Alexa 500's share of global traffic

SEO analytics

[dark_knight]

Hi, Does anybody have a rough idea of how much of the total traffic on the internet is generated by the "Alexa 500" websites?

Thanks.

[Quadrille]

Count the number of people with alexa toolbars, multiply by three, and subtract the number you first thought of :)

Seriously, I don't know - but I do know that it's not a representative sample of traffic, and therefore success in the "top 500 sites visited by would-be-geeks with Alexa Toolbars does not mean anything much in the real world.

[caveman]

The question is not about Alexa's user base. The question is, looking at the traffic of the *Web sites* that happen to be in Alexa's Top 500, what percent of total Web traffic do those 500 Web sites represent?

It's a good question, and often comes up in Internet media buy discussions, when large brands are trying to reach a significant percentage of total Web traffic/users as easily and efficiently as possible.

Now, if only we had the answer. ;-)

[dark_knight]

Yeah, to be clear what I really want to know is how much traffic is going strictly to the top 500 sites, and what's left for the rest of the web's 8 million other sites.

Even just to know what that # is for North America would be nice,
But I think the answer to this one is deep inside a vault inside a freedom-hating corporation.:)

[Quadrille]

But they are *not* the 'top 500' sites, they are the *alexa* top 500.

The user base matters because it is a self-selecting, skewed sample, and therefore cannot, in any way, be projected against Internet use as a whole.

Therefore any relationship between alexa faves and the real world is entirely coincidental.

[nomis5]

OK,knock Alexa as always. But don't those top 100 look like they might just be the top 100?I don't see my site there and neither do you. You see Yahoo, Google, Facebook, Bebo etc etc and lets face it, they ARE getting more hits than my site. If not, I'm being robbed!

[Quadrille]

Sure, they may well be 'the' top five. But you or I could have made our own list and we'd probably be just as close. That's what a coincidence is.

It's not about 'knocking' Alexa - it's about showing those who take it literally that it does NOT represent anything. It just doesn't.

If you want to base your estimates of Internet usage on the tastes of semi-geeks who think yet another toolbar is kewl, that's your choice. All I ask is that you make such a choice with your eyes open ;)

[edited by: Quadrille at 9:54 pm (utc) on April 13, 2008]

[nomis5]

OK. But does that mean that no one knows what the top 500 sites are? If so dark_knight's question (ignoring Alexa) just can't be answered?

Somewhere, some article convinced me that the top 50 sites took a ridiculous amount of the traffic, around 95%. The top 1,000 took over 98%. Does that ring a bell with anyone?

[Quadrille]

Nielsen Net Ratings are much more reliable than anything else you are likely to find - Nielsen is the company that sends media company shares through the roof is they smile, through the floor if they sneeze.

Not much help to us little folk, but always interesting and never bettered, to my knowledge.

[dark_knight]

Thank you for that clarification.

Does anybody know, then, what is the traffic share of the top 500 (or even 100) sites as per Nielsen Net Ratings. I just hope it ain't 98%, coz that would be freaky! ;)

[mhhfive]

well.. you could try to do an extremely rough calculation... that will be completely just sort order-of-magnitude... that said, here goes:

if you take this (very old) formula:
DUV = 7000000*AR^(-0.732)
where DUV == (Daily Unique Visitors)
and AR == (Alexa Rank)

and you sum the DUV for AR[1,500]...

you get a number like 116 million.

which you can try to compare with "Internet population" numbers -- which, for the US, was like 180 million in 2007.... which would be like 60-some percent of traffic...

However, the AlexaTop500 isn't all US.. so comparing it to the global Internet population for the top dozen or so countries.. which is like 700 million... so the AlexaTop500 would be 116/700 = 17% of the unique visitors or something like that..

Obviously, I'm using tons of numbers that aren't really derived from the same sources.. so take with an enormous grain of salt...

[dark_knight]

Thanks mhhfive,

That's actually a pretty helpful answer. So not as overwhelming as previously thought, which is encouraging.

[mhhfive]

the devil is in the details.. and my calculations may be pretty far off... That DUV formula is admittedly pretty sketchy and outdated... especially now that the Alexa rankings are computed differently now.

For example, just try the DUV for the top site. DUV(1)=7million

The top site supposedly has 7million unique visitors. I dunno how true that is... and then that kind of error propagates from 1-500..
DUV(2)=4.2million
DUV(3)=3.1million
...
DUV(500)=74,000

just giving you a reality check on these really bad estimates...