Forum Moderators: open
Thank you very much in advance.:)
Over the past few months Alexa has been like a new toy. It's easy to look at it's flaws & opportunity for abuse - but I don't see these as significant vs what the alternative is (nothing)
If it was trying to predict actual traffic numbers then the confidence level would not be so high vs the sample size - but in terms of creating a ranking of sites it is very useful.
This is a real tool for competitive analysis which will only become more relevant with time. In many ways it is revolutionary because so far nothing has existed to easily gather competitive data for the small guy - which is essential to drive competitive behavior on the web. Better data = better decisions = more money
just my 2cents
Could one argue that there are more Google toolbar searches than Alexa toolbar searches?
Could one argue that Google searchers are better or more deeper searchers and that therefore the suggested other sites might be more relevant?
Yeah and not only google. May be we'll see a new kind of game where the players are the toolbar se's ... mine is bigger. However, as long as the stats are calculated the same way as alexa does, they'll allways only show stats of THEIR users!
I highly doubt that a google bar user also has the alexa bar installed and vice versa. I don't mean the seo / se savy webmasters but the average toolbar user that installs a toolbar. The numbers of visitors, page views and stuff is just based on the particular toolbar usage and so just a bigger or smaller part of all users. If all se's that distribute a toolbar would calculate statistics of their usage, we probably would see the same popularity ratio like the numbers in our referer reports.
... if i'm not totally wrong ...
na, not me ...
>The traffic graphs have nothing to do with searching.
>They are actual site traffic recorded by people using
>the Alexa Toolbar.
That's exactly what i mean ... didn't say it clearly ...
I say the toolbar distributers only show the stats of their customer usage ... the machines where their toolbar is installed and so the stats are nonsense if you want to have a reall overview over your site's usage.
<added>slightly off topic: the first db driven site i programmed a few years ago had no double authentication check for the admin area. I just checked for the user field from the http header info and granted access if the username was okay and a saved session id (without any timeout) was found in the database. I was pretty surprised that after a while i saw hit's coming from alexa and crawling my admin area back and forth authenticating with a user names of some of my editors ... phhh, ... they had the toolbar installed while working in their admin area ... first i thought "dirty spyware" and after thinking a bit more i just changed my bad programming style)</added>
Totally inaccurate. I think the number of users for this search engine is too low to make valid comparisons.
If Google had such a service however....
I just cant accept it while so many Korean sites dominate the top 20. Korea is amongst the top 5 or 10 of online counties by population, but where are the German language sites and Canadian sites where online populations are of similar magnitude, and Italian sites when Italy which has a substantial number also?
Unless they have reasons for keeping this data confidential, it would not be hard to develop it.
"Number of downloads" does not cut it; many people download things, use them for a few days and then delete, or their computer crashes and they install again.
Seeing Alexa "calls home" whenever its used, providing data like:
1. No of unique users a day
2. No of page views a day
3. Rough breakdowns vis-a-vis geography (known IPs, or .com/.kr/.jp/.cn/.de etc)
would be a piece of cake, and would allow us to better assess the validity and reliability of the rankings.
If Alexa really wants to say their data is in any way valid this is the least they can do, and I cant see any reasons, on the top of my head, why they cannot do this, as its not that competitive intelligence, other than their sample base is still not yet ready for prime time or that they think people are assuming they have more users than they really are.
They use Google for the search engine.
They say 7 million have installed the toolbar. Compare that to 2million (last public figure) for Google, and less than 10,000 for most of the net ratings firms.
> Korea is amongst the top 5 or 10 of
The big difference is that there are very few big sites in Korea. Everyone is using the same set of isp's and start pages.
Ah that may explain it Brett. Plus also a very tiny minority speak English or Japanese or Chinese in south korea, so their choice of sites is not as great as English speaking populations too I would say, who can read sites from many English speaking countries - US (although Americans do speak and spell it funny /duck), Canada, NZ, UK, Aust, Sing, Philippines, Malaysia, and English sites in countries where English is a major second language.
Still not 100% convinced, but getting there. Info from Alexa on their magnitude and distribution of hits would be even more convincing!
yeh your right.
>2million (last public figure) for Google
any idea where that number was published?
Alexa rankings should be shown to your clients or advertisers if they are good compared to your competitors but trusting the rankings to believe that a good ranking means more number of visitors compared to your competitors would be like living in a world of illusion.
However, I am not disagreeing with Brett too. I respect him for his expertise and since he has mentioned that Alexa rankings are good for sites with rankings in the range of 1K-2K, I do not find any reason not to believe him. His own site is in this range and he should know better. I think if any other WebmasterWorld member has site in top 2K rankings, he/she can also confirm it.
My first guess is that you have to get up into the high traffic regions for the data to be consistent and good - for sites ranked 100K+, the visits from toolbar-equipped users may be insufficient to give a statistically representative profile. Any thoughts?
(logs say 1.65 page views per user)
This is not enough data to reliably translate Alexa rankings into visitor stats but if a couple of others would care to post equivalent data, where known, we might be able to get a rough idea.
My own observation is that 2 stats packages reporting on the same log files won't even produce the same results, so Alexa has little chance of measuring like a micrometer. I also note that the Alexa graph can fluctuate by 10s of thousands even if the daily stats don't fluctuate much at all. You might need a sample of longer than a week to get a realistic indicator.
Given all that, I like it. Alexa seems great for getting a rough idea where the alternative is no idea.
Last time I checked the average time spend online for Korean users was over twice as much as for american users. As I remember it the average time spend was approx 7 or 9 hours a month for US users and 17 for Korean!
