Alexa Changes Measurement Methodology

No longer reliant on toolbar traffic

[anallawalla]

We've been working hard behind the scenes for several months and are excited to announce that our new Alexa Rankings system is available.

Announcement link. [alexa.com]

We might have to find new things to say about the new and improved Alexa. There are no hints about where the new data sources lie, but I'd guess that aggregate data sharing arrangements have been made with large traffic sites such as ISPs and large portals.

[travelin cat]

Well that explains the almost 60% jump my main site made yesterday. I know Alexa is almost meaningless, but it does look good seeing the change in my browser.

[himalayaswater]

Heh looks good; many advertisers still use alexa before making final decision.

[travelin cat]

I think anyone contemplating the purchase of a domain would look at Alexa as another metric to consider.

[jdMorgan]

> I know Alexa is almost meaningless, but it does look good seeing the change in my browser.

If they are now aggregating domain-access data from ISPs and/or backbone providers rather than relying solely on their own Alexa toolbar data, then that means their numbers will be a lot less "meaningless."

Jim

[travelin cat]

Jim,

I hope (and think) you are correct. I'm loving my new number.

[jhood]

This is indeed good news. Alexa's ranking now almost perfectly matches Quantcast.

[rogerd]

I hope this is an improvement. On one fairly busy site (quantcast top 5K), I saw massive declines in Alexa rankings that were completely in opposition to my actual traffic logs. I pretty much stopped believing them after that.

The new numbers do look more accurate. In addition, it seems like they flushed the old bogus data. Good moves all around.

[radix]

I don't know if the new metrics data is reliable for sites with traffic from countries other than the US (or Canada, etc.).

I have a site that has been in Alexa's 200,000-300,000 range rank for years. The US traffic has been diminishing, traffic from neighboring countries (talking about Central Europe) growing on that site.

Quantcast shows almost the same rank as pre-change Alexa (in the 200,000-300,000 range).

Now Alexa ranks the site to about 1,000,000 (1 wk. avg.) and 600,000 (3 mos. avg.).

I see the same drop of rank for a competitor in the same traffic range with visitors from the same geo area.

However, leading sites of the country (in the 2,000-3,000 ranking range) did not get the drop.

[BillyS]

Cool - we moved up 50%. I have to say that most of these sites underestimate my traffic - we get four times the traffic usually published.

[radix]

Also, there is no real historical data available. The oldest one can get covers only 6 months. Will they add more months/years as the results of the new calculation roll in?

[gethan]

RogerD: "I hope this is an improvement. On one fairly busy site (quantcast top 5K), I saw massive declines in Alexa rankings that were completely in opposition to my actual traffic logs. I pretty much stopped believing them after that."

Observed almost the exact same thing - the new rankings are still lower than quantcast and my traffic suggest - but closer.

> Will they add more months/years as the results of the new calculation roll in?

I read somewhere that this will happen - they've probably purchased historical data as well.

[kaz]

The announcement says -

We are recalculating historic traffic data and will continue to add it over the coming weeks. We apologize for the inconvenience, but should have several years of data back on line shortly.

[TammyJo]

So true!

[turbosaab]

This could turn Alexa into something much more useful. Compete.com utilizes similar aggregate data, and for a free service it is actually quite revealing, even for relatively low-volume sites (the ones that Alexa doesn't graph because they aren't in the Top 100,000).

[steveb]

If they use the toolbar at all, it's still crap data. Compete and Quantcast are way better, and also have more specific data (Compete especially).

[Guido]

I'm not sure if they updated my data but
domain A: 30 u/v and 100 pages a day = rank 280,203
domain B: 400 u/v and 2700 pages a day = rank 277,964
It wasn't making any sense before, nothing changed.

[mcneely]

Ah yes, the mighty Alexa toolbar.
Just another piece of bloody backweb rubbish. A slow and cumbersome piece of adware.

Must not be working for them (the toolbar), if Alexa has to be throwing in with the likes of Aboutus, Compete and Builtwith.

[jojy]

I think alexa should provide tracking code to website owners.. this is the best way to gather website stats.. just like google analytics or quantcast do..

[travelin cat]

One thing I have noticed with this change is that our ranking has not changed once since the new algo came out. We used to be updated at least once a week previously.

[jojy]

i agree with you.. no ranking change from last 2 weeks..

[webvan]

slightly OT but might be related to that change, just noticed Alexa was not showing any new data since 05/19, any reason for that, can't see an existing topic about that.

[jojy]

I would like to update you here.. I disabled ia_archiver in robots.txt and my alexa traffic rank ramp up suddenly. I had 65,000 rank and now its 27,100. I am sure this is not due to alexa updates but this is the blocking of robot. This robot checks your website performance and several stats. I still have same traffic flow and page views. You should give a try.

[gibbergibber]

Webvan is right, Alexa hasn't updated for any sites since May 19. Strange.