Forum Moderators: bakedjake
CompWorld
Am I missing something here
Yes... that Alexa is almost irrelevant ;-)
The wayback archive is broken, the related sites are anything but, the feedback they feature is often self-posted, and the rankings are so easily manipulated that they are almost irrelevant.
Get a couple of cheap PCs, visit your site 3 times a day with each of your PCs. You can go from 500K to under 20K in a week.
This may be urban legend, but somebody once commented that the visits only count if they occur via the Alexa toolbar.
In any event, something like that proposed above happened to my site. I provide an evaluation copy of my software product there. Hackers posted a registration key for the product on several "crack" sites in China and elsewhere. My website got bombed with traffic for a week or so.
As a result, my Alexa position went from >300K to <100K and has since slowly drifted back to its old position.
if they occur via the Alexa toolbar
That's taken as assumed. Without the Alexa toolbar installed Alexa wouldn't know that you've been to that site.
You are affecting your own tracking I am afraid.
Hitwise is more accurate - just costs ten grand a year :-D
Dixon.
CompWorld
Small sample sizes are problematic because they are influenced by individual behavior. If the site owner and two employees install the toolbar and surf daily, the site rankings may see a big bump. This isn't as much of a problem for very high traffic sites.
Dissimilar sample populations affect results if the two groups install the Alexa toolbar at significantly different rates. A site geared to webmasters, for example, might see more Alexa toolbars installed than a site for elementary school teachers.
In short, I think Alexa is probably not a bad tool if you are comparing NBA.com and NFL.com, or CNN.com and ABCNews.com. For low volume sites, though, it's not particularly meaningful.
Thank you Dr. Heisenberg.
As extension, though, it should be pointed out that even when comparing popular sites only the reach number has any meaning.
When Alexa computes the overall ranking they combine reach (unique visitors) with unique pageviews (at about a 2/3 to 1/3 ratio). So if NBA.com and NFL.com get the same number of visitors but one has a lot lower usability requiring users to load many more pages to get the same detail of information, the lower usability site will show a much lower Alexa rating.
Putting pageviews into the rankings is one of the reasons forum sites get such low rankings.
Some questions though:
1)Is it true to reason Alexa works the way it works is because it is assumed that “link patterns” reflect “user patterns” on the www? (or at least largely reflects it)
Or is there some other reason(s) it works the way it does?
2)If yes to (1), how can Alexa claim that its rankings are accurate between rank 1 to 100,000 because you simply do not have access to various web logs / reports from webstats software like Webtrends which really can show actual user behaviour?
(Note: I have read Alexa’s disclaimers about how rankings above 100,000 can be considered distorted due to a variety of factors)
Still boggled - and Happy New Year!
b_ter
I think Alexa is being a wee bit optimistic if they claim accuracy even below 100K rankings.
Alexa actually has a good record of user behavior - but only for those users that have the toolbar installed. They have to extrapolate for everyone else. The smaller the sample, the riskier extrapolation is.
Links don't affect Alexa rankings or traffic numbers.
Sid
