

[John_Creed]

Whoever said that Alexa's ranking service is unique... there is definitely at least one other service out there (which I also use): ranking.com. [Mods please remove if against TOS].

Alexa is "unique" in that it offers a useful service and doesn't really have any serious competitors(That I know of).

[Dave_Hawley]

7mil daily, 20 mil installs.

...and there are how many PC's?

[JasonHamilton]

With the amount of antivirus / antispy software out there that removes Alexa, I must account for about 2 million of the downloads when I put Alexa back on my computer.

[Dave_Hawley]

Alexa is "unique" in that it offers a useful service and doesn't really have any serious competitors(That I know of).

Ever wondered why nobody else bothers? Most know that it is simply not possible to rank sites by traffic. Google would be about one of the few (if not the only one) that could gauge this with any degree of accuracy. This is only because they deliver a large % of traffic to a large number of sites, Alexa are not even close to Google in this regard.

One of the sites that Alexa claims to be ranked #15 returns a "HTTP Error 403 - Forbidden" doesn't that send you a loud and clear message?

I'm stunned to read here (from so called Web masters) that they give credence to this sham from Alexa. I feel very sorry for the owners of these sites you guys "master". I bet you also make claims like "We will get you in the top 10 results for your search terms" :)

[Brett_Tabke]

Actually, alexa may have a higher installed base of toolbars and Google. Amazon is a powerful distribution point and they use it effectively while staying off everyones radar.